Credit Unions: Greater Transparency Needed on Who Credit Unions  
Serve and on Senior Executive Compensation Arrangements 	 
(30-NOV-06, GAO-07-29). 					 
                                                                 
Legislative and regulatory changes have blurred distinctions	 
between credit unions and other depository institutions and	 
raised questions about the tax-exempt status of credit unions.	 
This report (1) assesses the effect of the Credit Union 	 
Membership Access Act on credit union membership and charters,	 
(2) reviews the National Credit Union Administration's (NCUA)	 
efforts to expand services to low- and moderate-income		 
individuals, (3) compares rates offered by credit unions with	 
comparably sized banks, (4) discusses unrelated business income  
tax issues, and (5) assesses transparency of credit union senior 
executive compensation. To address our objectives, we obtained	 
NCUA data on credit union membership, charter changes, efforts to
target those of modest means, and executive disclosure		 
requirements. We also analyzed Federal Reserve Board's Survey of 
Consumer Finances and Internal Revenue Service data.		 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-07-29						        
    ACCNO:   A63824						        
  TITLE:     Credit Unions: Greater Transparency Needed on Who Credit 
Unions Serve and on Senior Executive Compensation Arrangements	 
     DATE:   11/30/2006 
  SUBJECT:   Banking regulation 				 
	     Credit unions					 
	     Economically depressed areas			 
	     Executive compensation				 
	     Financial analysis 				 
	     Income statistics					 
	     Income taxes					 
	     Internal controls					 
	     Lending institutions				 
	     Loans						 
	     Tax exempt organizations				 
	     Tax exempt status					 
	     Transparency					 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Product.                                                 **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-07-29

   

     * [1]Report to the Chairman, Committee on Ways and Means, House of
       Representatives

          * [2]November 2006

     * [3]CREDIT UNIONS

          * [4]Greater Transparency Needed on Who Credit Unions Serve and on
            Senior Executive Compensation Arrangements

     * [5]Contents

          * [6]Results in Brief
          * [7]Background
          * [8]NCUA Rules Interpreting CUMAA Appear to Have Fueled Expansion
            of Community-Chartered Credit Unions

               * [9]NCUA Regulations Interpreting CUMAA Set the Stage for
                 Growth of Community Charter Credit Unions
               * [10]Growth of Federal Community Charter Credit Unions Has
                 Been the Result of Charter Conversions Rather than New
                 Credit Union Charter Approvals
               * [11]According to NCUA, Community Charters Allow Federal
                 Charters to Remain Viable and Can Provide Opportunities to
                 Diversify Membership

          * [12]NCUA Programs Target Individuals of Low- and Moderate-Income,
            but Limited Data Preclude an Evaluation of Actual Service to
            Those Individuals

               * [13]NCUA Encouraged Growth of Low Income Credit Unions and
                 Adoption of Underserved Areas, but Cannot Quantify Impact of
                 Programs on Use of Credit Union Services by Individuals of
                 Modest Means
               * [14]Federal Reserve Survey Data Suggest that Credit Unions
                 Continued to Serve a Lower Proportion of Low- and
                 Moderate-Income Households than Banks
               * [15]NCUA Has Pilot Survey to Measure Income of Credit Union
                 Members, but Will Not Be Able to Use Results to Determine
                 What Services Members at Various Income Levels Receive

          * [16]Credit Unions Offered Better Interest Rates on Some Products,
            but the Extent to Which the Benefits of Tax- Exempt Status Have
            Been Passed to Members Is Unclear
          * [17]Lack of Guidance and Criteria for Applying UBIT to State
            Credit Union Activities Makes Determining Compliance with the
            Requirement Difficult

               * [18]IRS Has Questioned Which State Credit Union Activities
                 Should Be Subject to UBIT
               * [19]Group Filings Make It More Difficult for IRS to
                 Scrutinize the Activities of Individual Credit Unions

          * [20]Alternatives Exist to Improve the Transparency of Credit
            Union Executive and Director Compensation

               * [21]Public Reporting of Executive Compensation for Federal
                 Credit Unions Is Limited and Regulator Scrutiny of
                 Compensation Is Primarily Reviewed for Safety and Soundness
                 Concerns
               * [22]Industry Surveys That Address Executive Compensation Are
                 Limited, and NCUA Has Released Its Compensation Data

          * [23]Conclusions
          * [24]Recommendations for Executive Action
          * [25]Agency Comments and Our Evaluation

     * [26]Objectives, Scope, and Methodology

          * [27]Effects of CUMAA and NCUA Regulations and NCUA Efforts to
            Serve Low- and Moderate- Income Individuals
          * [28]Comparison of Interest Rates Offered by Credit Unions With
            Those at Comparably Sized Banks
          * [29]Issues Related to the Application of UBIT to Credit Unions
          * [30]Information on Transparency and Compensation of Executive
            Compensation

     * [31]Analyses of Survey of Consumer Finances Data, 2001 and 2004
     * [32]Comparison of Interest Rates at Credit Unions and Banks
     * [33]Selected Salary Surveys for Credit Union and Bank Executives

          * [34]Credit Union Salary and Other Cash Compensation Data
          * [35]Bank Salary and Other Cash Compensation Data

     * [36]Comments from the National Credit Union Administration

          * [37]GAO Comments

     * [38]GAO Contact and Staff Acknowledgments

                 United States Government Accountability Office

Report to the Chairman, Committee on Ways and Means, House of
Representatives

November 2006

CREDIT UNIONS

 Greater Transparency Needed on Who Credit Unions Serve and on Senior Executive
                           Compensation Arrangements

                                       a

[39]GAO-07-29

CREDIT UNIONS

Greater Transparency Needed on Who Credit Unions Serve and on Senior
Executive Compensation Arrangements

  What GAO Found

Since the passage of the Credit Union Membership Access Act (CUMAA) in
1998, larger community-based credit unions have constituted a much greater
proportion of the industry. NCUA has approved federal community charters
with increasingly larger geographic areas and potential for economically
diverse membership. Much of the shift toward the larger community-based
credit unions was due to conversions from other charters. NCUA's approval
of these charters appears to have been triggered by changes in the
economic environment and financial services industry and to diversify
membership to accomplish goals such as increasing service to those of
modest means.

NCUA has established the low-income credit union program and allowed
adoption of "underserved areas" to increase credit union services to
individuals of modest means. Despite increased credit union participation
in these programs and the expansion of community charters, the 2004 and
2001 Survey of Consumer Finances indicated that credit unions lagged
behind banks in serving low- and moderate-income households. NCUA
officials told GAO that, given the nascent nature of its two initiatives
and the relatively recent shift to community charters, they did not yet
expect observable changes in the data. Also, NCUA recently has undertaken
a pilot effort to collect data on the income characteristics of credit
union members. Because limited data exist on the extent to which credit
unions serve those of modest means, any assessment would be enhanced if
NCUA were to move beyond its pilot and systematically collect income data.

Based on GAO analysis, credit unions typically had more favorable rates
than banks, particularly for consumer loans. For example, credit unions
auto loans were 1 to 2 percentage points lower than similarly sized banks,
on average. However, it was not clear the extent that the more favorable
rates fully reflected the tax subsidy that credit unions receive by
tax-exemption.

The Internal Revenue Service (IRS) has been reviewing state-chartered
credit union activities (federal credit unions are exempt) to determine
compliance with unrelated business income tax (UBIT) requirements, but
such determinations are difficult due to complicated criteria and because
many credit unions file group rather than individual returns. IRS stated
that it plans to issue technical guidance in the first quarter of 2007
that the agency believes will help ensure credit union compliance with
UBIT.

Finally, credit union executive compensation is not transparent. Federal
credit unions, unlike other tax-exempt organizations, do not file
information returns, which contain data on executive compensation, with
IRS. NCUA is collecting compensation data as part of its pilot, but it is
unclear whether NCUA will conduct future reviews. NCUA officials noted a
number of alternatives that could be used to increase transparency, such
as requiring federal credit unions to provide compensation information in
call reports or require that credit unions disclose compensation data at
annual meetings.

                 United States Government Accountability Office

                                    Contents

Letter                                                                   1 
                                [40]Results in Brief                        4 
                                [41]Background                              8 
                                [42]NCUA Rules Interpreting CUMAA Appear      
                                to Have Fueled Expansion                      
                                [43]of Community-Chartered Credit Unions   10 
                                [44]NCUA Programs Target Individuals of       
                                Low- and Moderate-Income,                     
                                but Limited Data Preclude an Evaluation       
                                of Actual Service to                          
                                [45]Those Individuals                      14 
                                Credit Unions Offered Better Interest         
                                Rates on Some Products, but                   
                                  [46]the Extent to Which the Benefits of     
                                              Tax-Exempt Status Have Been     
                                [47]Passed to Members Is Unclear           26 
                                [48]Lack of Guidance and Criteria for         
                                Applying UBIT to State Credit                 
                                Union Activities Makes Determining            
                                Compliance with the                           
                                [49]Requirement Difficult                  29 
                                [50]Alternatives Exist to Improve the         
                                Transparency of Credit Union                  
                                [51]Executive and Director Compensation    33 
                                [52]Conclusions                            39 
                                [53]Recommendations for Executive Action   41 
                                [54]Agency Comments and Our Evaluation     41 
[55]Appendixes                                                             
                    Appendix I: [56]Objectives, Scope, and Methodology     45 
                                [57]Effects of CUMAA and NCUA Regulations     
                                and NCUA Efforts to                           
                                [58]Serve Low- and Moderate-Income         45 
                                Individuals                                   
                                [59]Comparison of Interest Rates Offered      
                                by Credit Unions With Those                   
                                [60]at Comparably Sized Banks              47 
                                [61]Issues Related to the Application of   49 
                                UBIT to Credit Unions                         
                                [62]Information on Transparency and           
                                Compensation of Executive                     
                                [63]Compensation                           49 
                  Appendix II:  [64]Analyses of Survey of Consumer            
                                Finances Data, 2001 and                       
                                [65]2004                                   50 
                  Appendix III: [66]Comparison of Interest Rates at        57 
                                Credit Unions and Banks                       
                  Appendix IV:  [67]Selected Salary Surveys for Credit        
                                Union and Bank                                
                                [68]Executives                             76 
                    Appendix V: [69]Comments from the National Credit      81 
                                Union Administration                          
                                [70]GAO Comments                           97 
                  Appendix VI:  [71]GAO Contact and Staff Acknowledgments 103 

                                     Tables

                                    Contents

Table 1: [72]Number, Members, and Assets of Federal Credit Unions by 
            [73]Charter Type, from 2000 through 2005                       12 
Table 2: [74]Federal Credit Union Conversions to Community Charters,       
            [75]from 2000 through 2005                                     13 
Table 3: [76]Number of Credit Unions Approved to Expand into               
            [77]Underserved Areas, 2000-2005                               18 
Table 4: [78]UBIT Income Declared and Taxes Paid by State-Chartered        
            [79]Credit Unions, 2000-2004                                   30 
Table 5: [80]Definition of Income Categories Used for Community            
            [81]Reinvestment Act Examinations                              47 
Table 6: [82]Peer (Asset) Group Definitions, Used in Comparisons of        
            [83]Interest Rates between Credit Unions and Banks             48 
Table 7: [84]Percentages of Households Classified as Using Banks or        
            [85]Credit Unions, 2001 and 2004                               51 
Table 8: [86]Median Income Benchmarks Used for Primary                     
               (Household) and Additional (Family) Analyses of 2001 and       
[87]2004 SCF Data                                                       52 
Table 9: [88]Income Categories for Primary (Household) and                 
            Additional (Family) Analyses of 2001 and 2004 SCF                 
[89]Data                                                                53 
[90]Table 10: Percentages in Each Income Category for Primary              
             (Household) Analysis by Customer Type, 2001 SCF                  
[91]Data                                                                53 
[92]Table 11: Percentages in Each Income Category for Primary              
             (Household) Analysis by Customer Type, 2004 SCF                  
[93]Data                                                                54 
[94]Table 12: Percentages in Each Income Category for Additional           
          [95](Family) Analysis by Customer Type, 2001 SCF Data            54 
[96]Table 13: Percentages in Each Income Category for Additional           
          [97](Family) Analysis by Customer Type, 2004 SCF Data            55 
[98]Table 14: Median Income within Each Income Category for Primary        
             (Household) Analysis by Customer Type, 2001 SCF                  
[99]Data                                                                55 
[100]Table 15: Median Income within Each Income Category for Primary       
             (Household) Analysis by Customer Type, 2004 SCF                  
[101]Data                                                               56 
[102]Table 16: Median Income within Each Income Category for               
                Additional (Family) Analysis by Customer Type, 2001 SCF       
[103]Data                                                               56 

Figures

Contents

[104]Table 17: Median Income within Each Income Category for

            Additional (Family) Analysis by Customer Type, 2004 SCF

Data 56

[105]Figure 1: Low Income Credit Union Growth, 2000-2005 16 [106]Figure 2:
Comparison of Income Levels of Credit Union and Bank

Customers, from 2001 and 2004 SCF  Data[107] 22 Figure 3: Regular  Savings
Account Rates of Credit Unions and

Banks with Assets of $100 Million or Less and Assets

Greater than  $1  Billion,  from  2000  through  2005[108]  27  Figure  4:
Sixty-Month New Car Loan Rates of Credit Unions and

Banks with Assets of $100 Million or Less and Assets

Greater than  $1  Billion,  from  2000  through  2005[109]  28  Figure  5:
Thirty-Year Mortgage Loan Fixed Rates of Credit Unions

and Banks with Assets of $100 Million or Less and Assets

Greater than  $1  Billion,  from  2000  through  2005  28  [110]Figure  6:
Compensation  Information  Filed  in  IRS  Form  990  35  [111]Figure   7:
Comparison of Interest Rates for Savings Products at

December 31, 2000, by Asset Size 58 [112]Figure 8: Comparison of  Interest
Rates for Savings Products at

December 31, 2001, by Asset Size 59 [113]Figure 9: Comparison of  Interest
Rates for Savings Products at

December 31, 2002, by Asset Size 60 [114]Figure 10: Comparison of Interest
Rates for Savings Products at

December 31, 2003, by Asset Size 61 [115]Figure 11: Comparison of Interest
Rates for Savings Products at

December 31, 2004, by Asset Size 62 [116]Figure 12: Comparison of Interest
Rates for Savings Products at

December 31, 2005, by Asset Size 63 [117]Figure 13: Comparison of Interest
Rates of Consumer Loans at

December 31, 2000, by Asset Size 64 [118]Figure 14: Comparison of Interest
Rates of Consumer Loans at

December 31, 2001, by Asset Size 65 [119]Figure 15: Comparison of Interest
Rates of Consumer Loans at

December 31, 2002, by Asset Size 66 [120]Figure 16: Comparison of Interest
Rates of Consumer Loans at

December 31, 2003, by Asset Size 67 [121]Figure 17: Comparison of Interest
Rates of Consumer Loans at

December 31, 2004, by Asset Size 68 [122]Figure 18: Comparison of Interest
Rates of Consumer Loans at

December 31, 2005, by Asset Size 69

Contents

[123]Figure 19: Comparison of Interest Rates of Mortgage Products at
[124]December 31, 2000, by Asset Size 70 [125]Figure 20: Comparison of
Interest Rates of Mortgage Products at [126]December 31, 2001, by Asset
Size 71 [127]Figure 21: Comparison of Interest Rates of Mortgage Products
at [128]December 31, 2002, by Asset Size 72 [129]Figure 22: Comparison of
Interest Rates of Mortgage Products at [130]December 31, 2003, by Asset
Size 73 [131]Figure 23: Comparison of Interest Rates of Mortgage Products
at [132]December 31, 2004, by Asset Size 74 [133]Figure 24: Comparison of
Interest Rates of Mortgage Products at [134]December 31, 2005, by Asset
Size 75 [135]Figure 25: Credit Union Executive Average Base Salaries for
[136]2005 77 Figure 26: Bank Executive Average Base Salaries in 2004 79

                                 Abbreviations

     ABA          American Banker's Association                             
     CRA          Community Reinvestment Act                                
     CUMAA        Credit Union Membership Access Act                        
     CUSO         credit union service organizations                        
     FFIEC        Federal Financial Institutions Examination Council        
     IRS          Internal Revenue Service                                  
     LICU         Low Income Credit Union                                   
     MSA          Metropolitan Statistical Area                             
     NASCUS       National Association of State Credit Union Supervisors    
     NCUA         National Credit Union Administration                      
     SCF          Survey of Consumer Finances                               
     SEC          Securities and Exchange Commission                        
     UBIT         unrelated business income tax                             

This is a work of the U.S. government and is not subject to copyright
protection in the United States. It may be reproduced and distributed in
its entirety without further permission from GAO. However, because this
work may contain copyrighted images or other material, permission from the
copyright holder may be necessary if you wish to reproduce this material
separately.

A

United States Government Accountability Office Washington, D.C. 20548

November 30, 2006

The Honorable William M. Thomas Chairman Committee on Ways and Means House
of Representatives

Dear Mr. Chairman:

Changes in credit union membership restrictions and the blurring of
distinctions in the products and services offered by credit unions and
other depository institutions, as well as concerns about the extent to
which credit unions are serving people of modest means, have resulted in
questions about the tax-exempt status of credit unions. Unlike banks,
credit unions are (1) not-for-profit entities that build capital by
retaining earnings (they do not issue capital stock); (2) member-owned
cooperatives run by boards elected by the membership; (3) subject to field
of membership requirements that limit membership to persons sharing
certain circumstances, such as a common bond of occupation or association;
and (4) exempt from federal income tax. As a result of recent legal and
regulatory developments, the National Credit Union Administration (NCUA)
has approved progressively larger geographicbased fields of membership for
federal credit unions, including areas as large as whole counties or major
metropolitan areas. Additionally, credit unions are increasingly offering
products and services similar to those provided by banks, such as real
estate and business loans. With the expansion of credit union membership
and products, concerns have been raised whether the tax- exempt status of
credit unions provides an unfair competitive advantage over comparably
sized banks and the extent to which credit unions remain unique in terms
of the population they serve versus that served by other depository
institutions.

Credit unions can be chartered by the federal government or by a state
government. NCUA has oversight authority for federally chartered credit
unions; the states have primary oversight responsibility for
state-chartered credit unions. NCUA also provides share insurance to all
federally chartered and most state-chartered credit unions. As of December
2005, there were nearly 8,700 federally insured credit unions--about 5,400
federally chartered and 3,300 state-chartered--with about $679 billion in
total assets.^1

Under the Federal Credit Union Act, federal charters are subject to a
field of membership requirement.^2 The act provides for three types of
federal credit union charter--single common bond, multiple common bond,
and community.^3 A single common-bond credit union has a field of
membership that comprises one group having a common bond of occupation or
association. In a multiple common-bond charter, the field of membership
comprises more than one group, with each group having a common bond of
occupation or association (within the group).^4 In a community charter,
the membership comprises persons or organizations within a "well-defined
local community, neighborhood, or rural district."

Although they are exempt from federal income tax, both federally and
state-chartered credit unions do pay other taxes at the federal and state
levels. For example, both types of credit unions pay federal employment
taxes such as social security tax on behalf of their employees. According
to the National Association of State Credit Union Supervisors (NASCUS),
most states impose real property taxes on state-chartered credit unions;

^1According to NCUA, there are fewer than 500 nonfederally insured
(privately insured or uninsured) state-chartered credit unions.
Nonfederally insured credit unions are not subject to NCUA regulation.
This report focuses strictly on federally and state-chartered credit
unions that have member deposits insured by the National Credit Union
Share Insurance Fund.

^212 U.S.C. S 1759 (2000). State credit union chartering acts typically
include similar field of membership restrictions. For example, see Tex.
Fin. Code Ann. 122.051 (Vernon 2005).

^3The Credit Union Membership Access Act of 1998, Public Law 105-219 (Aug.
7, 1998), 112 STAT. 914, amended the Federal Credit Union Act to allow
multiple-group chartering, subject to limitations that NCUA must consider
when granting charters, and limited new community charter applications to
well-defined local communities.

^4In a multiple-bond credit union, the original number of members in each
group must be fewer than 3,000, unless a statutory exception applies. 12
U.S.C. S 1759(b), (d). The numerical requirement does not apply to persons
or organizations within an underserved local community, neighborhood, or
rural district. 12 U.S.C. S 1759(c)(2). The act also contains other
exceptions to the numerical requirement, as well as a grandfather
provision. See 12 U.S.C. SS 1759(d)(2), (c)(1).

and in a few states, federally chartered credit unions are subject to
state real property taxes.^5

As part of the House Ways and Means Committee's continuing oversight of
the tax-exempt sector, you asked us to review a variety of credit union
issues.^6 The objectives of this study were to (1) assess the effect of
the 1998 Credit Union Membership Access Act (CUMAA) on federal credit
union membership and charter expansion; (2) review NCUA's efforts to
expand credit union services to low- and moderate-income individuals; (3)
compare rates offered by credit unions with rates at comparably sized
banks, as one indicator of how tax-exemption might benefit credit union
members; (4) discuss issues associated with the application of the federal
unrelated business income tax (UBIT) to credit unions; and (5) assess the
transparency of credit union executives and board member compensation.

To assess the effect of CUMAA on federal credit union membership and
charter expansion, we obtained data from NCUA regarding charter types and
membership, including data on recent new charter approvals, conversions,
and expansions.^7 To review NCUA efforts to expand credit union services
to low- and moderate-income individuals, we obtained information from NCUA
on its low-income credit union program and recent underserved area
expansions.^8 We also obtained and analyzed the Board of Governors of the
Federal Reserve System's (Federal Reserve) 2004 Survey of Consumer
Finances (SCF) to assess the extent to which credit union members were of
low- and moderate-income households and discussed with NCUA officials
their effort to measure the income levels of credit union members. To
compare rates offered by credit unions with those of

^5NASCUS Profile Credit Union Supervisory and State Regulatory Structures,
2005-2006 Edition. NASCUS represents the 48 state governmental agencies
and U.S. territorial agencies that charter, regulate, and examine the
nation's state-chartered credit unions.

^6On November 3, 2005, the House Ways and Means Committee held a hearing
to review the credit union tax exemption in which GAO discussed issues
regarding the tax-exempt status of credit unions. See GAO, Financial
Institutions: Issues Regarding the Tax-Exempt Status of Credit Unions,
[137]GAO-06-220T (Washington, D.C.: Nov. 3, 2005).

^7Federally insured credit unions are required to report their potential
as well as actual membership to NCUA. Potential membership is an estimate
of the maximum number of members that could join a credit union.
Throughout this report, the term membership refers to actual credit union
members unless otherwise noted.

^8As discussed previously, CUMAA permits federal multiple-bond credit
unions to include underserved areas in their field of membership,
regardless of size and location.

comparably sized banks, we obtained and reviewed rate data on more than 15
loan and savings products to identify any systematic differences. To
facilitate our discussion of issues related to the application of UBIT to
credit unions, we reviewed provisions of the Internal Revenue Code,
Federal Credit Union Act, judicial decisions, and Internal Revenue Service
(IRS) documents concerning the unrelated business income tax (UBIT) and
the taxation of credit unions. We also obtained and analyzed IRS data
regarding its activities to identify state-chartered credit union
activities subject to UBIT. To provide information on transparency of
compensation of credit union executives and board members, we discussed
compensation reporting requirements of credit unions with IRS and obtained
publicly available data regarding the compensation of board members and
senior executives of credit unions and banks. Appendix I provides
additional information about our scope and methodology. We conducted our
work in Washington, D.C., and San Francisco from November 2005 through
November 2006 in accordance with generally accepted government auditing
standards.

  Results in Brief

Since the passage of CUMAA in 1998 and subsequent changes in NCUA
regulations, the credit union industry has experienced dramatic growth in
the number of credit unions with community-based charters. NCUA revised
its regulations after the passage of CUMAA, making it easier for federal
credit unions to qualify for community charters that served large
geographic areas such as entire counties. From 2000 through 2005, the
overall number of federally chartered credit unions declined, but the
number of federal community-chartered credit unions more than doubled
(from 523 to 1,115). The assets of all three types of federal credit
unions grew by about $140 billion between 2000 and 2005, and the assets
held by community credit unions nearly quadrupled (from $27.1 to $104.4
billion). Charter conversions largely drove the increase in numbers of
community credit unions, with more than 90 percent of the growth in
community charters resulting from conversions by multiple-bond credit
unions. According to NCUA officials, the changes were necessary to
maintain competitiveness against more expansive membership regulations in
some state charters, enhance the safety and soundness of credit unions,
and allow credit unions to serve more diverse memberships, including
individuals of modest means.

Although NCUA has taken actions to make credit union services available to
individuals of modest means, information that directly measures the income
levels of credit union members continues to be limited. In addition to
looking at community chartering as a way to serve those of modest means,
NCUA has established two initiatives to enhance the availability of credit
union services to individuals of modest means--the low-income credit union
program and permitting the adoption by credit unions of "underserved
areas," regardless of the credit union's location. Federal credit union
participation in both of the initiatives has risen since 2000--for
example, federal credit unions receiving NCUA approval to expand into
underserved areas increased from 40 in 2000 to 641 at the end of 2005.
However, the most recent available information on the income of credit
union members--the Federal Reserve's 2004 SCF--indicated that credit
unions continued to lag behind banks in the percentage of their customers
or members that were of low- and moderate-income households. Our analysis
of the 2004 SCF indicated that 31 percent of households that only and
primarily used credit unions were of modest means versus 41 percent for
households that only and primarily used banks. As an approximation of
income levels, SCF data have certain limitations for measuring the income
characteristics of credit union members and should be interpreted with
caution. NCUA commented that because of the relative newness of community
charter expansion and observed increase in participation of the low-income
and underserved area efforts, it would take time before changes in the
income profiles of credit union members could be reflected in the data.
Because of the limitations of available data such as the SCF, NCUA has
undertaken a pilot program to estimate the income levels of credit union
members. However, because of limitations in the sample size, the survey
will not allow NCUA to make statistically valid conclusions on member
income by specific charter type or about specific services provided to
members of various incomes.

Our analysis of interest rates for 15 loan and savings products indicated
that credit unions seem to offer more favorable rates than those of
comparably sized banks, particularly for consumer loans. For example,
rates that credit unions charged for car loans averaged about 1 to 2
percentage points lower than rates offered by similarly sized banks, and
credit union rates averaged 0.4 percentage points higher for regular
savings accounts. This difference was slightly more pronounced as the size
of the institutions increased. In contrast, rates that credit unions and
banks charged for mortgage loans were virtually the same; and, in limited
cases, banks offered better mortgage rates on average than similarly sized
credit unions. However, our analysis of deposit and loan rate data does
not fully identify how the tax-exemption of credit unions might benefit
credit union members. For example, tax-exemption may enable credit unions
to reduce fees they charge for services provided to members. In addition,
credit unions can finance additional services and add to desired or
required reserves through untaxed retained income.

IRS has been reviewing many types of state-chartered credit union
activities to determine if they should be subject to UBIT. Groups
representing state-chartered credit unions and the Credit Union National
Association have stated that IRS has not provided sufficient guidance on
what credit union activities are or are not subject to UBIT. According to
IRS officials, IRS is planning to issue technical advice on credit union
activities subject to UBIT. State-chartered credit unions paid more UBIT
in 2004 than in 2000, but many state-chartered credit unions are permitted
to be included in a group rather than an individual information
return--Return of Organization Exempt from Income Tax (Form 990). The
practice of providing aggregate returns effectively makes it more
difficult for IRS to scrutinize state-chartered credit union operations to
determine whether they are subject to UBIT. However, IRS officials told us
that the additional technical advice the agency will be providing will
specify its position on credit union activites that are subject to UBIT
and should improve credit union compliance with the statute.

Executive compensation for federal credit unions is not transparent,
largely because federal credit unions are not required to publicly file
information on executive compensation. The importance of transparency and
disclosure of executive compensation have become an important topic as
highlighted by the Securities and Exchange Commission's (SEC) recent rule
making for publicly held companies.^9 Similarly, for not-for-profits, the
disclosure of such information helps support oversight of these tax-exempt
entities. However, unlike most other tax-exempt organizations, federal
credit unions are not required to provide IRS with Form 990s that contain
publicly disclosed information such as executive compensation. NCUA legal
opinions have stated that member access to credit union records is
generally a matter of state law. In those opinions, NCUA observed that
members of federal credit unions are owners of the institution, similar to
shareholders of corporations, so the members should look to state
corporate law on matters such as access to federal credit union records.
In one opinion, the NCUA attorney referred to the general rule in most
states that shareholders are entitled to inspect corporate minutes and
other records for appropriate purposes. However, it was not clear to what
extent

^9SEC Final Rule, 17 C.F.R.  Parts 228, 229, 232,  239, 240, 245, 249  and
274, Release Nos. 338732; 34-54302; IC-27444; File No. S7-03-06.

Page 6 GAO-07-29 Credit Union Membership and Executive Compensation

credit unions or their members were aware of this general right or how
difficult or easy it would be for credit union members to obtain executive
compensation data. We identified a few credit union and bank trade group
surveys that address executive compensation for their respective
industries. While these surveys provided an indication of compensation
patterns, the surveys generally had methodological
limitations--self-selected samples, small sample sizes, or incomplete
information on total compensation and benefits--that precluded a direct
comparison of credit union executive compensation with that of similarly
sized banks. As part of its pilot program to obtain information on the
income characteristics of credit union customers, NCUA has also collected
data on credit union executive compensation and has reported the average
salaries of specific positions with the credit union industry. However,
the information, which is stratified into two subsets, reported salary
information for (1) credit unions with assets less than $50 million and
(2) credit unions with assets greater than $50 million. Additionally, this
information only provides a snapshot of executive compensation for a
single time period. Finally, NCUA officials told us that they are
exploring various options to provide greater transparency in credit union
executive compensation, such as amending NCUA's call report data to
require federal credit unions to submit compensation and benefit data for
senior executive officers or requiring that credit unions make credit
union salary information available to their members for inspection at
their public meetings.

This report includes two recommendations to NCUA's Chairman to
systematically track the performance of federal credit unions in providing
financial services to members of modest means and improve the transparency
of credit union executive compensation to enhance accountability of credit
unions to the public and to their members.

We provided a draft of this report to the Chairman of NCUA and the
Commissioner of IRS for their review and comment. We received written
comments from NCUA that are summarized below and reprinted in appendix V.
In addition, we received technical comments from IRS that have been
incorporated into this report as appropriate. In its written comments,
NCUA indicated that the agency's staff have recommended that the NCUA
board consider taking actions consistent with the recommendations made in
our report. However, NCUA indicated in its comment letter that it had
concerns with certain aspects of the draft report. Specifically, NCUA's
letter stated that it was inaccurate and inappropriate to measure the
success of federally chartered credit unions in serving persons of modest
means by reference only to the low- and moderate-income categories
associated with the Community Reinvestment Act (CRA). Additionally, NCUA's
letter stated that our income category benchmarks were inconsistent with
the specific definitions used by the other federal financial regulators
for CRA compliance. NCUA's letter also stated that the SCF was not
designed for reliable income comparisons between credit union members and
bank customers. As we noted in the report, neither the Federal Credit
Union Act nor NCUA have established definitions as to what constitutes
modest means. Thus, consistent with our 2003 report, we used low- and
moderate-income households as a proxy for persons of modest means for the
purposes of our analysis. Our analysis not only included comparisons
between credit unions and banks of low- and moderate-income households but
also middle and upper income households for both the 2001 and 2004 SCF.
This analysis shows that between 2001 and 2004 credit unions continued to
serve a higher proportion of middle- and upper-income households and a
smaller proportion of low- and moderate-income households than did banks.
The primary difference between our income categories and those used for
CRA purposes was the use of national rather than local Metropolitan
Statistical Area (MSA) median income as a benchmark for the various income
categories. We use the national measure since the SCF is a national
survey, and we did not have sufficient geographic information to conduct a
MSAlevel analysis. While we agree that the SCF was not specifically
designed to conduct comparative analyses of income levels of bank and
credit union customers, the SCF provides the best data currently available
to undertake such a comparison. SCF is a respected source of publicly
available data on financial institution and consumer demographics that is
nationally representative and was the only comprehensive source of
publicly available data that we could identify with information on
financial institutions and consumer demographics. Additional NCUA comments
are discussed at the end of this report and in appendix V.

  Background
Both federally and state-chartered credit unions are exempt from federal
income taxes.^10 However, their exempt status arises from different
provisions of federal law. Federal credit unions are specifically exempt
from federal and state income taxes under a provision of the Federal
Credit

^10As previously noted in the introductory paragraphs of this report,
federally and most statechartered credit unions are also exempt from state
income and franchise taxes and pay other taxes at the federal and state
levels.

Page 8 GAO-07-29 Credit Union Membership and Executive Compensation

Union Act.^11 State-chartered credit unions are exempt under a provision
of the Internal Revenue Code that describes as exempt, "Credit unions
without capital stock organized and operated for mutual purposes and
without profit."^12 The code also imposes UBIT on state-chartered credit
unions, but not on their federally chartered counterparts.^13

The tax-exempt status of credit unions originally was predicated on the
similarity of credit unions and mutual financial institutions. Section
11(a)(4) of the Revenue Act of 1916, the statutory forerunner of section
501(c)(14)(A), exempted from federal income tax "cooperative banks without
capital stock organized and operated for mutual purposes and without
profit." The exemption of credit unions stems from an opinion of the
Attorney General, 31 O.A.G. 176 (1917), holding that credit unions
organized under the laws of Massachusetts were so similar to cooperative
banks as to come within the scope of section 11(a)(4). IRS regulations
subsequently applied this ruling to credit unions generally.^14

While other institutions lost their exemption in the Revenue Act of 1951,
Congress specifically retained the exemption for credit unions by removing
cooperative banks, savings and loan societies, and building and loan
associations from exemption and inserting credit unions in their place.^15
The Senate Finance Committee report accompanying the Revenue Act of 1951
stated that the exemption of mutual savings banks was repealed to
establish parity with other banking institutions because the savings banks

^1112 U.S.C. S 1768.

^1226 U.S.C. S 501(c)(14)(A).

^1326 U.S.C. S 511(a)(2); see also 26 U.S.C. S 501(c)(1), which, in
combination with the exemption for federal credit unions under the Federal
Credit Union Act, excludes them from UBIT.

^14See also T.D. 3179, which amended Art. 515(3), section 231, Regs. 45,
Revenue Act of 1918 to read:

"Cooperative banks without capital stock organized and operated for mutual
purposes and without profit are exempt. Credit unions, such as those
organized under the laws of Massachusetts, being in substance and in fact
the same as cooperative banks, are likewise exempt from tax."

^15Pub. L. No. 82-183 S 313.

had become functionally similar to those other institutions.^16 According
to the Senate report, tax-exempt status gave mutual savings banks the
advantage of being able to finance growth out of untaxed retained
earnings, while competing corporations (commercial banks) paid tax on
income retained by the corporation. The report stated that the exemptions
for savings and loan associations had been repealed on the same ground.
The report did not state why the tax-exempt status of credit unions was
preserved.

Credit unions are an important, but relatively small segment of the
financial industry. According to NCUA and Federal Deposit Insurance
Corporation data, federally and state-chartered credit unions represented
7.5 percent of all deposits and shares insured by the federal government
as of December 31, 2005. Additionally, credit unions typically are much
smaller than banks and thrifts in terms of total assets. For example, NCUA
data indicated that approximately 88 percent of federally chartered credit
unions had $100 million or less in assets with 83 percent having assets
less than $50 million as of September 30, 2005. According to NCUA, the
average size of a federally chartered credit union was $73.2 million in
total assets and the median asset size was $11 million.

  NCUA Rules Interpreting CUMAA Appear to Have Fueled Expansion of
  Community-Chartered Credit Unions

Since the passage of CUMAA in 1998 and subsequent NCUA rule changes, NCUA
has approved community charters with increasingly larger geographic fields
of membership--for example, covering entire cities or multiple counties.
Since 2000, community-chartered credit unions have nearly tripled their
membership and nearly quadrupled their assets. Most of the new community
charters approved between 2000 and 2005 were charter conversions by
multiple-bond credit unions rather than new credit unions. According to
NCUA, community charters offer credit unions greater opportunity than
single- and multiple-bond credit unions to diversify their membership
base, thereby contributing to the institution's economic viability and
ability to serve all segments of the community, including those of modest
means.

^16S. Rep. No. 82-781 (1951) at 22, 25.

    NCUA Regulations Interpreting CUMAA Set the Stage for Growth of Community
    Charter Credit Unions

CUMAA is the most recent statute affecting field of membership
requirements of federally chartered credit unions. In 1998, the Supreme
Court determined that NCUA had erroneously interpreted the Federal Credit
Union Act to permit federally chartered credit unions to have multiple
common bonds.^17 In response, Congress passed a provision in CUMAA to
specifically permit multiple-bond credit unions subject to a general
limitation on the number of members sharing a particular common bond. Also
in CUMAA, Congress amended the provision of the act permitting the federal
community charter by changing the description of its field of membership
from "groups within a well-defined neighborhood, community, or rural
district" to "persons or organizations within a welldefined local
community, neighborhood, or rural district."^18

Subsequent to the passage of CUMAA, NCUA revised its regulations to
approve community charters consisting of larger geographic areas of
coverage and potential members. For example, NCUA recently approved one
credit union for a community charter covering the entirety of Los Angeles
County. Thus, an estimated 9.6 million persons who live, worship, and go
to school or work in the county and businesses and other legal entities
within county boundaries qualify for membership in this credit union.^19
We reported in 2003 that previous NCUA regulations required credit unions
to document that residents of a proposed community area interacted or had
common interests.^20 Credit unions seeking to serve a single political
jurisdiction (e.g., a city or a county) with more than 300,000 residents
were required to submit more extensive paper work. However, NCUA revised
its regulations in 2003, defining a local community as any city, county,
or political equivalent in a single political jurisdiction,

^17National Credit Union Administration v. First National Bank & Trust
Company 522

U.S. 479 (1998).

^18Pub. L. No. 105-219 S 101(b)(3); see 12 U.S.C. S 1759 (1994).

^19According to NCUA officials, the 9.6 million potential membership
figure for this credit union was based on the 2001 Census Bureau estimate
of the population of Los Angeles County. Potentially, the credit union's
membership could be larger than the population of Los Angeles County since
individuals who live outside the county but worship or work in the county
would be eligible to join the credit union.

^20See GAO, Credit Unions: Financial Condition Has Improved, but
Opportunities Exist to Enhance Oversight and Share Insurance Management,
[138]GAO-04-91 (Washington, D.C.: Oct. 27, 2003).

regardless of population size and eliminated the documentation
requirements.

As shown in table 1, the number of community-chartered federal credit
unions doubled from 2000 through 2005, while the number of multiple-bond
credit unions declined by about 22 percent. In spite of the recent
decline, multiple-bond credit unions remain the largest group of federally
chartered credit unions in number and in total membership and assets.
However, community-chartered credit unions overtook multiple-bond credit
unions as the largest of the three federal charter types in terms of
average membership and average size in terms of assets beginning in 2003.

 Table 1: Number, Members, and Assets of Federal Credit Unions by Charter Type,
                             from 2000 through 2005

Federal Credit Unions       2000    2001    2002    2003    2004      2005 
Number                                                                     
Multiple Single            3,045   2,933   2,842   2,684   2,534   2,385   
                              2,512   2,401   2,256   2,106   1,987   1,893   
Community                    523     783     855     986    1,051    1,115 
All charters                6,080   6,117   5,953   5,776   5,572    5,393 
Actual members (in                                                         
millions)                                                                  
Single                       7.0     7.1     7.0     7.0     6.8       7.3 
Multiple                    30.8    29.5    29.1    27.8    26.8      26.0 
Community                    5.2     7.2     8.4    11.4    13.3      14.6 
All charters                42.9    43.8    44.6    46.2    46.9      47.9 
Assets (in billions)                                                       
Single                      $43.3   $49.4   $54.4   $58.9   $62.1    $70.8 
Multiple                   $168.8  $180.2  $196.6  $202.2  $204.7   $202.6 
Community                   $27.1   $40.5   $50.3   $75.4   $91.9   $104.4 
All charters               $239.2  $270.1  $301.2  $336.6  $358.7   $377.8 
Average membership (in                                                     
thousands)                                                                 
Single                       2.8     2.9     3.1     3.3     3.4       3.9 
Multiple                    10.1    10.1    10.3    10.3    10.6      10.9 
Community                    9.9     9.2     9.8    11.6    12.6      13.1 
All charters                 7.1     7.2     7.5     8.0     8.4       8.9 
Average assets (in                                                         
millions)                                                                  
Single                      $17.3   $20.6   $24.1   $28.0   $31.3    $37.4 
Multiple                    $55.4   $61.4   $69.2   $75.3   $80.8    $85.0 

(Continued From Previous Page)
Federal Credit Unions      2000     2001    2002    2003    2004      2005 
Community                 $51.8    $51.8    $58.8   $76.5   $87.5    $93.6 
All charters              $39.3    $44.2    $50.6   $58.3   $64.4    $70.1 

                                 Source: NCUA.

    Growth of Federal Community Charter Credit Unions Has Been the Result of
    Charter Conversions Rather than New Credit Union Charter Approvals

To a large degree, the increase in number, membership, and assets of
community charter credit unions can be attributed to charter conversions
rather than to new credit union charter approvals. Between 2000 and 2005,
NCUA approved 616 applications for federal community charters. Of these
616 approved federal community charters, 600 were conversions from single-
or multiple-bond credit unions while only 16 were for new credit union
charters. As shown in table 2, the vast majority of the conversions to
community charters--549 or about 92 percent--involved multiple-bond credit
unions.

Table 2: Federal Credit Union Conversions to Community Charters, from 2000
through 2005

Year                From single-bond From multiple-bond              Total 
2000                                       8        81                  89 
2001                                      10        83                  93 
2002                                       5        86                  91 
2003                                      14        121                135 
2004                                       5        75                  80 
2005                                       9        103                112 
Total                                     51        549                600 

                                 Source: NCUA.

    According to NCUA, Community Charters Allow Federal Charters to Remain
    Viable and Can Provide Opportunities to Diversify Membership

NCUA officials indicated that changes in chartering policy have been
triggered by factors such as the continued viability of federal credit
unions in a changing economic environment and financial industry
developments. NCUA believes that community charter expansion allows
federal credit unions to attract a more diverse membership base. According
to the officials, this in turn can enhance a credit union's economic
viability, safety and soundness, as well as provide greater opportunities
to serve members of modest means. For example, officials explained that
single- and
multiple-bond credit unions often tend to be organized around employer or
occupationally based associations, which in turn creates greater economic
risk exposure since the membership base is intertwined with the economic
cycles of a particular employer or occupation.^21 Additionally, NCUA
officials noted that employer or occupational bonds result in a greater
concentration of members with middle rather than lower incomes. Since
community charters are organized around geographically based associations,
credit unions would be able to provide individuals from a broad range of
occupations and income levels in these communities with access to their
products and services. However, community based credit unions would be
vulnerable to regional downturns in the economy.

  NCUA Programs Target Individuals of Low- and Moderate-Income, but Limited Data
  Preclude an Evaluation of Actual Service to Those Individuals

NCUA has established and increased participation in two programs and
policies that are specifically designed to make credit union services more
available to individuals of low- and moderate-income. NCUA's Low Income
Credit Union (LICU) program is designed to assist credit unions that can
demonstrate that a majority of their members have a median household
income less than 80 percent of the national household income or make less
than 80 percent of the average for all wage earners. NCUA also has made it
easier for federal credit unions, regardless of location, to expand their
fields of membership into underserved areas (areas experiencing economic
difficulty). Although federal credit unions increasingly have participated
in these efforts in recent years, lack of data on the income levels of
credit union members has made it difficult to determine how effective
these programs have been in providing services to individuals of modest
means. But, the limited existing data on income levels of credit union
customers suggest that credit unions continue to lag behind banks in the
proportion of customers that are of low-and moderate-income. NCUA has
undertaken a pilot effort to capture information on the income
characteristics of credit union members, but the data will not allow NCUA
to reach statistically valid conclusions by charter type.

^21For example, a Federal Reserve Bank of Atlanta research paper concluded
that "there are material benefits of credit union membership
diversification and that these benefits derive from expanded investment
opportunities and reduced concentration risk." See W. Scott Frame, Gordon
V. Karels, and Christine McClatchey, "The Effect of the Common Bond and
Membership Expansion on Credit Union Risk," Federal Reserve Bank Atlanta
Working Paper No. 2001-10, April 2001.

Page 14 GAO-07-29 Credit Union Membership and Executive Compensation

    NCUA Encouraged Growth of Low Income Credit Unions and Adoption of
    Underserved Areas, but Cannot Quantify Impact of Programs on Use of Credit
    Union Services by Individuals of Modest Means

As we reported in 2003, it has been generally accepted that credit unions
have a historical emphasis on serving people with "small" or "modest"
means. Congressional findings contained in CUMAA linked the tax-exempt
status of credit unions, in part, to their "specified mission of meeting
the credit and savings needs of consumers, especially persons of modest
means."^22 NCUA incorporated this emphasis into its current strategic
plan, which gives its mission as "facilitating the availability of credit
union services to all eligible consumers, especially those of modest means
through a regulatory environment that fosters a safe and sound credit
union system." According to NCUA officials, the changes in chartering
requirements should allow credit unions to serve a more diverse
membership, including those of modest means.

In addition to approving more community charters, NCUA has encouraged
credit union activity in other areas in an attempt to make credit union
services more available to low-income individuals and underserved areas.
According to NCUA, its LICU program is designed to assist credit unions
serving predominantly low-income members in obtaining technical and
financial services.^23 Credit unions that receive a low-income designation
receive certain opportunities, such as the following:

     o greater authority to accept deposits from nonmembers such as voluntary
       health and welfare organizations;
     o access to low-interest loans, deposits, and technical assistance
       through participation in NCUA's Community Development Revolving Loan
       Fund;

^22Pub. L. No. 105-219 S 2(4).

^23Section 701.34 of NCUA's Rules and Regulations defines the term
"low-income members" as those members who make less than 80 percent of the
average for all wage earners as established by the Bureau of Labor
Statistics or whose annual household income falls at or below 80 percent
of the median household income for the nation as established by the Census
Bureau.

Page 15 GAO-07-29 Credit Union Membership and Executive Compensation

     o ability to offer uninsured secondary capital accounts and include
       these accounts in the credit union's net worth for the purposes of
       meeting its regulatory capital requirements;^24 and
     o a waiver of the aggregate loan limit for member business loans.

From 2000 through 2005,  the number of  LICUs grew from  632 to 1,032,  an
increase of more than 63 percent (see fig. 1).

      Figure 1: Low Income Credit Union Growth, 2000-2005

Number of credit unions1,200

600

400

200

        0 2000 2001 2002 20032004 2005 Year

Source: NCUA.

Credit union expansion into underserved areas also has increased in recent
years. From 1994 through 1998, NCUA rules permitted federal credit unions,
regardless of charter type, to include residents in low-income communities
and associations in their fields of membership.

^24A "secondary capital instrument" is either unsecured debt or debt that
has a lower priority than that of another debt on the same asset. These
subordinated debt instruments are not backed or guaranteed by the federal
share insurance funds.

Page 16 GAO-07-29 Credit Union Membership and Executive Compensation

In 1998, CUMAA expressly recognized that multiple-bond credit unions would
be authorized to serve persons or organizations within an area that was
underserved. The Federal Credit Union Act defines an underserved area as a
local community, neighborhood, or rural district that is an "investment
area" as defined by the Community Development Banking and Financial
Institutions Act of 1994--that is, experiencing poverty, low income, or
unemployment.^25 NCUA's Chartering and Field of Membership Manual
(Interpretive Ruling and Policy Statement 03-1 or IRPS 03-1) allowed
credit unions to include underserved areas in their fields of membership,
without regard to location or changes to their charter type.^26 For
example, NCUA recently approved a credit union in the state of Maryland to
serve residents within an area of Washington, D.C., determined to be
"underserved." Between 2000 and 2005, the number of credit unions
receiving NCUA approval to adopt underserved areas grew from 40 to 641. As
shown in table 3, the largest proportion of the 641 credit unions approved
through year-end 2005 were multiple-bond credit unions (410 or 64
percent), followed by community-chartered credit unions (196 or 31
percent).

^25Section 103(16) of the 1994 act defines "investment area" as follows:
"The term `investment area' means a geographic area (or areas) including
an Indian reservation that--

(A) (i) meets objective criteria of economic distress developed by the
Fund, which may include the percentage of low-income families or the
extent of poverty, the rate of unemployment or underemployment, rural
population outmigration, lag in population growth, and extent of blight
and disinvestment; and (ii) has significant unmet needs for loans or
equity investments; or (B) encompasses or is located in an empowerment
zone or enterprise community designated under section 1391 of title 26."
12 U.S.C. S 4702(16).

^26According to NCUA's Chartering and Field of Membership Manual, a
federal credit union that desires to include an underserved community in
its field of membership must first develop a business plan specifying how
it will serve the community. At a minimum, the business plan, must
identify the credit and depository needs of the community and detail how
the credit union plans to serve those needs. The credit union will be
expected to regularly review the business plan to determine if the
community is being adequately served. The regional director may require
periodic service status reports from a credit union about the underserved
area to ensure that the needs of the community are being met as well as
requiring such reports before NCUA allows a federal credit union to add an
additional underserved area.

  Table 3: Number of Credit Unions Approved to Expand into Underserved Areas,
                                   2000-2005

    Multiple-bond federal   Community-chartered  Single-bond    Total federal 
                                                     federal    credit unions 
                                                      credit                  
       credit unions with federal credit unions  unions with with approved    
                          with                   underserved underserved      
Year-end   underserved     underserved areas        areas           areasa 
                    areas                                                     
2000                22                    17            0               40 
2001               120                    66            6              196 
2002               238                   127           14              386 
2003               314                   150           18              489 
2004               369                   170           24              570 
2005               410                   196           28              641 

Source: NCUA.

^aAs of December 2005, seven state-chartered credit unions formerly had
been federally chartered and had approved underserved areas.

However, recent changes in NCUA policies may limit the growth of the
underserved areas program. In connection with a lawsuit instituted in
November 2005, NCUA stopped permitting single-bond and community federal
credit unions to include underserved areas in their fields of membership.
This had the effect of allowing access only for multiple-bond credit
unions, which is permitted specifically in a provision of the Federal
Credit Union Act.^27 In the lawsuit, the American Banker's Association
(ABA) challenged NCUA's approval of community-chartered credit unions
adding underserved areas to their field of membership. ABA argued that
NCUA misinterpreted the Federal Credit Union Act by allowing a community
federal credit union to expand into several communities in Utah. ABA
contended that the Federal Credit Union Act allows multiplebond credit
unions, but not community-chartered credit unions, to add underserved
areas to their fields of membership. In response, NCUA subsequently
amended its chartering regulations to limit the adoption of underserved
areas to multiple-bond credit unions. NCUA's final rule, incorporating
these amendments, took effect on July 28, 2006.^28 On July 20, 2006, ABA
announced that it had agreed to dismiss its lawsuit.

^2712 U.S.C. S 1759(c)(2). ^2871 Fed. Reg. 36667 (Jun. 28, 2006).

Page 18 GAO-07-29 Credit Union Membership and Executive Compensation

Despite the expansion into underserved areas and the LICU program, NCUA
cannot specifically quantify the extent to which these programs have
increased use of credit union services by individuals of modest means. As
we reported in 2003 and will discuss in the following sections, limited
data are available that specifically measure the income levels of credit
union members and the services used by individuals of modest means. As a
result, although NCUA data indicate increased adoption of underserved
areas and increased participation in the LICU program, data do not exist
to specifically show the extent to which these programs have increased
services provided to individuals of modest means.

    Federal Reserve Survey Data Suggest that Credit Unions Continued to Serve a
    Lower Proportion of Lowand Moderate-Income Households than Banks

Despite the shift toward community charters and the increase in the number
of credit unions participating in NCUA's low-income and underserved
programs, our analysis of data from the Federal Reserve's 2004 Survey of
Consumer Finances (SCF) indicated that credit unions had a lower
proportion of customers who were of low- and moderate-income than did
banks.^29 These results were similar to the results of our analysis of the
Federal Reserve's 2001 SCF data, which we discussed in our 2003 report.^30

29The SCF is conducted every 3 years and is intended to provide detailed
information on the balance sheet, pension, income, and other demographic
characteristics of U.S. households and their use of financial
institutions. We used the term "household" rather than "family," since the
reporting unit of SCF more closely resembles the Census Bureau's
definition of "household" than its definition of "family." It should be
noted that SCF was not specifically designed to conduct comparative
analyses of income levels of bank and credit union customers; however, SCF
provides the best data currently available to undertake such a comparison.
See appendix II for more information on our methodology and analysis of
the SCF.

^30GAO, Credit Unions: Financial Condition Has Improved, but Opportunities
Exist to Enhance Oversight and Share Insurance Management, [139]GAO-04-91
(Washington, D.C.: Oct. 27, 2003). The analysis presented in this section
employs the same methodology as in our 2003 report.

We combined the 2004 SCF data into two main groups--households that only
and primarily used credit unions (credit union customers) and households
that only and primarily used banks (bank customers).^31 We then computed
the proportions of credit union customers and bank customers in each of
these four income categories--low, moderate, middle, and upper. We based
our income categories on criteria that financial regulators use to assess
compliance with the Community Reinvestment Act, which is intended to
encourage depository institutions to help meet the credit needs of the
communities that they serve. Specifically, (1) a lowincome household had
less than 50 percent of the national median household income; (2) a
moderate-income household had an income of at least 50 percent, but less
than 80 percent, of the national median household income; (3) a
middle-income household had an income of at least 80 percent, but less
than 120 percent, of the national median household income; and (4) an
upper-income household had an income of at least 120 percent of the
national median household income. We estimated that 14 percent of credit
union customers were of low-income and 17 percent were of moderate-income,
compared with 24 percent and 16 percent for banks. We found the difference
between the proportion of low-income customers at banks and credit unions
to be statistically significant (that is, the evidence suggested that the
difference between the two was not simply the result of chance). Moreover,
we estimated that 20 percent of credit union customers were of
middle-income and 49 percent were of upper-income, compared with 18
percent and 41 percent for banks. We found the difference between the
proportion of upper-income customers at banks and credit unions to be
statistically significant as well.

In an effort to assess the extent to which credit unions served people of
"modest means," we combined households with low- or moderate-incomes into
one group (as a proxy for modest means) and combined households

^31We based our methodology for determining these classifications on work
that Jinkook Lee, a professor and researcher at Ohio State University,
performed. See Jinkook Lee and William A. Kelly Jr., "Who Uses Credit
Unions?" (Prepared for the Filene Research Institute and the Center for
Credit Union Research, 1999, 2001). Individuals who "primarily" used
credit unions placed more than 50 percent of their assets in credit unions
and those who "primarily" used banks placed more than 50 percent of their
assets in banks. The term "use" refers to a household's placement of
assets in a checking, savings, or money market account.

with middle or upper incomes into another group.^32 We found that 31
percent of credit union customers were of "modest means," compared with 41
percent of bank customers, suggesting that banks served a higher
proportion of people of "modest means." The difference between banks and
credit unions was statistically significant.

As shown in figure 2, the proportion of credit union customers that were
in the upper-income category grew from 2001 to 2004. This increase, from
43 percent to 49 percent, was statistically significant. Thus, the
statistically significant difference between banks and credit unions in
serving people of "modest means" that we documented in our 2003 report
using 2001 data appears to have persisted in the 2004 data. Moreover, we
found the decline from 2001 to 2004 in the proportion of credit union
customers in the "modest means" category to be statistically significant.
Additionally, the relatively high percentage of households in the
moderate- and middleincome categories that used credit unions (37 percent)
in the 2004 SCF may be reflective of credit union membership traditionally
being based on occupational- or employer-based fields of membership.

^32As in our 2003 report, we were unable to find a definition of "modest
means." Thus, we used the group consisting of low- and moderate-income
households as a proxy for the purposes of our analysis.

Page 21 GAO-07-29 Credit Union Membership and Executive Compensation

Figure 2: Comparison of Income Levels of Credit Union and Bank  Customers,
from 2001 and 2004 SCF Data

      Percentage

Upper income

Middle income

Moderate income

Low income Sources: GAO and Federal Reserve.

Note: Totals may not equal 100 percent due to rounding.

However, NCUA officials told us that since growth in the agency's programs
to expand services to lower-income persons and undeserved areas are
relatively recent, it was probably too soon to expect any changes in the
SCF data, with respect to customer income. Further, NCUA felt that it
would take time for any results to appear in the data, as credit unions
seeking to expand into new areas and reaching new types of customers would
face a learning curve in their efforts. Additionally, NCUA officials
stated that since most of the conversions to the community charter
occurred within the last 5 years, within a reasonable period they expected
to see a change in the customers these credit unions were serving. It
should also be kept in mind that the latest available data from SCF are
2-years old, so any more recent changes would not be reflected in our
analysis.

As we noted in our 2003 report, limitations in SCF data preclude its use
in drawing definitive conclusions about the income characteristics of
credit union members. Additional information--especially about the income
levels of credit union members receiving consumer loans and other credit
union services--would be required to assess more completely whom credit
unions serve. As further noted in our 2003 report, NCUA has noted that
credit union members were likely to have higher incomes than nonmembers
because credit unions are occupationally based. As NCUA and others have
noted--because of the statutory limitations on who can join federal credit
unions--credit union membership is largely based on employment, and credit
unions are restricted to the income composition of the individuals within
fields of membership containing employed individuals. However, as we noted
earlier, SCF provides the best data currently available regarding the
income characteristics of credit union members.

To determine how sensitive our results were to our income categorization,
we used median family income in addition to median household income to
analyze the 2001 and 2004 SCF data.^33 We found similar results using both
median family and household income. Recognizing the limitations of the SCF
and other available data, our 2003 report suggested that Congress consider
requiring NCUA to obtain data on the extent that credit unions provided
loans and other services to low- and moderate-income households within
each federally insured credit union's field of membership.^34

    NCUA Has Pilot Survey to Measure Income of Credit Union Members, but Will
    Not Be Able to Use Results to Determine What Services Members at Various
    Income Levels Receive

In response to your Committee's concerns regarding the lack of available
information to evaluate credit union member income and services, NUCA
undertook a data collection effort to profile federal credit union member
income information, identify the credit union services offered to credit
union members, and provide information on the compensation of credit union
executives. (We discuss executive compensation in more detail later in
this report). As of August 31, 2006, NCUA had completed its data
collection phase, as agreed with the Office of Management and Budget

^33See appendix II for greater detail on the SCF analyses we performed. In
SCF, a household unit is divided into a ``primary economic unit'' (PEU)
and everyone else in the household unit. The PEU is intended to be the
economically dominant single individual or couple (whether married or
living together as partners) and all other persons in the household unit
who are financially interdependent with that economically dominant person
or couple. The Census Bureau's definition of "family" excludes the
possibility of one-person household units, but its definition of
"household" allows for them.

^34 [140]GAO-04-91 , p. 83.

under the Paperwork Reduction Act, which is intended to minimize the
paperwork burden for nonprofit institutions.^35

NCUA took a random sample of 481 federal credit unions and relied on two
different data collection methods to determine member incomes. NCUA
officials told us they intended to compare the results of the two methods
to determine the extent of any income differences and identify which of
the two approaches might be relied upon in the future. One method involved
obtaining information such as a credit union member's zip code from NCUA's
Automated Integrated Regulatory Examination System to make projections of
median household income. The other method involved using the street
address and zip codes of credit union members and applying a software
package that uses geo-coding to determine median family income averages.
The officials told us that the software package is widely used in the
banking industry to help make income determinations for fair lending
examinations.

NCUA also gathered information from the credit unions on the type of
services the institutions offer to their members, including services that
may be of value to members with lower incomes or little financial
experience. Using the same sample of credit unions, NCUA collected
information on whether or not certain services are provided by the credit
union. For example, NCUA gathered information on the extent to which the
sampled credit unions offer low-balance checking accounts and whether they
offer some type of financial literacy training.

NCUA officials stated to us that there are limitations of the data
collection effort. First, although the information collected represents a
statistically valid random sample of the federal credit union population
and will provide information on the income levels of overall federal
credit union members, the data will not enable NCUA to make statistically
valid conclusions by charter type or make conclusions about the extent of
credit union services being provided to various income levels. The
officials explained that to do so would require a larger and more
time-consuming data collection effort, requiring an increase in sample
from the current effort of 481 to a sample of almost 1,200 credit unions.
According to the officials, a larger sample would not allow them to meet
their goal to report results by year-end 2006.

^35The Paperwork Reduction Act of 1995 (Pub. L. 104-13, 44 U.S.C. Chapter
35) was intended, among other things, to minimize the paperwork burden for
nonprofit institutions that results from the collection of information by
or for the federal government.

Page 24 GAO-07-29 Credit Union Membership and Executive Compensation

NCUA indicated that despite these limitations, the data collected will add
to the agency's knowledge and should be valuable in deciding what actions,
if any, might be appropriate over the longer term. At the time of our
discussions, NCUA had not developed benchmarks to use as a measure for a
"modest means" category related to member income data. NCUA indicated that
its data collection effort will help the agency better understand the
concept of "modest means" in relation to geographically dispersed,
limited, and diverse fields of membership. NCUA's data collection effort
represents 61 percent of all credit unions because the regulator has
oversight authority for federally chartered credit unions, while state
governments have responsibility for overseeing the remaining 39 percent of
the credit unions (state-chartered credit unions). Finally, while NCUA's
data collection effort will be useful for establishing a baseline, NCUA
officials stated that there are no plans to gather the information on a
continual or routine basis.

In response to a March 2006 congressional request, National Association of
State Credit Union Supervisors (NASCUS) officials told us they and state
regulatory agencies have initiated a data collection effort for
statechartered credit unions. NASCUS will collect some information similar
to that collected in the NCUA pilot, such as membership income and
executive compensation.^36 However, NASCUS also will collect data in two
additional areas: credit union service organizations (CUSO) and UBIT.^37
NASCUS is using a methodology similar to NCUA's to determine member income
levels. According to NASCUS, they have developed a representative sample
by applying different weights to unique state credit union
characteristics, including size, field of membership, and charter type.
Credit unions selected in the representative sample will respond to a
questionnaire developed by state regulatory agencies. The questionnaire
addresses membership, CUSOs, UBIT and executive compensation. As of
September 2006, the officials indicated that the data collection effort
had started, and that they expected the results to be available in the
first quarter of 2007.

^36To determine executive compensation for state-chartered credit unions,
NASCUS is requesting information from credit unions in the states that
file a consolidated IRS Form 990. State-chartered credit unions are
required to file this form, which includes executive compensation
information.

^37A credit union service organization is a corporation, a limited
liability corporation, or limited partnership owned by one or more credit
unions that provides services such as insurance, securities, or real
estate brokerage, primarily to credit unions or members of affiliated
credit unions.

  Credit Unions Offered Better Interest Rates on Some Products, but the Extent
  to Which the Benefits of Tax-Exempt Status Have Been Passed to Members Is
  Unclear

Our analysis showed that credit unions tended to offer better interest
rates than similarly sized banks for a variety of products and loans, but
rate data alone cannot be used to determine the extent to which the
benefits of tax exemption have been passed to members. We obtained and
analyzed rate data for various savings products offered by credit unions
and banks from 2000 through 2005 and found that credit unions on average
offered higher rates than comparably sized banks.^38 Similarly, the rate
data that we obtained for various loan products indicated that on average
credit unions tended to offer lower interest rates than comparably sized
banks, particularly for consumer loans such as automobile loans. However,
it is important to note that interest rates during the period covered by
our analysis were at historic lows.

As seen in figure 3, rates offered by credit unions from 2000 through 2005
on regular savings accounts on average were higher than those offered by
similarly sized banks. However, the differences among the rates of
comparably sized credit unions and banks tended to get larger as the size
of the institutions increased. For example, for institutions with assets
of less than $100 million, the difference between credit unions and banks
averaged about 0.15 of a percentage point in this period, while the
difference for institutions with assets greater than $1 billion averaged
almost 0.7 of a percentage point. More recently, the gap in rates between
larger credit unions and banks closed considerably; in the greater than $1
billion asset range, the gap was more than 1 percentage point in 2000, but
about one-half of 1 percent in 2005. We observed similar trends throughout
the period for other savings products such as money market accounts and
certificates of deposit (see app. III).

^38We engaged Datatrac Corporation--a market research and information
technology company specializing in the financial services industry--to
gather and analyze data on loan and savings products that thousands of
credit unions and banks offered from 2000 through 2005. Datatrac
calculated average rates for each of the products for all institutions
(divided into five distinct asset classes) over this period. See appendix
III for more information on our rate analysis methodology and results.

Page 26 GAO-07-29 Credit Union Membership and Executive Compensation

Figure 3: Regular Savings  Account Rates of Credit  Unions and Banks  with
Assets of $100 Million  or Less and Assets  Greater than $1 Billion,  from
2000 through 2005

Sources: GAO and Datatrac.

The difference between credit unions and banks was more pronounced for
consumer loans. For example, over the 6-year period, the rates that credit
unions charged for 60-month new car loans tended to be lower than the
rates charged by similarly sized banks, by 1 or 2 percentage points. As
shown in figure 4, the trend was consistent for the larger asset category
as well. However, unlike savings products, the rate differences between
credit unions and banks for car loans widened in 2005. These trends held
true for rates for other consumer products such as credit cards.

Figure 4: Sixty-Month New Car Loan  Rates of Credit Unions and Banks  with
Assets of $100 Million  or Less and Assets  Greater than $1 Billion,  from
2000 through 2005

Sources: GAO and Datatrac.

Although credit unions charged lower interest rates for consumer loans,
similarly sized credit unions and banks charged virtually identical rates
on larger loans such as mortgages, from 2000 through 2005 (see fig. 5). In
some limited instances, banks offered lower rates than similarly sized
credit unions. Also, larger institutions in general offered lower rates
than smaller institutions.

Figure 5: Thirty-Year Mortgage Loan Fixed Rates of Credit Unions and Banks
with Assets of $100  Million or Less and  Assets Greater than $1  Billion,
from 2000 through 2005

Sources: GAO and Datatrac.

^aData based on responses of less than 10 institutions.

Our analysis of deposit and loan rate data for credit unions and banks
does not fully identify how the tax-exemption of credit unions might
benefit credit union members. First, there may be other reasons for
differences in rates beside tax differences. For example, loan rates may
differ because of differences in borrower characteristics, such as
creditworthiness, or because of geographic market differences. In
addition, tax-exemption may benefit members in other ways than through
loan and deposit rates. Credit unions might also charge lower fees than
they otherwise would for services and products provided to members. We did
not identify any comprehensive studies or data sources that addressed
differences in fees charged by credit union and banks on a national basis.
Unlike banks, credit unions can finance additional services and add to
desired or required reserves through untaxed retained income. As a result
of tax-exemption, credit unions may retain more income to add to reserves
or to finance additional services than they would if they were taxed.

  Lack of Guidance and Criteria for Applying UBIT to State Credit Union
  Activities Makes Determining Compliance with the Requirement Difficult

As stated earlier, state-chartered credit unions are subject to tax on
unrelated business income while federal credit unions specifically are
exempt. IRS has several ongoing examinations of state-chartered credit
unions to determine which of their activities are subject to UBIT. Credit
union trade groups have stated the need for guidance regarding the
activities that IRS has determined to be subject to UBIT; IRS has stated
that it plans to issue technical advice in 2007 after completing its
reviews. Furthermore, the practice of allowing group statewide filings has
made it more difficult for the IRS to scrutinize the activities of
individual institutions to ensure compliance with the UBIT statute. IRS
officials asserted that the agency plans to issue technical advice on the
application of UBIT to state credit union activities, which they stated
should improve credit union compliance with the statute.

IRS Has Questioned Which State Credit Union Activities Should Be Subject to UBIT

UBIT is a tax on income derived by a tax-exempt entity from a trade or
business that is regularly carried on and not substantially related to the
exercise or performance of the purpose or function constituting the basis
for the entity's exemption. Under the Internal Revenue Code,
statechartered credit unions are subject to UBIT, but federal credit
unions are not subject to the tax because they are exempt federal
instrumentalities under a provision of the code.^39 As shown in table 4,
the amount of income subject to UBIT reported by state-chartered credit
unions and the related taxes paid nearly doubled from 2000 through 2004
and totaled more than $5 million over this period.

Table 4: UBIT  Income Declared  and Taxes Paid  by State-Chartered  Credit
Unions, 2000-2004

Tax year                            Total UBIT incomea    Total taxes paid 
2000                                        $2,565,056            $691,094 
2001                                         3,642,903             927,896 
2002                                         9,820,709           1,281,938 
2003                                         3,788,412           1,050,767 
2004                                         4,646,782           1,377,726 
Total                                      $24,463,862          $5,329,421 

Source: IRS.

^aNegative income (UBIT losses) are not included in the UBIT income
figures.

As credit unions have increased the types of products and services that
they offer, certain product offerings by state-chartered credit unions
have resulted in IRS examining whether state-chartered credit unions are
using their tax-exempt status to conduct business unrelated to their
exempt purposes. In November 2005, an IRS commissioner informed the
Congress of an IRS review of certain activities of state-chartered credit
unions for purposes of UBIT.^40 The IRS has been reviewing activities that
included the following:

     o the sale of optional credit life insurance and credit disability
       insurance to members that would pay off the loan balances with the
       organization, if the borrower died or became disabled;
     o the sale of "guaranteed auto protection" insurance, which pays the
       automobile loan balance in the event of loss or destruction of a
       vehicle to the extent it exceeds the value of the vehicle;

^39See 26 U.S.C. SS 501, 511; 12 U.S.C. S 1768.

^40Statement of Steven T. Miller, Commissioner, Tax-Exempt and Government
Entities Division, Internal Revenue Service, in testimony before the full
committee of the House Committee on Ways and Means (Nov. 3, 2005).

Page 30 GAO-07-29 Credit Union Membership and Executive Compensation

     o the sale of automobile warranties;
     o the sale of cancer insurance;
     o the sale of accidental death and dismemberment insurance;
     o ATM fees charged to nonmembers;
     o the sale of health or dental insurance;
     o the marketing of mutual funds to members; and
     o the marketing of other insurance and financial products.^41

According to IRS officials, the agency had 50 ongoing examinations of
state-chartered credit unions for UBIT purposes as of September 2006.

Determining the applicability of UBIT to state credit union activities is
a complicated proposition because it depends on the relationship of the
activities to credit unions' tax-exempt purposes or functions. However, as
IRS stated in a Technical Advice Memorandum, neither the Internal Revenue
Code nor IRS regulations enumerate the functions of a credit union exempt
under section 501(c)(14) of the code.^42 The tax-exemption is based on
what can be described as structural features, specifically the
institution's mutuality and nonprofit operations, whether it is organized
and operated in accordance with state law, and whether its members share a
common bond.^43

Groups Representing state-chartered credit unions and the Credit Union
National Association have stated that IRS has not provided sufficient
guidance on which credit union activities are or are not subject to UBIT.
According to IRS officials, IRS is planning to issue specific information
in the form of Technical Advice Memoranda as a result of its examination
of

^41Statement of Steven T. Miller. ^42Technical Advise Memorandum 9548001
(Mar. 23, 1995). ^43See G.C.M. 34612, Exempt Status of State Chartered
Credit Unions (Sept. 14, 1971); see also

G.C.M.
           37467Exemption from Tax On Corporations - Credit Unions (Mar. 21,
           1978),

G.C.M.
           38348.

Page 31 GAO-07-29 Credit Union Membership and Executive Compensation

credit union UBIT activities in the first quarter of 2007.^44 IRS believes
that the Technical Advice Memoranda will more clearly articulate specific
activities of state-chartered credit unions that can be subject to the tax
and improve compliance with the statute.

    Group Filings Make It More Difficult for IRS to Scrutinize the Activities of
    Individual Credit Unions

Not-for-profit entities that generate more than $1,000 in unrelated
business income are required to disclose on Form 990 that they have
generated such income and whether they have filed an Exempt Organization
Business Income Tax Return (Form 990-T) with IRS declaring the income and
related UBIT liability. State-chartered credit unions are required to file
Form 990, and Form 990-T if they generate unrelated business income in
excess of $1,000. However, according to IRS officials, an IRS ruling
allows state regulatory authority in some states to file forms 990 on a
groupwide basis; that is; one form can be filed on behalf of all
state-chartered credit unions in a particular state.^45 Thus, a
state-chartered credit union located in one of those states could generate
taxable unrelated business income without having to file a Form 990,
individually. According to IRS officials, the group ruling applies to 34
states, and group returns in about 21 of those states have been filed in
recent years. Of the 21 states, only 1 has asserted on the Form 990 that
UBIT was generated in that state.

IRS stated it would be able to positively verify if an individual credit
union declaring unrelated business income in excess of $1,000 on its Form
990 also filed a Form 990-T. However, the agency currently does not have a
process in place to review group returns to ensure that the credit unions
filed the Form 990-T. As a result, IRS cannot systematically determine if
credit unions that were included in group returns and generated unrelated
business income properly filed a Form 990-T declaring such income and paid
UBIT. According to IRS officials, the group exemption process was
instituted to relieve IRS of the burden of individually processing a large
number of applications from organizations sharing a common affiliation

^44Technical Advice Memoranda are issued in response to technical or
procedural question that develop during proceedings on the interpretation
and proper application of tax law, tax treaties, regulations, revenue
rulings, notices, or other precedents published by the Office of Chief
Counsel. Proceedings include the examination of a taxpayer's return.

^45IRS requires that every year, each local organization (in this case,
credit union) authorize in writing the central organization that prepares
the group return to include it in the group return and must declare, under
penalty of perjury, that the information it submits to be included in the
group return is true and complete.

and that are operated for the same general purpose. For example, IRS noted
that some organizations, including churches and veterans organizations,
have a great number of subordinate organizations that are similarly
situated. IRS officials agreed that requiring all state-chartered credit
unions to file an individual Form 990 could enhance the agency's ability
to scrutinize the activities of individual credit unions to determine
whether they were subject to UBIT. However, officials also noted that it
was not clear if the benefits of eliminating the group filing exemption
would exceed the costs--both to IRS as well as to the individual credit
unions. Specifically, officials noted that credit unions that are
currently included in group returns would each need to file for
recognition as a taxexempt organization and incur annual costs to prepare
and file individual Form 990. Moreover, IRS officials noted that they
expect that the Technical Advice Memoranda that the agency is planning to
issue in early 2007 would improve credit union compliance with UBIT filing
requirements.

  Alternatives Exist to Improve the Transparency of Credit Union Executive and
  Director Compensation

Federal credit union executive compensation is not transparent. Federal
credit unions are not required to file reports, including the IRS Form 990
required for most other tax-exempt organizations that would provide
information on executive and director compensation. NCUA legal opinions
have stated that member access to credit union records is generally a
matter of state law but that federal credit union members "have inspection
rights similar to those enjoyed by a shareholder in a corporation" and
that "the general rule in most jurisdictions is that a shareholder is
entitled to inspect corporate minutes and other records as long as he has
a proper, nonvexatious purpose."^46 However, we could not determine to
what extent credit unions and credit union members were aware of this
information. We identified a number of credit union and bank executive
compensation surveys, but data and methodological limitations precluded us
from making direct comparisons of executive compensation. NCUA has
collected executive compensation information for federal credit unions as
part of its efforts to assess who credit unions serve.

 ^46NCUA Letters 06-0127B (Feb. 6, 2006); 96-0541 (Jun. 14, 1996); and 89-0525
                                (Jun. 8, 1989).

      Page 33 GAO-07-29 Credit Union Membership and Executive Compensation

    Public Reporting of Executive Compensation for Federal Credit Unions Is
    Limited and Regulator Scrutiny of Compensation Is Primarily Reviewed for
    Safety and Soundness Concerns

The issue of transparency and disclosure of executive compensation has
become an important topic, both for tax-exempt entities and publicly held
companies. Credit union members bear some similarity to public company
shareholders in that they are "owners" and vote for boards of directors
that are entrusted to oversee executive compensation. The importance of
disclosure of executive and director compensation was illustrated in
recent changes adopted by SEC in July 2006 to increase transparency and
disclosure by public companies and reflect the increasing focus on
corporate governance and director independence. According to SEC, the
objective was to provide investors with a clearer and more complete
picture of the compensation earned by a company's principal executive
officer, principal financial officer, highest paid executive officers, and
members of its board of directors.

In contrast, credit union executive compensation is not transparent
because credit unions are not required to file publicly available reports
such as the IRS Form 990 that disclose executive compensation data. For
tax-exempt organizations, IRS has noted that some members of the public
rely on Form 990 as the primary or sole source of information about a
particular organization. Most tax-exempt organizations with gross receipts
that are normally more than $25,000 are required to file the Form 990
annually. IRS also uses these forms to select organizations for
examinations. Figure 6 shows the compensation information collected on the
Form 990.

Figure 6: Compensation Information Filed in IRS Form 990

Source: IRS.

On August 23, 1988, IRS issued a determination that federal credit unions
are not required to annually file a Form 990 because of their status as
taxexempt instrumentalities under section 501(c)(1) of the Internal
Revenue Code. Also, as noted previously, some state-chartered credit
unions file through a group filing process (21 states in 2004). For these
states, IRS receives only the name and addresses of individual credit
unions. As a result, scrutiny of the compensation of credit union
executives and other key personnel is limited.

Additionally, boards of directors of credit unions receive limited
compensation because the directors serve in nonpaid positions. According
to the Federal Credit Union Act, no member of a federal credit union board
may be compensated; however, a federal credit union may compensate one
individual who serves as an officer of the board.^47 Although the credit
union may not pay its board of directors a salary, it may provide or
reimburse directors for such things as mileage; insurance, including life,
health, and accident insurance; and travel expenses. In contrast, bank
boards of directors may receive fees such as an annual retainer for
serving on the board, profit sharing, professional fees, and other
bonuses. Also, according to one bank survey, about half of the banks that
responded indicated that their compensation fees were based strictly upon
attendance.

According to NCUA, executive compensation is assessed during the credit
union's examination to determine its reasonableness as it relates to
safety and soundness concerns. As stated in our 2003 report, NCUA recently
moved from an examination and supervision approach that primarily was
focused on reviewing transactions to an approach that focuses resources on
high-risk areas within a credit union.^48 To complement the risk-focused
approach and allow NCUA to better allocate its resources, the agency
adopted a risk-based examination program in July 2001. NCUA officials
explained that supervisory examinations, including reviews of credit union
executive compensation, follow a risk-focused approach. The officials told
us that examiners would review executive compensation in instances of
safety or soundness concerns, such as compensation arrangements that
significantly exceeded compensation paid to persons with similar
responsibilities in credit unions of similar size and in similar
locations. NCUA also stated that since it has not found a systemwide issue
with executive compensation, it has not considered it necessary to collect
or aggregate executive compensation data. Additionally, we found NCUA's

^4712 U.S.C. SS1761(c) and 1761a. ^48 [141]GAO-04-91 , p. 42.

Page 36 GAO-07-29 Credit Union Membership and Executive Compensation

guidance on compensation similar in content with the Federal Financial
Institutions Examination Council's (FFIEC) Interagency Guidelines
Establishing Standards for Safety and Soundness.^49

At various times, credit unions and others have questioned whether members
have the right to obtain and inspect credit union information, including
salary data. NCUA legal opinions have stated that member access to credit
union records is generally a matter of state law, and federal credit
unions should look to the appropriate state corporate law. In a letter
dated June 14, 1996, NCUA's Associate General Counsel said that federal
credit union members "have inspection rights similar to those enjoyed by a
shareholder in a corporation" and that "state law determines the types of
information and documents, and the degree of access, available to
shareholders/members."^50 The letter stated that "the general rule in most
jurisdictions is that a shareholder is entitled to inspect corporate
minutes and other records as long as he has a proper, nonvexatious
purpose." However, it is unclear to what extent federal credit union
members and credit union personnel are aware of a member's right to
inspect records, or how difficult or easy it would be for credit union
members to obtain information such as salaries.

    Industry Surveys That Address Executive Compensation Are Limited, and NCUA
    Has Released Its Compensation Data

We could not identify any surveys or studies that directly compared credit
union executive compensation with compensation provided by similarly sized
banks. However, we did identify a few credit union and bank trade group
surveys that address executive compensation for their respective
industries.^51 Credit union and bank trade group officials told us that
these surveys generally are used to help their industries gauge comparable
pay by job title and an institution's asset size.

^49According to its Web site, FFIEC is a formal interagency body empowered
to prescribe uniform principles, standards, and report forms for the
federal examination of financial institutions by the Federal Reserve, the
Federal Deposit Insurance Corporation, NCUA, the Office of the Comptroller
of the Currency, and the Office of Thrift Supervision and to make
recommendations to promote uniformity in the supervision of financial
institutions.

^50Letter from Richard S. Schulman, NCUA Associate General Counsel, Re:
Aberdeen Proving Ground Federal Credit Union, NCUA 96-0541 (Jun. 14,
1996).

^51Credit Union National Association, 2005-2006 Complete Credit Union
Staff Salary Survey and America's Community Bankers, 2005 32^nd Annual
Compensation Survey.

Several limitations exist that preclude us from directly comparing credit
union and bank executive compensation. While these surveys provided
information on the cash compensation by job title, institution, and asset
size, the surveys did not provide detailed information on the other forms
of compensation received to allow a direct comparison of credit union and
bank executive compensation. Some benefits include items such as
retirement plans, stock options (for bank executives), employment
contracts, severance pay, and perks such as vehicle allowances. Due to the
lack of consistency and availability of data beyond cash compensation in
these surveys, it is difficult to make any overall comparisons between
credit union and bank executive compensation. Other limitations to these
surveys include, in some instances, low response rates for the three
executive positions (chief executive officer, chief financial officer, and
chief operating officer). Further, the data collected in these surveys
were based on self-reported information from the survey participants.
Appendix IV provides more detail on the survey limitations and the results
of executive compensation for credit unions and banks that responded to
their respective surveys.

As mentioned previously, NCUA has collected credit union executive
compensation data and reported compensation information on the topthree
executive positions--chief executive officer, chief financial officer, and
chief operating officer. NCUA collected 2005 compensation information from
the IRS Form 1099 and Form W-2 (wages and salary data). NCUA officials
told us that the sample size will enable them to project industry averages
for the federal credit union population and be stratified into two
statistically valid subsets based on the asset size of the credit unions
surveyed. However, the NCUA effort provides a snapshot of federal credit
union compensation for a single year, 2005, and it is unclear whether NCUA
will conduct future reviews of credit union executive compensation.

NCUA also suggested alternative methods of collecting compensation
information and increasing the transparency of the information. During our
review, NCUA indicated that it was considering amending the quarterly
"call reports" that all federally insured credit unions are required to
submit to NCUA to include compensation and benefit data for senior
executive officers. Call reports are available for public inspection, and
NCUA routinely reviews them. Currently, the call report collects only
aggregate data on employee compensation and benefits. Additionally, NCUA
officials indicated that requiring credit unions to disclose credit union
salary information to members during public meetings would be another
alternative for  increasing the  transparency  of executive  and  director
compensation.

                                  Conclusions

Since the passage of CUMAA and subsequent changes to NCUA regulations that
permitted credit unions to serve larger geographic areas and enlarged
fields of membership, community-chartered federal credit unions have grown
in number and asset size. As a result, the common bonds of occupation or
employment that traditionally existed between credit union members have
become attenuated, blurring one of the historical distinctions between
credit unions and other depository institutions. But, credit unions do
retain distinctions in terms of structure and governance, and they retain
their tax-exempt status.

One perceived rationale for the credit union tax exemption, expressed by
Congress, is the notion that credit unions serve individuals of small or
modest means. Yet, it is difficult to determine to what extent credit
unions actually serve individuals of modest means. Although NCUA has
established programs to expand services to this group, the relative
newness of the programs, combined with the absence of long-term,
continuing, and systematic collection of data on the income of credit
union members, currently preclude an assessment both of the programs'
effectiveness and overall industry performance. However, limited data
(SCF) suggest that in both 2001 and 2004, credit unions had a smaller
proportion of low- and moderate-income customers than banks. NCUA
officials have noted that it may be too soon for data to fully reflect
NCUA initiatives and industry activities and that growth in the community
charter will allow credit unions to draw members from larger and more
diverse populations, including people of modest means.

While NCUA has taken steps to identify the income levels of credit union
members, several limitations in NCUA's data collection effort will make it
difficult to fully assess the extent to which credit unions have been
serving low- and moderate-income populations. Notably, the data will not
stratify information about member incomes by specific charter types or
identify the specific financial services that credit union members have
been using. Obtaining more detailed information on credit union member
income and the financial services they used could help NCUA track the
performance of credit unions and help monitor progress over time.
Furthermore, this information would provide Congress and the public with
clear evidence that, as CUMAA notes, credit unions were accomplishing
their "specified mission" of "meeting the credit and savings needs of
consumers, especially persons of modest means." However, the NCUA effort,
while laudable, currently is confined to a pilot project. The value and
utility of the information collected would be greatly enhanced if NCUA
were to move beyond a pilot and continue the data collection effort and
address some of the limitations of the pilot.

Although state-chartered credit unions have increased the amount of UBIT
paid in recent years, determining which credit union activities are
subject to the UBIT is difficult. IRS is currently conducting examinations
of statechartered credit unions and plans to release technical advice
early in 2007 that the agency believes will more clearly explain which
credit union activities are subject to the UBIT. While state-chartered
credit unions are required to file information returns (Form 990), the
group that are filed constrain IRS's ability to scrutinize credit union
activities related to UBIT because they convey little information about
individual credit unions. However, IRS is planning to issue technical
advice describing specific state credit union activities that may be
subject to the UBIT to help ensure state credit union payment of the tax.

Finally, the transparency of executive compensation is an important issue
for private and public companies alike. In the private sector, SEC's
recent efforts to increase the transparency of publicly held companies
underscore the importance of enhancing accountability and greater
disclosure of information. In contrast, credit union executive
compensation is not transparent due to the lack of information available
to the public. Increased public opportunities to review executive salaries
would promote greater credit union accountability, similar to requirements
for publicly held companies. While the Form 990 is an avenue for
increasing both the quantity and transparency of publicly available
information about executive compensation at credit unions, federal credit
unions are not required to file the form. However, the public could be
given other opportunities to review credit union activities. For example,
NCUA could require all federally insured credit unions to include
compensation and benefit data for senior executive officers in the call
reports that are submitted on a quarterly basis--an option that NCUA
officials indicated was under current consideration. Or, NCUA could
require federal credit unions to disclose or make available credit union
records, such as senior executive salary information, to members during
annual meetings.

  Recommendations for Executive Action

To help ensure that credit unions are fulfilling their tax-exempt mission
of providing financial services to their members, especially those of low
or moderate incomes, we recommend that the Chairman of NCUA systematically
obtain information on the income levels of federal credit union members to
allow NCUA to track and monitor the progress of credit unions in serving
low- and moderate-income populations. NCUA's recent pilot survey to
measure the income of credit union members could serve as a starting point
to obtain more detailed information on credit union member income.
Ideally, NCUA should expand its survey to allow the agency to monitor
member income characteristics by credit union charter type, obtain
information on the financial services that low- and moderateincome members
actually use, and monitor progress over time.

To increase the transparency of executive compensation and enhance
accountability of credit unions, we recommend that the Chairman of NCUA
take action to ensure that information on federal credit union executive
compensation is available to credit union members and the public for
review and inspection. To achieve this, NCUA may want to consider options
such as requiring federal credit unions to include specific information on
executive compensation in call reports or issuing regulations that would
require all federal credit unions to make executive compensation
information available to members of credit unions at annual meetings.

  Agency Comments and Our Evaluation

We provided a draft of this report to the Chairman of NCUA and the
Commissioner of IRS for their review and comment. We received written
comments from NCUA that are summarized below and reprinted in appendix V.
In addition, we received technical comments from IRS that have been
incorporated into this report as appropriate.

In its comment letter, NCUA indicated that the agency's staff have
recommended that the NCUA board consider taking actions consistent with
the recommendations made in our report. NCUA, however, expressed concerns
with certain important aspects of the draft report. In particular, NCUA
stated in its letter that a meaningful comparison between federally
chartered credit unions and other financial institutions should include an
in-depth assessment of their structural and governance differences. NCUA
also noted that the substantive differences among federal credit unions in
charter types and fields of membership significantly impact, among other
things, who credit unions serve and how they operate and provide services.

We agree that there are important structural and governance differences
between credit unions and other depository institutions, which are
highlighted in the report. For example, page one of the draft and current
report notes that credit unions, unlike banks, are (1) not-for-profit
entities that build capital by retaining earnings (they do not issue
capital stock); (2) member-owned cooperatives run by boards elected by the
membership; (3) subject to field of membership requirements that limit
membership to persons sharing certain circumstances, such as a common bond
of occupation or association; and (4) exempt from federal income tax.
Additionally, we agree that differences in charter types and fields of
membership are important factors that should be considered in assessing
who credit unions serve. However, as we note in the report, statistically
reliable data on credit union members by charter type and field of
membership were not available at the time of our review. The lack of this
type of data was the primary basis for the report's recommendation that
NCUA systematically obtain information on the income levels of federal
credit union members. We are encouraged by NCUA's pilot effort to obtain
information on the income levels of federal credit union members and
continue to believe the value of the information collected would be
greatly enhanced if NCUA were to continue its data collection efforts and
address some of the limitations of the pilot.^52 Specifically, NCUA's data
collection efforts could be strengthened by (1) providing benchmark data,
such as general population income statistics or other appropriate
measures, to allow comparisons with the data collected on the income
levels of credit union members; (2) obtaining data on the extent of
services offered by credit unions (e.g., free checking accounts, no charge
ATMs, low-cost wire transfers, etc.) are being used by income category;
(3) expanding the data collection effort to allow the results to be
projectable by charter type; and

(4) conducting the study on a systematic or periodic basis to assess the
extent of progress over time.

NCUA's letter also stated that it was inaccurate and inappropriate to
measure the success of federally chartered credit unions in serving
persons of modest means by reference only to the low- and moderate-income
categories associated with the Community Reinvestment Act. Specifically,
NCUA noted that there was legal and historical evidence that the term
modest means, as used by Congress in the context of the Federal Credit
Union Act, is intended to include a broader range of individuals than
those

^52See NCUA, Member Service Assessment  Pilot Program, A Study of  Federal
Credit Union  Service  (Washington,  D.C.: Nov.  3,  2006),  available  at
www.ncua.gov .

Page 42 GAO-07-29 Credit Union Membership and Executive Compensation

in low- and moderate-income categories. As we noted in the report, neither
the Federal Credit Union Act nor NCUA have established definitions as to
what constitutes modest means. Thus, we used the group consisting of low-
and moderate-income households as a proxy for persons of modest means for
the purposes of our analysis. This allowed us to use the definitions
established for the Community Reinvestment Act as the basis for income
categories used on our analysis. Our analysis not only included
comparisons between credit unions and banks of low- and moderateincome
households but also middle and upper income households for both the 2001
and 2004 SCF. This analysis shows that between 2001 and 2004 credit unions
continued to serve a higher proportion of middle- and upperincome
households and a smaller proportion of low- and moderate-income households
than did banks.

In its letter, NCUA noted that our income category benchmarks were
inconsistent with the specific definitions of the CRA categories the other
federal financial regulators used--specifically the use of national versus
local median income for our benchmarks. Because the most comprehensive and
statistically reliable data available on the income characteristics of
credit union and bank customers at the time of our review--the Federal
Reserve's Survey of Consumer Finances--were nationally representative, we
used national median income measures as the basis for our income
categories whereas the categories used for the Community Reinvestment Act
are based on more local measures.

NCUA's letter also expressed concerns about the reliability of conclusions
reached using the Federal Reserve's Survey of Consumer Finances data.
Specifically, NCUA noted that the SCF was not designed for reliable income
comparisons between credit union members and bank customers. As we noted
in our draft and current report, we agree that the SCF was not
specifically designed to conduct comparative analyses of income levels of
bank and credit union customers; however, SCF provides the best data
currently available to undertake such a comparison. As we reported in
2003, we analyzed the SCF because it is a respected source of publicly
available data on financial institution and consumer demographics that is
nationally representative and because it was the only comprehensive source
of publicly available data that we could identify with information on
financial institutions and consumer demographics. Moreover, our draft and
current report noted limitations in SCF data that preclude drawing
definitive conclusions about the income characteristics of credit union
members. NCUA also provided additional detailed written comments as an
enclosure to its letter, which we have reprinted in appendix V with our
responses.

As we agreed with your office, unless you publicly announce the contents
of this report earlier, we plan no further distribution until 30 days from
the date of this letter. At that time, we will send copies of the report
to the Ranking Member, House Committee on Ways and Means; other interested
congressional committees and subcommittees; the Chairman, NCUA; and the
Commissioner, IRS. We will make copies available to others upon request.
In addition, the report will be available at no charge on the GAO Web site
at http://www.gao.gov.

If you have any questions concerning this report, please contact Yvonne D.
Jones at (202) 512-8678. Contact points for our Office of Congressional
Relations and Public Affairs may be found on the last page of this report.
See appendix VI for a list of other staff who contributed to the report.

Sincerely yours,

Yvonne D. Jones
Director, Financial Markets and Community Investment

     Appendix I
     Objectives, Scope, and Methodology

Our report objectives were to (1) assess the effect of the 1998 Credit
Union Membership Access Act (CUMAA) on federal credit union membership and
charter expansion, (2) review the National Credit Union Administration's
(NCUA) efforts to expand credit union services to low- and moderateincome
individuals, (3) compare rates offered by credit unions to comparably
sized banks as one indicator of how tax-exemption might benefit credit
union members, (4) discuss issues associated with the application of the
federal unrelated business income tax (UBIT) to credit unions, and (5)
assess the transparency of credit union executives and board member
compensation.

  Effects of CUMAA and NCUA Regulations and NCUA Efforts to Serve Low- and
  Moderate-Income Individuals

To study the impact of CUMAA on federal credit union membership and
charter expansion, we reviewed and analyzed the legislative history for
CUMAA and compared its provisions with NCUA interpretive rulings and
policy statements in effect before and after the enactment of CUMAA. In
addition, we interviewed NCUA officials and industry representatives and
met with credit union and banking trade groups including the National
Association of Federal Credit Unions, National Association of State Credit
Union Supervisors (NASCUS), Credit Union National Association, America's
Community Bankers, and Independent Community Bankers to obtain their
viewpoints on how CUMAA and NCUA regulation affected credit union
chartering and field of membership. To obtain information about state
credit union chartering and fields of membership, we held discussions and
reviewed documentation provided by NASCUS. Finally, we obtained electronic
files from NCUA that contained annual call report financial data (Form
5300) of all federally chartered credit unions for yearends 2000 through
2005. The information included the number of credit unions, actual and
"potential" membership (that is, people within a credit union's field of
membership but not members of the credit union), assets, charter
approvals, charter conversions, and charter expansions.

To identify the results of NCUA programs intended to expand credit union
services to low- and moderate-income individuals and underserved areas, we
analyzed NCUA call report data for the low-income-designated credit unions
and credit unions that expanded into underserved areas for yearends 2000
through 2005. The data included information on the number of credit unions
participating in these programs, their asset size, and their membership.

We reviewed NCUA-established procedures for verifying the accuracy of the
Form 5300 database and found that the data are verified on a yearly

Appendix I Objectives, Scope, and Methodology

basis, either during each credit union's examination or through off-site
supervision. In addition, we cross checked the December 2000 to December
2002 data that we recently received with the same data in our 2003 report.
We determined that the data were sufficiently reliable for the purposes of
this report.

Further, we analyzed existing data on the income levels of credit union
customers. Specifically, we analyzed both the 2001 and 2004 releases of
the Board of Governors of the Federal Reserve System's (Federal Reserve)
Survey of Consumer Finances (SCF). The SCF is conducted every 3 years and
is intended to provide detailed information on the balance sheets,
pension, incomes, and demographics of U.S. households and their use of
financial institutions.^1 Because some households use both banks and
credit unions, we performed our analyses based on the assumption that
households can be divided into four user categories--those who use credit
unions only, those who primarily use credit unions, those who use banks
only, and those who primarily use banks.^2 "Primarily use" banks (or
credit unions) means placing more than 50 percent of a household's assets
in banks (or credit unions). As in our prior report, we created four
income categories that are based on those used by financial regulators as
part of Community Reinvestment Act examinations--low, moderate, middle,
and upper--to classify these households (see table 5).^3 As in our 2003
report, we were unable to find a definition of "modest means"; thus, to
assess the extent to which credit unions served people of "modest means,"
we combined households with low- or moderate-incomes into one group as a
proxy for modest means.

^1We use the term, "household," rather than "family," since the subject
group of the SCF more closely resembles the U.S. Census Bureau's
definition of "household" than its definition of "family."

^2Our analysis was based on an approach developed by Jinkook Lee of Ohio
State University. See Jinkook Lee and William A. Kelly Jr., "Who Uses
Credit Unions?" (Prepared for the Filene Research Institute and the Center
for Credit Union Research, 1999, 2001).

^3The Community Reinvestment Act is intended to encourage depository
institutions to help meet the credit needs of the communities in which
they operate, including low- and moderate-income neighborhoods, consistent
with safe and sound banking operations. It was enacted by the Congress in
1977 (12 U.S.C. 2901).

                 Appendix I Objectives, Scope, and Methodology

Table 5: Definition of Income  Categories Used for Community  Reinvestment
Act Examinations

                             Categories Definitions

Low income Income  less than  50 percent of  the Metropolitan  Statistical
Area's (MSA) median income

Moderate income Income at least 50 percent and less than 80 percent of the
MSA''s median income

Middle income Income at least 80 percent and less than 120 percent of  the
MSA's median income

Upper income  Income at  least 120  percent or  more of  the MSA's  median
income

Source: 12 C.F.R. 228.12(n).

Finally we discussed with NCUA officials the design and methodology of its
ongoing pilot project to measure the income levels of federal credit union
members. We also discussed with NASCUS officials their effort to measure
the income levels of state-chartered credit union members.

  Comparison of Interest Rates Offered by Credit Unions With Those at Comparably
  Sized Banks

To compare the rates of credit unions with those at comparably sized
banks, we engaged the services of Datatrac Corporation--a market research,
information technology company specializing in the financial services
industry--to provide data on 15 loan and savings products offered by
credit unions and banks.^4 Datatrac calculated the average rates for each
of these products by five distinct peer groups for asset size, for about
2,000 credit unions and 4,000 banks (see table 6).

^4Datatrac is a privately held company that specializes in financial
industry research. Specifically, Datatrac monitors and analyzes rate
trends on popular deposit and lending products for thousands of financial
institutions nationwide. Institutions voluntarily provide data to Datatrac
on a weekly basis for inclusion in the company's database.

Page 47 GAO-07-29 Credit Union Membership and Executive Compensation
Appendix I Objectives, Scope, and Methodology

Table 6: Peer (Asset) Group  Definitions, Used in Comparisons of  Interest
Rates between Credit Unions and Banks

Group Asset size of institution                                            
I     Total assets of $100 million or less                                 
II    Total assets greater than $100 million, but less than or equal to    
         $250 million                                                         
III   Total assets greater than $250 million, but less than or equal to    
         $500 million                                                         
IV    Total assets greater than $500 million, but less than or equal to $1 
         billion                                                              
V     Total assets greater than $1 billion, but less than or equal to the  
         asset size,                                                          
         rounded up to the nearest billion dollars, of the largest credit     
         union                                                                

Source: GAO.

We established the peer groups based on the institution's size as measured
by total assets for banks and credit unions. Datatrac obtained asset
information for each institution by combining information in its database
with call report data for each institution. Datatrac computed average
rates for institutions overall and for all institutions within analysis
groups. In computing these simple averages, individual institution rates
were not weighted to reflect loan volume or other measures of size.

Datatrac provided us with an electronic file containing information for
2000 through 2005. The information included (1) institution type, (2)
average rate, (3) maximum rate, (4) minimum rate, (5) standard deviations,
(6) product name, (7) quarter and year, and (8) institution counts. We
interviewed Datatrac officials to confirm that they followed industry
accepted protocols to ensure data integrity, including input and
processing controls. We also reviewed Datatrac's methodological
documentation. In addition, we conducted reasonableness checks on the data
we received and identified data gaps in the year-end 2003 information.
Datatrac examined its processing procedures and explained to us that its
cut-off date was incorrectly designated 1 week later than planned. At the
same time, Datatrac also verified that the same problem did not exist in
any other quarters of the years 2000 through 2005. Datatrac provided us an
updated electronic file reflecting the corrections. We determined that the
revised data were sufficiently reliable for the purposes of this report.

                 Appendix I Objectives, Scope, and Methodology

  Issues Related to the Application of UBIT to Credit Unions

To review issues related to the application of UBIT to credit unions, we
reviewed the legislative history of UBIT and the historical basis for the
taxexempt status of credit unions and met with representatives of the
Internal Revenue Service (IRS) to discuss UBIT filing and reporting
requirements. We also discussed with IRS officials their examinations of
unrelated business activity at state-chartered credit unions and
development of policies and procedures in this area. We also obtained
information from IRS on the number and types (group versus individual) of
Return of Organization Exempt from Income Tax (Form 990) and Exempt
Organization Business Income Tax Return (Form 990T) filings by
statechartered credit unions and the amount of unrelated business income
reported and taxes paid by state-chartered credit unions for tax years
from 2000 through 2004.

  Information on Transparency and Compensation of Executive Compensation

To provide information on the transparency and compensation of credit
union executives and board members, including an assessment of the
availability of compensation data to credit union members and a comparison
of executive compensation at credit unions and comparably sized banks, we
interviewed officials at NCUA and IRS to discuss executive compensation
reporting requirements. We obtained and analyzed examiner guidance on
compensation from NCUA and the other federal banking regulators--the
Federal Deposit Insurance Corporation, Federal Reserve, Office of the
Comptroller of the Currency, and Office of Thrift Supervision. We also met
with credit union and banking trade groups including the National
Association of Federal Credit Unions, NASCUS, Credit Union National
Association, America's Community Bankers, and Independent Community
Bankers to identify publicly available data regarding the compensation of
credit union and bank senior executives. We reviewed and analyzed selected
credit union and bank compensation surveys. For more information on the
surveys and our analysis, see appendix IV. We also met with NCUA to
discuss their efforts to collect federal credit union executive
compensation.

Appendix II
Analyses of Survey of Consumer Finances Data, 2001 and 2004

Using the methodology that we employed in our prior report, data from the
2001 and 2004 releases of the Federal Reserve SCF that we analyzed
indicated that credit unions continued to serve a lower proportion of
lowincome households than banks for the years analyzed.^1 As we reported
in 2003, we analyzed the SCF because it is a respected source of publicly
available data on financial institution and consumer demographics that is
nationally representative and because it was the only comprehensive source
of publicly available data with information on financial institutions and
consumer demographics that we could identify. While it is the best
publicly available data that we could identify, there are limitations in
SCF data that preclude drawing definitive conclusions about the income
characteristics of credit union members. In an effort to provide greater
context, in this appendix, we also present the results of additional
analyses of the 2001 and 2004 SCF data that we conducted.

The SCF is conducted every 3 years and is intended to provide detailed
information on the balance sheet, pension, income, and other demographics
of U.S. households and their use of financial institutions.^2 The survey
is based on approximately 4,500 interviews and represents a sample of more
than 100 million households. For each of the 2001 and 2004 SCF releases,
we combined the SCF data into two main groups-- households that only and
primarily used credit unions (credit union customers) and households that
only and primarily used banks (bank customers).^3 Our analyses of 2001 and
2004 SCF data indicated that, among households that used a financial
institution, those households that we

^1See GAO, Credit Unions: Financial Condition Has Improved, but
Opportunities Exist to Enhance Oversight and Share Insurance Management,
[145]GAO-04-91 (Washington, D.C.: Oct. 27, 2003).

^2We use the term "household" rather than "family" since the reporting
unit of the SCF more closely resembles the U.S. Census Bureau's definition
of "household" than its definition of "family." The Census Bureau's
definition of "family" excludes the possibility of one-person household
units, but its definition of "household" allows for them. See Brian K.
Bucks, Arthur B. Kennickell, and Kevin B. Moore, "Recent Changes in U.S.
Family Finances: Evidence from the 2001 and 2004 Survey of Consumer
Finances," Federal Reserve Bulletin, Mar. 22, 2006, A3.

^3Those who "primarily" used credit unions placed more than 50 percent of
their assets in credit unions, and those who "primarily" used banks placed
more than 50 percent of their assets in banks. The term "use" refers to a
household's placement of assets in a checking, savings, or money market
account. Our methodology for determining these classifications was based
on work that Jinkook Lee, a professor and researcher at Ohio State
University, performed. See Jinkook Lee and William A. Kelly Jr., "Who Uses
Credit Unions?" (Prepared for the Filene Research Institute and the Center
for Credit Union Research, 1999, 2001).

Appendix II Analyses of Survey of Consumer Finances Data, 2001 and 2004

identified as being bank customers outnumbered those that we identified as
being credit union customers by a large margin (see table 7).^4 Because
such a high percentage of the U.S. population represented by the SCF only
used banks, the data obtained from the SCF are particularly useful for
describing characteristics of bank users but much less precise for
describing smaller population groups, such as those that only used credit
unions. It should be noted that SCF was not specifically designed to
conduct comparative analyses of income levels of bank and credit union
customers, and the pool of bank customers is not necessarily comparable to
the pool of credit union customers.

Table 7: Percentages  of Households  Classified as Using  Banks or  Credit
Unions, 2001 and 2004

                               Percentage of households (among all households
                                               using a financial institution)
Financial institution usage         2001 SCF data            2004 SCF data 
Only used credit unions                                8                 8 
Primarily used credit unions                          13                14 
Primarily used banks                                  17                15 
Only used banks                                       62                63 

Sources: GAO and Federal Reserve.

We found that credit union customers had a higher median income than bank
customers in both the 2001 and 2004 SCF releases. In the 2001 SCF, the
median income of all households was $39,000; bank customers had a median
income of $40,000 and credit union customers had a median income of
$44,000. In the 2004 SCF, the median income of all households was $42,000;
bank customers had a median income of $43,000 and credit union customers
had a median income of $50,000.

We computed the proportions of credit union customers and bank customers
in each of four income categories--low, moderate, middle, and upper. As in
our 2003 report, we based our income groups on income categories used by
financial regulators for federal Community

^4In our analyses of SCF data, we specify banks to include both commercial
banks and savings and loan institutions. Percentages reflect the
households using financial institutions as a percentage of all financial
institution users and exclude those households that did not use a
financial institution (sometimes referred to as "unbanked").

Page 51 GAO-07-29 Credit Union Membership and Executive Compensation
Appendix II Analyses of Survey of Consumer Finances Data, 2001 and 2004

Reinvestment Act examinations in an effort to provide a consistent
framework given that "modest means" is not clearly defined.^5 For our
primary analysis of 2001 and 2004 SCF data, we used 2000 and 2003 median
household income as reported by the U.S. Census Bureau; for our additional
analyses of 2001 and 2004 SCF data, we used 2000 and 2003 median family
income as reported by the U.S. Census Bureau (see tables 8 and 9). It
should be noted that the categories that we use here, which we introduced
in our 2003 report, are based on a national median income measure whereas
the categories used for Community Reinvestment Act are based on more local
measures.

Table 8:  Median  Income  Benchmarks  Used  for  Primary  (Household)  and
Additional (Family) Analyses of 2001 and 2004 SCF Data

                                      2001 median SCF             2004 median 
                                     household income 2004 median         SCF 
              2001 median SCF                          SCF family   household 
Analysis     family income                              income      income 
Primary                N/A                 $42,151         N/A     $43,318 
Additional         $50,890                     N/A     $52,680         N/A 

                      Sources: GAO and U.S. Census Bureau.

^5See appendix I for more information.

Appendix II Analyses of Survey of Consumer Finances Data, 2001 and 2004

Table 9: Income Categories for Primary (Household) and Additional (Family)
Analyses of 2001 and 2004 SCF Data

Income category           Primary analysis       Additional analysis       
Low                       Income less than 50    Income less than 50       
                             percent of the median  percent of the median     
                             household income       family income (2001: less 
                             (2001: less than       than $25,445; 2004: less  
                             $21,076; 2004: less    than $26,340)             
                             than $21,659)                                    
Moderate                  Income at least 50                               
                             percent and less than  Income at least 50        
                             80 percent of the      percent and less than 80  
                             median household       percent of the median     
                             income (2001: at least family income (2001: at   
                             $21,076 but less than  least $25,445 but less    
                             $33,721; 2004: at      than $40,712; 2004; at    
                             least                  least $26,340             
Middle     $21,659 but less than $34,654)  but less than $42,114) Income
              Income at least 80 percent and  at least 80 percent and less
              less than 120 percent of the    than 120 percent of the median
              median household income (2001:  family income (2001: at least
              at least $33,721 but less than  $40,712 but less than $61,068;
              $50,581; 2004: at least         2004: at least $42,144
               $34,654 but less than $51,982) but less than $63,216)
Upper      Income at least 120 percent or  Income at least 120 percent or
              more of the median household    more of the median family
              income (2001: at least $50,581; income (2001: at least $61,068;
              2004: at least $51,982)         2004: at least $63,216)

Source: GAO.

As noted earlier in the report, our (primary) analysis of 2004 SCF data
suggested that credit unions served a lower proportion of households of
modest means (low- and moderate-income households, collectively) than
banks, a result consistent with the finding in our 2003 report analyzing
the 2001 SCF data (see tables 10 and 11).

Table 10:  Percentages in  Each Income  Category for  Primary  (Household)
Analysis by Customer Type, 2001 SCF Data

                                          All SCF  Credit union          Bank 
Income category                    respondents    customers      customers 
Low                                       27.0            16.4        25.7 
Moderate                                  16.6            19.3        16.1 
Middle                                    17.6            21.7        17.5 
Upper                                     38.9            42.6        40.7 

Sources: GAO and Federal Reserve.

Note: Income benchmark is the median household income for 2000.

Appendix II Analyses of Survey of Consumer Finances Data, 2001 and 2004

Table 11:  Percentages in  Each Income  Category for  Primary  (Household)
Analysis by Customer Type, 2004 SCF Data

                                      All SCF    Credit union            Bank 
Income category           respondents          customers         customers 
Low                                   24.9               14.5         24.2 
Moderate                              16.6               16.6         16.4 
Middle                                17.8               20.2         18.1 
Upper                                 40.7               48.8         41.3 

Sources: GAO and Federal Reserve.

Note: Income benchmark is the median household income for 2003.

In an effort to determine how sensitive these results were to our income
categorization, we also used the median family income for 2000 and 2003 to
analyze the 2001 and 2004 SCF data. As shown in tables 12 and 13, the
results from our additional analyses were similar to those of our primary
analysis. While the median family income was higher than the median
household income in each year, the results continue to suggest that a
greater proportion of bank than credit union customers were of modest
means. This difference between banks and credit unions was statistically
significantly different from zero in the 2004 SCF; there was also a
statistically significant decline in the proportion of credit union
customers of modest means between the 2001 and 2004 SCF data. Thus, while
the results of our analyses should not be considered definitive, they do
suggest that any impact from the recent efforts by NCUA to increase credit
union membership among the underserved and low- and moderate-income
households have not yet appeared in the data.

Table 12:  Percentages in  Each Income  Category for  Additional  (Family)
Analysis by Customer Type, 2001 SCF Data

                                         All SCF   Credit union          Bank 
Income category                   respondents    customers       customers 
Low                                      33.8             23.2        32.3 
Moderate                                 18.6             23.6        18.2 
Middle                                   16.7             21.6        16.5 
Upper                                    30.9             31.7        33.1 

Sources: GAO and Federal Reserve.

Note: Income benchmark is the median family income for 2000.

Appendix II Analyses of Survey of Consumer Finances Data, 2001 and 2004

Table 13:  Percentages in  Each Income  Category for  Additional  (Family)
Analysis by Customer Type, 2004 SCF Data

                                         All SCF   Credit union          Bank 
Income category                   respondents    customers       customers 
Low                                      31.6             20.2        31.0 
Moderate                                 18.6             21.1        18.3 
Middle                                   18.1             20.6        18.6 
Upper                                    31.7             38.2        32.2 

Sources: GAO and Federal Reserve.

Note: Income benchmark is the median family income for 2003.

We also considered the median income of bank and credit union customers
within each of our income categories for both the primary and additional
analyses to assess whether there were any notable differences between
credit union and bank customers (see tables 14 through 17). We found that
the income characteristics of the customers tended to be similar; however,
the median income in the upper-income category tended to be higher for
bank customers.

Table  14:  Median  Income  within   Each  Income  Category  for   Primary
(Household) Analysis by Customer Type, 2001 SCF Data

                                                Median income within category
Income category            Credit union customers           Bank customers 
Low                                            $13,000             $13,000 
Moderate                                        28,000              27,000 
Middle                                          40,000              41,000 
Upper                                           74,000              83,000 

Sources: GAO and Federal Reserve.

Note: Income benchmark is the median household income for 2000.

Appendix II Analyses of Survey of Consumer Finances Data, 2001 and 2004

Table  15:  Median  Income  within   Each  Income  Category  for   Primary
(Household) Analysis by Customer Type, 2004 SCF Data

                                                Median income within category
Income category            Credit union customers           Bank customers 
Low                                            $13,000             $13,000 
Moderate                                        28,000              28,000 
Middle                                          42,000              43,000 
Upper                                           86,000              86,000 

Sources: GAO and Federal Reserve.

Note: Income benchmark is the median household income for 2003.

Table 16:  Median  Income  within  Each  Income  Category  for  Additional
(Family) Analysis by Customer Type, 2001 SCF Data

                                                Median income within category
Income category            Credit union customers           Bank customers 
Low                                            $16,000             $15,000 
Moderate                                        33,000              33,000 
Middle                                          51,000              50,000 
Upper                                           83,000              98,000 

Sources: GAO and Federal Reserve.

Note: Income benchmark is the median family income for 2000.

Table 17:  Median  Income  within  Each  Income  Category  for  Additional
(Family) Analysis by Customer Type, 2004 SCF Data

                                                Median income within category
Income category            Credit union customers           Bank customers 
Low                                            $17,000             $15,000 
Moderate                                        34,000              34,000 
Middle                                          52,000              51,000 
Upper                                           97,000             101,000 

Sources: GAO and Federal Reserve.

Note: Income benchmark is the median family income for 2003.

Appendix III

Comparison of Interest Rates at Credit Unions and Banks

Data that we obtained indicate that credit unions offer more favorable
rates on average than similarly sized banks for a number of savings
products and consumer loans. However, similarly sized credit unions and
banks appeared to offer virtually the same rates on mortgage loans, such
as 15- and 30-year fixed-rate mortgages. We engaged the services of
Datatrac Corporation--a market research and information technology
company, specializing in the financial services industry--to gather and
analyze data on loan and savings rates for 15 loan and savings products (5
consumer loan, 3 mortgage loan, and 7 savings products) that were offered
from 2000 through 2005 at about 2,000 credit unions and 4,000 banks.^1
Financial institutions voluntarily provide data to Datatrac on a weekly
basis for inclusion in the company's database. Therefore, information
presented is not necessarily statistically representative of the entire
banking and credit union industry.

Datatrac calculated the average rates for each of these products by five
distinct asset size peer groups:

     o total assets of $100 million or less;
     o total assets greater than $100 million, but less than or equal to $250
       million;
     o total assets greater than $250 million, but less than or equal to $500
       million;
     o total assets greater than $500 million, but less than or equal to $1
       billion; and
     o total assets greater than $1 billion, but less than or equal to the
       asset size, rounded up to the nearest billion dollars, of the largest
       credit union.

Datatrac computed average rates for institutions overall and for all
institutions within analysis groups. In computing these simple averages,
individual institution rates were not weighted to reflect loan volume or
other measures of size. While Datatrac Corporation's database contained

^1Datatrac is a privately held company that specializes in financial
industry research. Specifically, Datatrac monitors and analyzes rate
trends on popular deposit and lending products for thousands of financial
institutions nationwide.

Page 57 GAO-07-29 Credit Union Membership and Executive Compensation

      Appendix III Comparison of Interest Rates at Credit Unions and Banks

data provided by about 2,000 credit unions and 4,000 banks, data were not
always obtained from all the credit unions and banks for every product
and/or time period in each of the five asset groupings. We identify all
instances in which the information presented was based on rate data
provided by less than 10 institutions. Additionally, because averages
based on a small number of institutions may be unreliable, we did not
report instances when rate data was provided by less than 5 institutions.

Figures 7 through 23 provide a detailed comparison of rates on savings and
loan products offered by credit unions to those at similarly sized banks
for the 6-year period spanning from 2000 to 2005.

  Figure 7: Comparison of Interest Rates for Savings Products at December 31,
                              2000, by Asset Size

                           Sources: GAO and Datatrac.

Appendix III Comparison of Interest Rates at Credit Unions and Banks

  Figure 8: Comparison of Interest Rates for Savings Products at December 31,
                              2001, by Asset Size

                           Sources: GAO and Datatrac.

Appendix III Comparison of Interest Rates at Credit Unions and Banks

  Figure 9: Comparison of Interest Rates for Savings Products at December 31,
                              2002, by Asset Size

                           Sources: GAO and Datatrac.

Appendix III Comparison of Interest Rates at Credit Unions and Banks

                           Sources: GAO and Datatrac.

Appendix III Comparison of Interest Rates at Credit Unions and Banks

                           Sources: GAO and Datatrac.

Appendix III Comparison of Interest Rates at Credit Unions and Banks

                           Sources: GAO and Datatrac.

      Appendix III Comparison of Interest Rates at Credit Unions and Banks

Sources: GAO and Datatrac.

Note: Data is not reported for  credit card (classic) and banks with  $100
million or less due  to an insufficient  number of reporting  institutions
(less than 5).

^aData is based on responses of less than 10 institutions.

Appendix III Comparison of Interest Rates at Credit Unions and Banks

                           Sources: GAO and Datatrac.

Appendix III Comparison of Interest Rates at Credit Unions and Banks

                           Sources: GAO and Datatrac.

Appendix III Comparison of Interest Rates at Credit Unions and Banks

                           Sources: GAO and Datatrac.

Appendix III Comparison of Interest Rates at Credit Unions and Banks

                           Sources: GAO and Datatrac.

Appendix III Comparison of Interest Rates at Credit Unions and Banks

Institution assets range

                           Sources: GAO and Datatrac.

Appendix III Comparison of Interest Rates at Credit Unions and Banks

Sources: GAO and Datatrac.

^aData is based on responses of less than 10 institutions.

Appendix IV
Selected Salary Surveys for Credit Union and Bank Executives

  Credit Union Salary and Other Cash Compensation Data

Credit union and bank survey information we obtained provides an
indication of executive base salaries for the respective industries. The
credit union and bank salary survey data we identified had a key
limitation--the information was not directly comparable because of
differences in the underlying sampling strategies and data gathering
methodologies. Also, while both surveys report the types of cash
compensation received for their industry executives (i.e., salary and
bonuses), we were not able to identify and compare other forms of benefits
that an executive might typically receive in a compensation package.

There were a number of other limitations in the data that we identified.
In some instances, the information collected for each of the surveys
involved a sample of members belonging to their respective trade group
associations. The data collection periods for each of the surveys were
different. For instance, the credit union survey collected salary
information between January and May 2005, while the bank survey collected
information during 2004. The bank survey also provides general information
on other benefits such as savings incentive plans, pension plans, and paid
time off benefits for which we do not have comparable information in the
credit union survey. Also, cash compensation reported for the credit union
survey includes base salary, incentives, and bonuses, while the bank
survey reported base salary, bonus, and profit sharing compensation.
Finally, the cash compensation information presented for these surveys are
grouped in different asset size ranges. The credit union survey presents
information based on 13 asset size categories, while the bank survey
presents information based on 7 asset size categories.

Credit Union Salary and Other Cash Compensation Data

According to the Credit Union National Association's 2005 to 2006 Complete
Credit Union Staff Salary Survey, the average base salary of credit union
presidents, chief executive officers (CEO), and managers for those credit
unions responding increased 4.8 percent from the previous year's survey.
In addition to base salary, more than half (55 percent) of credit union
presidents, CEOs, and managers also received other forms of cash
compensation such as incentives or bonuses. For CEOs, incentives averaged
$9,634, while bonuses averaged $4,993. The survey also noted that bonuses
continue to be more common than incentives (45 percent compared with 5
percent receiving these payments, respectively in 2004). As shown in
figure 25, the average credit union base salary for the CEO position was
about $78,000 while the average base salaries for the chief financial
officers (CFO) and chief operations officers (COO) was approximately
$73,000 and $64,000, respectively. However, national

    Appendix IV Selected Salary Surveys for Credit Union and Bank Executives

averages should be viewed with care since executive salaries also vary  by
region and the size of the credit union.

Figure 25: Credit Union Executive Average Base Salaries for 2005

President/CEO/Manager salariesChief financial officer salariesChief
operating officer salaries

Dollars in thousandsDollars in thousandsDollars in thousands350 350

Source: GAO analysis of CUNA survey data.

Similarly, according to the survey for those credit unions responding,
credit union executives, including CFOs and COOs, experienced about a 2
percent increase in average salary over the previous year. Of those that
responded to the survey, approximately 27 percent of CFOs received
incentives, which averaged $5,963, and 38 percent received a bonus which
averaged $4,650. Additionally, approximately 21 percent of COOs received
incentives which averaged $5,678, while 47 percent received a bonus that
averaged $3,578.

The number of responses for the survey questions on the three credit union
executive positions also varied from question to question and across the
different asset categories. For instance, a total of 773 credit unions

    Appendix IV Selected Salary Surveys for Credit Union and Bank Executives

  Bank Salary and Other Cash Compensation Data

responded to the president/CEO question, but the responses by asset
category ranged from a low of 16 responses by credit unions with assets of
$1 to $2 million to a high of 113 responses by credit unions with assets
of $100 to $200 million.^1 A total of 330 credit unions responded to the
CFO question, while the responses by asset category ranged from a low of 2
responses by credit unions with assets of $5 to $10 million to a high of
74 responses by credit unions with assets of $100 to $200 million.
Finally, 268 credit unions responded to the COO question, while the
responses by asset category ranged from a low of 2 responses by credit
unions with assets of $5 to $10 million to a high of 65 responses by
credit unions with assets of $100 to $200 million.

Bank Salary and Other Cash Compensation Data

According to America's Community Bankers 2005 Compensation Survey, the
national average base salary for those banks responding to the survey for
CEOs was up 13.2 percent from the 2004 reported average. The average
bonus/profit sharing payment for CEOs was $73,129. Similarly, the national
average base salary for those banks responding for CFOs was up 10.8
percent from 2004, while the average bonus/profit sharing compensation was
$28,700. The base salary for those banks responding for COOs was up
8.8 percent from 2004, while the average bonus/profit sharing compensation
was $32,697. As shown in figure 26, the average bank base salary for the
CEO position was about $213,000 while the average base salaries for the
CFO and COO was approximately $121,000 and $141,000 respectively. As
mentioned previously, national averages should be viewed with care since
executive salaries also vary by region and by asset size.

^1According to the Credit Union National Association's salary survey, as
part of the methodology, nearly all affiliated credit unions with $100
million or more in assets were sent a survey. Stratified random samples of
credit unions with $1 million to $100 million in assets were also sent the
survey. Thus, larger credit unions were given a greater chance of being
selected for the survey to ensure a high degree of accuracy for these
credit unions and weighted to adjust for the overrepresentation of the
larger credit unions. Weighting is a standard survey analysis procedure
designed to adjust estimates to account for different rates of selection
within sample strata, which ensures that results are not biased by a
specific group of credit unions.

    Appendix IV Selected Salary Surveys for Credit Union and Bank Executives

    President/CEO/Manager salaries Chief financial officer salaries Chief
    operating officer salaries

Dollars in thousands Dollars in thousands Dollars in thousands 450 450 450

400

                                    400 400

350

                                    350 350

300

                                    300 300

250

                                    250 250

200

200

200

150

150

150

100

100

100

50

50

50

0

0

0

[50 or  less] ^51-100  301-500201-300  101-200 [501-1,000]^1,000+  [50  or
less] ^51-100  301-500201-300  101-200  [501-1,000]^1,000+  [50  or  less]
^51-100 301-500201-300 101-200 [501-1,000]^1,000+

Bank asset  size  (dollars  in  millions)  Bank  asset  size  (dollars  in
millions) Bank asset size (dollars in millions) Average

Source: GAO analysis of ACB survey data.

The bank executive survey responses also varied by the total number of
respondents and by the different asset categories.^2 For instance, a total
of 358 banks responded to the president/CEO question, while the responses
by asset category ranged from a low of 16 banks with assets up to $50
million to a high of 74 banks with assets of $101 to $200 million. A total
of 256 banks responded to the CFO question, while the number of responses
by asset category ranged from a low of 2 banks with assets up to $50
million to a high of 49 responses by banks with assets of $501 million to
$1 billion. Finally, 187 banks responded to the COO question, while the
response rates by asset category ranged from a low of 8 banks with assets

^2The America's Community Bankers survey cautions against comparing peer
group data among the various asset sizes due to the differences in bank
types. That is, banks with assets up to $50 million disproportionately are
mutual institutions, while banks with assets of more than $1 billion
disproportionately are stock banks.

      Page 79 GAO-07-29 Credit Union Membership and Executive Compensation

Appendix IV Selected Salary Surveys for Credit Union and Bank Executives

up to $50 million to a high of  38 banks in both the $101 to $200  million
and $501 million  to $1  billion categories. Due  to the  small number  of
responses in some  instances, the results  of this data  should be  viewed
with caution.

Appendix V

Comments from the National Credit Union Administration

See comment 1.

                               November 14, 2006

Yvonne D. Jones Director, Financial Markets

and Community Investments Government Accountability Office 441 G St., NW
Washington, DC 20548

Dear Ms. Jones:

Thank you for the opportunity to review and comment on the draft GAO
Report (Report) entitled "Greater Transparency Needed on Who Credit Unions
Serve and on Senior Executive Compensation Arrangements." On behalf of the
National Credit Union Administration (NCUA), I would like to express our
appreciation for the professionalism of your staff and our gratitude for
the dialogue that occurred through the course of GAO's study. NCUA
believes that dialogue was helpful in developing a better mutual
understanding of the complexity of the issues addressed in the Report and
the conflicts that arise when considering the mission and purpose of
federal credit unions (FCU) in the context of today's financial
marketplace.

It is unfortunate GAO did not have available at the time of drafting the
Report the results of NCUA's Member Service Assessment Pilot Program
(MSAP) (Enclosure 1), since MSAP includes significant new data on FCUs.
Importantly, MSAP provides meaningful and accurate information on FCU
membership profiles, as well as an assessment of the data collected. This
assessment is critical for an objective analysis of the data. It also
demonstrates any conclusions reached must consider FCU structure and
operations, and the significant differences between other financial
institutions and FCU charter types.

As outlined in greater detail in the enclosed response to the Report
(Enclosure 2), NCUA does have continued concerns with certain important
aspects of the Report. NCUA believes that a meaningful comparison between
FCUs and other financial institutions must include an in-depth assessment
of their structural and governance differences. Furthermore, comparisons
among FCUs must consider charter types and field of membership
differences. These substantive differences significantly impact who credit
unions serve, how they operate and provide services, how they develop and
maintain their net worth and working capital, and how they affect the
continued viability of the FCU system. Such a framework is missing in the
Report, thus limiting its reliability.

NCUA also believes it is inaccurate and inappropriate to measure the
success of FCUs in serving persons of modest means by reference only to
the low-and

Appendix V Comments from the National Credit Union Administration

moderate-income categories associated with the Community Reinvestment Act
(CRA), as these categories only extend to families at or below 80 percent
of the median income. There is ample legal and historical evidence that
the term modest means, as used by Congress in the context of the FCU Act,
is intended to include both below average wage earners and a broader class
of working individuals generally. Further, GAO elected to use income
category benchmarks that are inconsistent with the specific definitions of
the CRA categories used by the other federal financial regulators. Using
broad income categories and equating modest means to low- and
moderate-income individuals precludes a valid assessment of the economic
demographics of FCU membership.

Additionally, NCUA has serious concerns about the reliability of
conclusions reached using the Federal Reserve's Survey of Consumer Finance
(SCF) data. The SCF was not designed for reliable income comparisons
between credit union members and bank customers. Other concerns, addressed
in Enclosure 2, include the importance of credit union membership limits,
the effects of recent trends in community chartering, and proper
recognition of NCUA's efforts to target services to lower income
individuals.

Regarding the recommendations made in the Report, NCUA staff recommended
in MSAP that the NCUA Board consider whether it is appropriate to gather
additional membership data to further enhance NCUA's efforts in expanding
credit union service to low- and moderate-income individuals. NCUA staff
also recommended that the NCUA Board consider evaluating alternative
approaches to collecting and aggregating executive compensation on an FCU
system basis.

Notwithstanding the continued concerns listed above and described in
greater detail in Enclosure 2, I again want to emphasize our great
appreciation for the efforts of your staff and their willingness to
consider our concerns and engage in open and meaningful dialogue.

                                   Sincerely,

                      J. Leonard Skiles Executive Director

 Enclosures: 
 
1. Report to the NCUA Board on the Member Service Assessment Pilot
                     Program (MSAP), dated November 3, 2006

2. NCUA's Detailed Response to GAO's Draft Report GAO-07-29

                                       2

See comment 2.

Appendix V Comments from the National Credit Union Administration

NCUA's Detailed Response to GAO Draft Report GAO-07-29 "Greater
Transparency Needed on Who Credit Unions Serve and on Senior Executive
Compensation Arrangements"

The following discussion addresses our primary concerns with the GAO Draft
Report, GAO-07-29 (Report). These concerns include: (1) inaccurate use of
low-and moderate-income as a proxy for modest means; (2) inappropriate use
of income categories ostensibly based on CRA categories; (3) improper
reliance on the Federal Reserve's Survey of Consumer Finance; (4)
insufficient discussion of the structure and framework of FCUs; (5)
insufficient discussion of NCUA's efforts to enhance service to low- and
moderate-income individuals; and (6) incomplete data on executive
compensation.

1. GAO's Definition of "modest means"

It is inaccurate for GAO to define the term modest means as only including
low- and moderate-income individuals. To use a proxy definition for modest
means, although convenient for drafting the Report, contradicts clear
congressional intent and disregards important statutory mandates on whom
FCUs can serve. NCUA strongly believes that using the terms modest means
and low- and moderateincome individuals interchangeably creates confusion
and a perception inconsistent with statutory intent and regulatory
policies put in place to achieve that intent.

While the Report recognizes in footnote 30 on page 26 that there is no
commonly accepted definition of modest means, the following statement on
page 6 equates low- and moderate-income to modest means:

[T]he Federal Reserve's 2004 Survey of Consumer Finance (SCF)

-indicates that credit unions continued to lag behind banks in the
percentage of their customers or members that were of low- and
moderate-income households. Our analysis of the 2004 SCF indicated that 32
percent of households that only and primarily used credit unions were of
modest means (emphasis added). . . . .

The history of the Credit Union Membership Access Act of 1998 (CUMAA)^1
demonstrates congressional intent when the term "modest means" was used.
This term was first introduced in proposed amendments to the FCU Act in
1998 describing the mission of credit unions. Although these amendments
were not adopted in the final version of CUMAA, the House Report
accompanying the proposed bill noted: "Section 204 reaffirms the
continuing and affirmative obligation of insured credit unions to meet the
financial services needs of persons

1

Pub. L. No. 105-219, 112 Stat. 913 (August 7, 1998).

-1 -Enclosure 2

Appendix V Comments from the National Credit Union Administration

of modest means, including those with low- and moderate-incomes,
consistent with safe and sound operation."^2

The Senate Report followed a similar usage in referring to section 204 of
the bill. Specifically, the Senate Report also discussed the calling of
credit unions to serve the entire range of membership and to provide
"affordable credit union services to all individuals of modest means,
including those with low- and moderate-incomes, within the field of
membership of such credit union."^3

These congressional views reflect the clear understanding that the term
modest means indicates a meaning broader than individuals with low- and
moderateincome, and those that meet the definition of modest means must
also be within the field of membership (FOM). In this respect, the term,
though not specifically defined, conforms explicitly with its earlier
counterpart, "small means," as a shorthand reference to members of the
broad working class.

CUMAA also served notice that outreach programs to reach low- and
moderateincome individuals, and the support for credit unions designated
to serve lowincome memberships, should still continue. Additional
authorities granted to lowincome designated credit unions, and the ability
for multiple common bond FCUs to adopt underserved areas are also
consistent with a more expansive definition for modest means.

2. Use of CRA-type definitions for income levels

The Report, in footnote 27 on page 24, provides an explanation for the use
of the Federal Reserve's Survey of Consumer Finance (SCF) and income
categories, and states:

We [GAO] based our groups on income categories used by financial
regulators for federal Community Reinvestment Act examinations intended to
encourage depository institutions to help meet credit needs in all areas
of the communities that they serve:

(1) a low-income household had an income of less than 50 percent of the
national median household income; (2) a moderate-income had an income of
at least 50 percent of but less than 80 percent of the national median
household income; (3) a middle-income household had an income of at least
80 percent of but less than 120 percent of the national median household
income; and (4) an upper-income household had an income of at least 120
percent of the national median household income.

2

H.R.
           REP. NO. 105-472, at 22 (1998)(emphasis added).

S.
           REP. NO. 105-193, at 11 (1998)(emphasis added).

3

                                -2 -Enclosure 2

Appendix V Comments from the National Credit Union Administration

                                 See comment 3.

                                 See comment 4.

This footnote does not accurately reflect the income categories
established by the federal financial regulators for CRA examinations and
contradicts Table 5 in Appendix I of the Report. The income categories
identified in the Code of Federal Regulations for CRA purposes are based
on median family income as a percent of metropolitan statistical area
(local area) median family income. ^4 The income categories utilized in
the Report use median household income as a percent of national (not local
area) median household income. Although the Report utilizes median family
income in its additional analysis, this not only contradicts the SCF's
methodology, but also does not correct for CRA inconsistency.
Consequently, the statement that the income levels used are similar to
those used in other governmental programs is misleading and implies the
analyses are based on CRA income categories when, in fact, the income
categories are GAO-defined.

Additionally, footnote 27 illustrates CRA is intended to "encourage
depository institutions to help meet credit needs in all areas of the
communities that they serve. . ." Given 80 percent of FCUs are
occupational or associational based, the CRA-type categories have limited,
if any, applicability for the assessment of FCUs.

3. Basing Assessment on the Federal Reserve's Survey of Consumer Finance

NCUA recognizes the lack of reliable data to serve as a basis for valid
conclusions regarding income distribution of FCU members at the time of
the drafting of the Report. NCUA also accepts that the SCF was the only
source of data available that provided income figures, albeit of limited
application, for FCU members.^5 As correctly pointed out by GAO, the SCF
was not designed to analyze credit union member income distribution or
make comparisons between credit union members and bank customers. For
example, the SCF does not provide proportional representation of credit
union members and bank customers necessary to develop valid conclusions
pertaining to income distribution. Notwithstanding these known
deficiencies, the SCF is the primary source for the conclusions reached in
the Report, which has the potential for misleading assessments about whom
credit unions serve compared to banks.

Additionally, throughout this study NCUA discussed with GAO various means
of presenting the SCF data. In NCUA's view, the use of a single chart, on
page 27 in the body of the Report, using broad income categories, limits
the reader's ability to draw objective conclusions. Although Appendix I
includes additional

^4 See 12 C.F.R. SS 228.12(b) and (m)(Federal Reserve), 345.12(b) and
(m)(FDIC), 25.12(b) and (m)(OCC), and 563e.12(b) and (m)(OTS).

5

The number of households primarily using credit unions included in the
2004 SCF is only 14 percent of those surveyed. The number of FCU member
households included in this small number is unknown. See page 63 of the
Report.

-3 -Enclosure 2

Appendix V Comments from the National Credit Union Administration

comparisons, they are also insufficient for providing a comprehensive view
of member incomes.

The use of additional tables, in the body of the report, depicting the
same data in various ways would have allowed a more complete view of
member incomes. For example, the Federal Reserve uses income percentiles
in its assessment of the SCF, which provides a more objective presentation
of income distribution than the broad income categories used in the
Report. Table 1 presents the data used in the Report based on these income
percentiles.

                                   Table 1^6

           Percentage of Members/Customers within Income Percentile           
              Primary and Only Users                           2003           
Percentile of Income   Annual Income Ranges   Credit Unions      Banks     
Income pct < 20            $0 to $18,900          11.5%          19.2%     
Income pct: 20-39.9     $18,901 to $33,900        18.8%          20.0%     
Income pct: 40-59.9     $33,901 to $53,600        23.2%          20.7%     
Income pct: 60-79.9     $53,601 to $89,300        24.8%          20.2%     
Income pct: 80-89.9     $89,301 to $129,400       13.2%          9.4%      
Income pct: 90-100          > $129,400             8.5%          10.6%     

Further, including both average and median incomes for comparative
purposes, rather than using only median as reflected in the Report,
provides for a more complete view of member incomes. According the SCF
results and as demonstrated in Table 2, while credit union members have
the highest median income, bank customers have the highest average income.

                                   Table 2^7

                  Median and Average Income Comparison                        
                                      2003                                    
                       Total SCF         Credit Union           Banks         
       Median           $42,000            $50,000              $43,000       
      Average           $68,778            $62,572              $74,211       

There are also technical inconsistencies in the Report's methodology. For
example, as stated on page 28: "To determine how sensitive our [GAO's]
results were to our income categorization, we used median family income in
addition to median household income to analyze the 2001 and 2004 SCF data.
We found similar results using both median family and household income."
However, this comparison does not accomplish its stated objective. The use
of the median

6

Table compiled by NCUA to illustrate other alternatives for SCF data
analysis.

7

Table compiled by NCUA to illustrate other alternatives for SCF data
analysis using the median income included in the Report.

-4 -Enclosure 2

See comment 5.

Appendix V Comments from the National Credit Union Administration

family income for comparison is inconsistent with the SCF methodology,
which utilized median household income. Therefore, this comparison does
not add validity to the results of the study since it only changes the
comparative benchmark.

4. Providing a limited framework for credit union membership assessment

Although the Report correctly recognizes that credit unions retain their
distinction in terms of structure and governance, it does not provide a
framework that would allow for an appropriate interpretation of the
assessments presented. For example, factual background information about
credit unions and their important differences from banks, which is vital
for an understanding of this issue, is not adequately addressed. To fully
understand and assess any data that attempts to compare credit union
members with depositors in other types of financial institutions, the
Report should include discussion of the following:

A. Statutory limitations on FCU membership

MSAP data confirms the importance of the statutory mandate concerning
common bond when assessing membership profiles. It also confirms that
comparisons with other financial institutions, as well as among different
chartertypes of FCUs, are difficult. FCUs are chartered as cooperatives to
serve individuals only within their FOM. They are, therefore, limited in
whom they can serve and are restricted to the income composition of the
individuals within their allowed FOM. It is misleading to draw definitive
conclusions about the success of FCUs in serving individuals and groups
outside their traditional membership base without fully focusing on their
authorized FOMs. This is particularly important in view of the fact that,
as of December 2005, approximately 80 percent of all FCUs had single-or
multiple-common bond charter types based on occupation or association. The
implication of this FOM concentration, based primarily on working
individuals, is far reaching within the context of assessing the
membership profile of FCUs.

Understanding statutory limitations on who can join FCUs is critical in
conducting an objective assessment of the FCU system membership profile,
any policy consideration on who benefits from credit union services, and
the impact of FCUs on the financial sector. The statutory limitations also
emphasize the differences between FCUs and banks and draw into question
the reasonableness of any general comparison between income distribution
of FCU members and bank customers. To conduct a reasonable comparative
assessment of whom FCUs and banks serve, both types of institutions would
need to have a similar structure and other characteristics. Although
community-chartered FCUs and community banks may share some similarities
relative to location, structurally, communitychartered FCUs remain
cooperatives with the limitations of building capital/net worth,
geographic constraints, and numerous other restrictions.

-5 -Enclosure 2

Appendix V Comments from the National Credit Union Administration

                                 See comment 6.

                                 See comment 7.

                                 See comment 8.

B. Composition of the FCUs

The Report provides an extensive review of the characteristics and growth
patterns of community-chartered FCUs. However, the proportion of FCUs that
are community-chartered, the need for charter conversions to ensure
continued viability, and the challenges community-chartered FCUs face when
converting from a single or multiple common bond charter, as well as other
issues, are not thoroughly addressed. For example:

    1. Despite recent growth in FCU community charters, they still only
       represent approximately 20 percent of FCUs and 30 percent of FCU
       membership. This is a significant portion of the FCU system, but, as
       noted in the Report, this growth has primarily been within the last
       five years. Additionally, it should be emphasized that much of this
       growth is a result of FCUs converting from an already existing
       occupational or associational FOM. Instead, the Report concentrates on
       the growth of this subset when characterizing the entire FCU system,
       in particular the perceived "blurring" of the distinction between FCUs
       and other depository institutions.
    2. A thorough assessment of the causes for the recent community charter
       conversions is not provided. The primary reason for these conversions
       has been to ensure continued viability of FCUs in changing economic
       and financial industry environments. A review of several examples
       documents this point. Clearview FCU (formally US Airways); Bethpage
       FCU (formally Grumman); JAX FCU (formally Jacksonville Naval Base);
       and New Cumberland FCU (formally New Cumberland Army Depot Defense
       Distribution Center) all converted to community charters in response
       to changes in their primary sponsors.
    3. The time necessary to successfully implement a different business
       model when converting to a community charter is not adequately
       addressed. This is critical since the cutoff for the SCF data is 2003,
       yet the period under review extends to and includes 2005.
       Consequently, the SCF does not allow for an assessment of any
       appreciable changes based on the recent growth of FCU community
       charters, as the majority of the conversions to a community charter
       have occurred since 2000, and 192 have occurred since 2003. Since the
       growth of community charters is discussed at length, it should also be
       fully explained that relative to the overall issue of reaching out to
       low- and moderate-income individuals, the impact of this growth can
       not be expected to be represented in the SCF data. Because the SCF
       does not overlay the time period of the review, its relevance is
       further diminished.
    4. The intent of NCUA's regulations pertaining to community charters is
       not accurately described. On page 1 the Report states: "As a result of
       recent

-6 -Enclosure 2

Appendix V Comments from the National Credit Union Administration

                                 See comment 9.

                                See comment 10.

                                See comment 11.

legal and regulatory developments, field of membership requirements for
credit unions have been relaxed - member groups now can include anyone who
lives, works, worships, or attends school in areas as large as whole
counties or major metropolitan areas." This statement suggests that the
affinity requirement (lives, works, worships...) of NCUA's field of
membership rules and the geographic limits on community charters are
recent developments. That suggestion is not accurate. Both of these NCUA
regulatory policies predate CUMAA. NCUA did grant community charters prior
to CUMAA that encompassed whole counties and metropolitan areas. It is
true that the documentation requirements for single political
jurisdictions were reduced through regulatory amendments that post-dated
CUMAA, but that change was based on NCUA's experience in chartering
communities constituting a single political jurisdiction.

    1. The size and extent of the community charters approved by NCUA are not
       appropriately represented. By using the approval of Los Angeles
       County, on page 13, as an example of a community charter conversion,
       it misrepresents the size of the community charter conversions
       commonly authorized. The data provided to GAO reflects the average
       population size for those community charter conversions approved
       during the period from 2000 to 2005 was 304,886, and the median size
       was 125,000.
    2. The Report states in the Highlights, as well as on page 6 and
       elsewhere, that NCUA's change in chartering policy is "triggered
       partly by concerns about competing with states with more expansive
       credit union chartering rules. . ." It is inaccurate to indicate that
       FOM parity with state-chartered credit unions is a primary objective
       when revising FOM policies for FCUs. Although this issue has surfaced
       during the regulatory comment period on proposed policy changes, it
       has not been a factor in NCUA's policy making.

C. The size and market share comparison of credit unions and banks

Although the Report attempts to compare credit unions to banks, it does
not provide a framework for an objective analysis, which, in addition to
the membership limitations discussed above, should reflect the relative
industry position of the two types of financial institutions.

As with all institutions in the financial industry, FCUs have evolved to
ensure their continued viability. Since 1934, dramatic changes in the
overall economic environment in which FCUs must operate have occurred.
These changes have required that FCUs adapt in order to meet the financial
needs and expectations of their members. Specifically, in the last forty
years, changing demographics in the United States were characterized both
by the loss of numerous well-paying blue collar jobs in the manufacturing
sector and an increasing disparity in the

-7 -Enclosure 2

Appendix V Comments from the National Credit Union Administration

                                See comment 12.

                                See comment 13.

income range between persons in the working class and the upper class.
Operational evolution can be seen at several levels, including the
offering of a wider range of services to a more broadly defined FOM.
Fundamentally, however, even though some FOMs are broader today, FCUs have
adhered to and preserved the integrity of both the common bond and their
cooperative structure, which is reflected in regulatory policies.

In addition, the types of services FCUs now increasingly offer have
changed. As with the common bond, FCUs have found it necessary to adapt in
order to meet member expectations and demand for products and services. On
page 1 the Report states "credit unions are now allowed to offer many
products and services similar to those provided by banks, such as real
estate and business loans." Such a conclusion, however, fails to
adequately assess the changing economic environment. Further, this
statement misrepresents the services credit unions have historically
provided. FCUs, for example, have been offering member business loans
since their inception, often providing loans to entrepreneurs initiating a
small business. As to the issue of mortgage lending, the FCU Act first
authorized mortgage lending for FCUs in 1978. State-chartered credit
unions in several states, most notably in the New England area, have
provided this type of lending since the 1950s.

In regard to rate comparisons, the Report recognizes the rate differences
between banks and credit unions on savings and lending products. However,
it should further recognize the interest rate environment during the
period of the GAO review when interest rates were at historic lows. An
assessment of the interest rate environment alone may have explained the
reason for the decreasing gap in the rate paid on savings. This analysis
is also crucial in assessing the mortgage rates since these loans of
long-term maturity significantly affect the asset/liability management and
ultimately the safety and soundness of a financial institution.

Additionally, as shown in Table 3, credit unions are an important, but
relatively small, segment of the financial industry. This size disparity
draws into question the appropriateness of the comparison and conclusions
in the Report.

                                -8 -Enclosure 2

Appendix V Comments from the National Credit Union Administration

                                See comment 14.

                                See comment 15.

                                    Table 3

                    Banks and Other Financial       Credit Unions Insured by
                 Institutions Insured by Federal    NCUSIF (Federally-Insured
     Total        Deposit Insurance Corporation8    Credit Unions)9
     Federally                        Total         Federal  State   Total        
      Insured  Commercial  Savings    FDIC     %of  Credit  Credit  NCUSIF   %of  
Year Deposits    Banks      Banks    Insured  Total  Union   Union  Insured Total 
2005 7,718,597 6,073,333  1,067,845 7,141,178 92.5% 321,831 255,588 577,419 7.5%  
2004 7,140,323 5,592,825    991,376 6,584,201 92.2% 308,318 247,804 556,122 7.8%  
2003 6,482,630 5,028,866   925,423  5,954,289 91.8% 291,485 236,856 528,341 8.2%  
                            Dollars shown in millions

NCUA also has concerns relating to the asset groups used in the Report for
the comparison between banks and credit unions. The smallest group size
used for comparative purposes in the Report is $100 million or less in
assets. It is not disclosed, however, that approximately 88 percent of
FCUs fall into that category, with 80 percent having assets less than $50
million as of September 30, 2005. It should also be noted that the average
asset size of FCUs is $73.2 million with the median asset size just $11
million.

5. NCUA's efforts to target credit union services to low-and
moderateincome individuals

One of GAO's stated objectives was to review NCUA's efforts to expand
credit union services to individuals of low-and moderate-income. The
Report correctly focuses on two principal programs in this context: (1)
NCUA's Low-Income Credit Union (LICU) program; and (2) NCUA's strategic
efforts to encourage FCUs to expand services into specifically designated
underserved areas. It also correctly notes that NCUA's support for these
programs has resulted in increased participation in both programs by FCUs
in recent years.

It is, however, inaccurate and inappropriate to use these programs to
define and assess service to people of modest means as they are
specifically targeted to low-income individuals. The legislative history
of the law creating the LICU program indicates that its purpose was "to
encourage saving and provide access to credit for low-income persons, and
to bring consumer education into poverty areas.. . ."^10 This
congressional action reflects recognition that the low-income segment of
the community is less financially capable, without assistance or special
consideration, of supporting a credit union bound by the traditional
constraints of common bond and cooperative structure. With its focus on
the new term "low income," Congress acknowledged that the traditional FCU

8

Information obtained from FDIC Statistics on Banking: A Statistical
Profile of the United States Banking Industry as published by FDIC,
Division of Insurance and Research, for 2003, 2004, and 2005.

9

Information obtained from Yearend Statistics for FICUs as published by the
National Credit Union Administration for 2003, 2004, and 2005.

10

115 Cong. Rec. S13997 (May 27, 1969) (statement of Sen. Scott).

-9- Enclosure 2

See comment 16.

Appendix V Comments from the National Credit Union Administration

membership base must necessarily be different, and broader. Although
Congress recognized the difference, it did not believe an amendment to the
overall statutory purpose for FCUs, which at that time was service to
persons of "small means," was required.

Instead, Congress implicitly endorsed FCU service to the traditional
membership base and specifically directed that NCUA should supply its own
definition of low income for purposes of implementing the provisions of
the new law. By regulation, NCUA did so, specifying that the term low
income means individuals who make less than either 80 percent of the
average for all wage earners, as established by the Bureau of Labor
Statistics, or whose household income is at or below 80 percent of the
national median household income as established by the Census Bureau.^11

To qualify for low-income designation, a credit union must have more than
50 percent of its membership consisting of individuals defined as low
income. This was a specific initiative by NCUA to recognize credit unions
that predominately served a low-income population but were challenged in
providing additional services and/or programs to their members. This
initiative opened opportunities for these credit unions to obtain
additional capital from philanthropic organizations and assistance from
the Department of the Treasury's Community Development Financial
Institution Fund (CDFI), the NCUA's Community Development Revolving Loan
Fund (CDRLF), and other organizations to enhance and expand services to
the low-income population.

Page 20 of the Report accurately describes the other unique
characteristics of LICUs and correctly notes LICUs grew in number between
2000 and 2005, from 632 to 1,032, a 63 percent increase. This result was
achieved with NCUA's vigorous encouragement and evidences dramatic success
in NCUA's effort to increase service to low-income members. Although NCUA
has not collected income and service usage data, the descriptive analyses
conducted by NCUA on the data collected in MSAP reflect LICUs and FCUs
with underserved areas are serving a relatively greater proportion of low-
and moderate-income individuals than the FCU system as a whole.

A more comprehensive analysis of the reasons that underlie NCUA's recent
policy change concerning expansion into underserved areas is warranted.
The American Bankers Association sued NCUA, challenging the decision to
allow a community based FCU to expand its service into an underserved
area. The fundamental issue in the case was the authority of NCUA to
authorize any FCU, regardless of charter type, to expand into underserved
areas. The Report should explain that settlement of the lawsuit resulted
in the prohibition of single-bond and community FCUs from adopting
underserved areas. This prohibition, which

11

12 C.F.R. S 701.34(a)(2). As originally implemented, NCUA's rule used 70
percent of median as the relevant percentage indicator of "low income."
The rule was changed to its current usage of 80 percent in 1993.

-10 -Enclosure 2

See comment 17.

Appendix V Comments from the National Credit Union Administration

is contrary to congressional intent, inhibits the ability of both types of
FCUs to increase service to low- and moderate-income individuals who are
outside the credit union's FOM.

In addition, the Report on page 22 uses Washington, D.C. as an example of
an underserved area approved by NCUA without regard to location. This
presentation is misleading. It is not explained that once an FCU
identifies an area meeting the underserved requirement, as defined in
Section 103(16) of the Community Development Banking and Financial
Institutions Act of 1994,^12 it must apply to NCUA to add the area to its
field of membership. A detailed marketing plan, emphasizing how the FCU
plans to reach out and serve all individuals in the underserved area, must
be submitted. A detailed business plan must also be submitted indicating
how the FCU will meet the needs of the individuals in the underserved area
by describing the products (e.g., free checking, micro-credit loans) and
services (e.g., bilingual staff, financial education seminars) the credit
union offers or is planning to offer. Once approved to serve a specific
underserved area, the credit union must maintain or open an office or
service facility in the underserved area within two years.

Other outreach initiatives by NCUA to increase service to underserved
individuals have not been sufficiently acknowledged or described. NCUA has
initiated several programs focused on assisting LICUs and on providing all
credit unions with best practices to consider when converting to community
charters or adding underserved areas. Since 1987, NCUA has administered
the CDRLF, which was established by Congress, to provide technical
assistance grants and low-cost loans for any LICU interested in enhancing
service to its membership. Under NCUA's auspices, the CDRLF has granted
273 loans totaling $40.5 million, and 1,923 grants totaling $5.8 million.

In addition to the CDRLF, the Access Across America initiative, announced
in February of 2002, incorporated NCUA's activities for small and
low-income designated credit unions, as well as those FCUs adopting
underserved areas. The program was designed to partner with federal
government agencies and other organizations to identify and facilitate use
of resources available for credit unions to assist in their efforts to
serve low- and moderate-income individuals. Workshops continue to provide
partnering opportunities with federal government agencies, as well as
non-profit and private organizations. This initiative has resulted in NCUA
entering into Memoranda of Agreement with the Internal Revenue Service,
Operation Hope, and the Department of Agriculture, each of which committed
to provide assistance in sharing opportunities with participating credit
unions. Moreover, NCUA maintains a working relationship with the
Department of Health and Human Services, CDFI, and Fannie Mae to provide
opportunities for credit unions to expand the products and services
particularly useful to those members with low- and moderate-incomes.

12

Pub. L. 103-325, 108 Stat. 2163 (Sept. 23, 1994)(codified at 12 U.S.C. SS
4701 et seq.).

-11 -Enclosure 2

Appendix V Comments from the National Credit Union Administration

As an adjunct to the Access Across America initiative, the Partnering and
Leadership Successes program was introduced in 2003 to provide best
practices in serving members and marketing to potential members in all
credit unions, especially in underserved areas and communities. The agency
coordinates widely attended workshops where a mix of credit unions present
programs focused on serving low- and moderate-income individuals. A few of
these programs include partnering opportunities with the Neighborhood
Reinvestment Corporation, Latino outreach, and micro-business lending
opportunities with the Small Business Administration.

In conjunction with these workshops, numerous Letters to Credit Unions
have been published that augment the workshops, providing information to
the credit union system about opportunities available to enhance service
and marketing to individuals in underserved areas.^13 Two early examples
of these letters include the February 2002 Letter to Federal Credit
Unions, Letter No. 02-FCU-02 titled Partnership Opportunities with IRS,
which introduced the credit union system to the Volunteer Income Tax
Assistance program, and the September 2001 Letter to Federal Credit
Unions, Letter No. 01-FCU-06 titled Financial Education Curriculum, which
announced FDIC's new Money Smart Financial Education Curriculum.

The overall objective of NCUA's initiatives is to provide increased
opportunities for FCUs to diversify their membership profile and to assist
small and low-income designated credit unions as they manage their
operations in compliance with the increasing number of complex laws and
regulations. If successful, the viability of some low-income designated
FCUs will be preserved, thus further enhancing the opportunity for low-
and moderate-income individuals in their FOM to join and participate in
the financial services offered by small and low-income designated FCUs.

Each of these initiatives was in direct response to CUMAA. But these types
of initiatives have long been a part of NCUA's, or its predecessor
agency's, regulatory fabric. There have been others, such as the 1960s era
initiative, undertaken jointly with the Office of Economic Opportunity, to
establish FCUs to serve low-income communities, the drive to increase the
number of LICUs, and the regulatory encouragement to add underserved
areas.

More recently, in 1993, NCUA created the Office of Community Development
Credit Unions which is dedicated to ensuring the long-term viability of
small and low-income designated credit unions. Today this activity is
handled by the Office of Small Credit Union Initiatives (OSCUI), which has
expanded considerably in terms of staff, resources, and programs.

OSCUI conducts regional and national training workshops on a variety of
topics to help small and low-income designated credit unions succeed. For
example, in

NCUA Home Page - http://www.ncua.gov - Letters to Credit Unions, 2001 to
2005.

                                -12 -Enclosure 2

Appendix V Comments from the National Credit Union Administration

                                See comment 18.

                                See comment 19.

2006 to date, OSCUI has held fifteen national workshops covering subjects
such as establishing financial literacy programs, disaster recovery
planning, and compliance with the Bank Secrecy Act. In addition to the
national workshops, OSCUI coordinates with NCUA's regional offices to
conduct smaller roundtable training sessions focused on the needs of small
and low-income designated credit union officials.

6. Transparency of Executive Compensation

NCUA agrees with the conclusion that credit union executive compensation
is not readily transparent. Absent compensation information captured by
IRS Form 990, it can be difficult for FCU members to ascertain the exact
compensation and benefits received by their executives. In the past, NCUA,
while not objecting to disclosure of this information, has deferred to
applicable state law on whether compensation and benefit information
should be disclosed.

As the Report points out, staff have indicated more efficient methods to
capture and disseminate executive compensation information in lieu of
filing Form 990. Such methods include: (1) amending NCUA's regulations to
require FCUs to include executive compensation information in their annual
reports; (2) requiring the reporting of such information in NCUA's
quarterly call reports; or (3) amending the standard FCU Bylaws to require
disclosure of compensation information during an FCU's annual membership
meeting. These and other methods may be considered by the NCUA Board in
evaluating the transparency of executive compensation.

While NCUA agrees FCU executive compensation is not readily transparent,
several matters in the Report warrant clarification. They include:

    1. Despite the absence of a standardized reporting mechanism, NCUA does
       not ignore the issue of executive compensation. Contrary to the
       implication on page 45, NCUA does assess executive compensation during
       the examination process primarily to determine its reasonableness as
       it relates to safety and soundness. There has never been a systemwide
       issue relating to executive compensation. As such, NCUA has not
       considered it necessary to collect or aggregate executive compensation
       data.
    2. On Page 42, it is implied that MSAP is deficient because it does not
       collect executive compensation information for banks, thereby
       preventing a direct comparison between FCUs and banks. However, it is
       not within NCUA's authority to collect data from banks or thrifts.
       Additionally, since this is not a safety and soundness issue for the
       credit union system, NCUA's authority to collect executive
       compensation extends only to FCUs.
    3. Comparing executive compensation of FCUs and banks was not a stated
       objective for GAO's study. Attempting to make a direct comparison is
       not

-13 -Enclosure 2

Appendix V Comments from the National Credit Union Administration

                                See comment 20.

                                See comment 21.

                                See comment 22.

only irrelevant to the issue of transparency, but is impossible given the
differences in the forms of compensation available to FCU versus bank
executives. For example, as the Report notes, stock options and stock
bonuses are routinely paid to bank executives, but are unavailable to
credit union executives. Nevertheless, the discussion of this matter seems
to imply that somehow credit union executive compensation may be askew.
Only by delving into the data provided in Appendix IV of the Report is it
clear that credit union executives on average make significantly less than
their banking counterparts.

    1. Since the Report addressed comparisons between senior officers of
       credit unions and banks, it should have also included a more detailed
       comparison between directors of credit unions and banks. It neither
       discusses nor includes any data regarding the compensation paid to
       directors of banks, which in some instances can be rather lucrative.
       At least some discussion would have been appropriate, especially since
       FCU boards are comprised of volunteers.^14 Including such data and
       discussion would have made for a more thorough and accurate comparison
       of executive compensation.
    2. The Report states on page 48 that MSAP will not stratify executive
       compensation by asset size of credit unions. This is not accurate.
       MSAP compensation data can be stratified into two statistically valid
       subsets based on asset size of the credit unions surveyed. In
       addition, limited descriptive conclusions can be derived from the data
       about other asset subgroups.

7. Conclusion

As referenced in MSAP and this response, NCUA recognizes the difficulty in
addressing the issues of membership profiles and the transparency of
executive compensation in the absence of comprehensive data. NCUA also
understands that the time allotted for completion of the Report did not
allow for consideration of the MSAP data and similar data being compiled
by NASCUS. Although the Report includes significant new detail and
qualifies its reliance on the SCF, NCUA anticipates the general
conclusions reached will be reported without the appropriate qualifiers.
In order to assure a complete and thorough understanding of the FCU
system, NCUA suggests that GAO include in its Report the information and
data contained in MSAP. It is also suggested that the completeness of the
Report would be further enhanced by inclusion of the data now being
collected by NASCUS, thus allowing for a thorough assessment of the entire
credit union system.

14

Pursuant to the FCU Act, no member of an FCU board may be compensated;
however, an FCU may compensate one individual who serves as an officer of
the board. For example, if the credit union's paid CEO is also a member of
the board. See 12 U.S.C. SS 1761(c) and 1761a.

-14 -Enclosure 2

       Appendix V Comments from the National Credit Union Administration

The  following  are   GAO's  comments   on  the   National  Credit   Union
Administration's letter dated November 14, 2006.

1. As noted in NCUA's letter, we did not receive the results of its pilot

  GAO Comments

survey on the membership profile of federal credit unions (Member Service
Assessment Pilot Program) in time to include it as part of our study. The
report can be found at NCUA's website www.ncua.gov.

    1. NCUA questioned GAO's use of low-and moderate-income as a proxy for
       the term modest means. As we note in our 2003 and current report,
       neither the legislative history of the Federal Credit Union Act, as
       amended, nor NCUA have established definitions as to what constitutes
       modest means. As a result, we used the low- and moderate-income
       categories that we defined in our 2003 report, which are based on what
       the other federal financial regulators use for Community Reinvestment
       Act purposes, as a proxy for modest means. Moreover, both citations
       identified by NCUA in the House and Senate reports for the bill that
       ultimately was enacted as CUMAA specifically identify low- and
       moderate-income as components of what is referred to as modest means.
       We agree that the term modest means also indicates a meaning broader
       than individuals with low- and moderate-income. Further, our analysis
       included comparisons between credit unions and banks of households
       with middle- or upper-incomes. This analysis showed that between 2001
       and 2004 credit unions continued to serve a higher proportion of
       middle- and upper-income households and a smaller proportion of low-
       and moderate-income households than did banks.
    2. NCUA stated that the text in footnote 27 of the draft report did not
       accurately reflect the income categories that the federal financial
       regulators established for CRA examinations. The text in question has
       been moved up into the body of the report and modified to more clearly
       state that our categories were based on, but not identical to, that
       used by the other federal financial regulators for CRA purposes. The
       primary difference between our income categories and those used for
       CRA purposes was the use of national median income rather than local
       metropolitan statistical area median income as a benchmark for the
       various income categories. We use the national measure since the SCF
       is a national survey. Further, we agree with NCUA's assertion that
       occupational and associational based credit unions have restricted
       membership bases, which limit their ability to serve all income
       categories. However, as we note in the report, although the number of

Appendix V Comments from the National Credit Union Administration

credit unions with single or multiple common bonds have been decreasing
since 2000 and the number of credit unions with more inclusive community
charters have been increasing, 2001 and 2004 SCF data indicated that
credit unions continue to serve a higher proportion of middle- and
upper-income households than banks.

    1. NCUA questioned our use of SCF data as the primary source for
       conclusions reached in the report regarding the income characteristics
       of credit union members. We believe that the report as stated clearly
       outlines the limitations of SCF data in conducting the analysis, but
       as we noted in our prior report, the SCF is the only source of
       comprehensive data to conduct such an analysis. We agree that there
       are other ways of analyzing and presenting these data. However, we
       believe that figure 2 in our report provides a valid comparison of
       bank and credit union customers in the SCF data. In addition, it uses
       the methodology of our 2003 report, which allows us to directly
       compare the results of our 2003 report with our current report. We
       focus on the median income, as we did in our prior report, since this
       measure is less susceptible to the influence of extreme values than
       the mean. As noted in the report, we performed an additional analysis
       using the median family income to provide additional context to our
       analysis within the same methodological framework.
    2. NCUA suggested that our report does not provide a framework for
       understanding the effect of statutory limitations on federal credit
       unions when comparing the income distribution of federal credit union
       members and bank customers. We explicitly acknowledged the importance
       of these limitations in our 2003 report and have added some additional
       text to reflect these limitations in our current report. Nevertheless,
       we believe that our analysis of SCF data on the income levels of
       credit union members versus bank customers provides important
       contextual information on the extent, if any, that credit union
       members are different from individuals that use banks. The lack of
       data on the income distribution of credit union members by charter
       type was one of the primary factors behind our recommendation that
       NCUA expand its pilot survey to allow the agency to systematically
       obtain and monitor credit union member income data by charter type.
    3. NCUA stated that the report does not thoroughly address the proportion
       of federal credit unions that are community chartered. We believe our
       report addresses this issue correctly, as originally presented. Both
       in table 1 of our report and the related text, we note

Appendix V Comments from the National Credit Union Administration

that despite the growth in community charters, multiple-bond credit unions
remain the largest group of federally chartered credit unions in number,
total membership, and assets. However, as we noted in our report, it is
important to emphasize that community-chartered credit unions overtook
multiple-bond credit unions as the largest of the three federal charter
types, in terms of average membership and average size in terms of assets,
beginning in 2003.

    1. NCUA stated that the report does not thoroughly address the agency's
       position on the need for charter conversions to ensure continued
       viability. We believe our report addresses this issue correctly, as
       originally presented. As noted in our report, we attributed to NCUA
       some of the causes for growth in the community charter, including the
       agency's belief that community charter expansion allows federal credit
       unions to attract a more diverse membership base that can enhance a
       credit union's economic viability or safety and soundness as well as
       provide greater opportunities to serve members of modest means. We
       further note in our report that NCUA explained that single- and
       multiple-bond credit unions often tend to be organized around employer
       or occupationally based associations, which in turn creates greater
       economic risk exposure since the membership base is intertwined with
       the economic cycles of a particular employer or occupation. Finally,
       we cite a Federal Reserve Bank of Atlanta research paper, which
       concluded that there are material benefits of credit union membership
       diversification and that these benefits derive from expanded
       investment opportunities and reduced concentration risk.
    2. NCUA stated that the time necessary to successfully implement a
       different business model when converting to a community charter is not
       adequately addressed. We believe our report addresses this issue
       correctly, as originally presented. Specifically, the report cites
       NCUA's belief that it would take time for any results to appear in the
       SCF data as credit unions seeking to expand into new areas and
       reaching new types of customers would face a learning curve in their
       efforts. Our report further notes that the latest available data from
       SCF are 2-years old, so any more recent changes would not be reflected
       in our analysis.
    3. NCUA stated that the intent of NCUA's regulations pertaining to
       community charters was not accurately described. Specifically, NCUA
       stated that introductory text in the draft report suggested that the
       affinity requirements of NCUA's field of membership rules and the
       geographic limits on community charters are recent developments.

Appendix V Comments from the National Credit Union Administration

NCUA noted that both of these regulatory policies predated CUMAA. We have
clarified the text of our report to reduce the potential for confusion by
stating that since the passage of CUMAA, NCUA has approved progressively
larger geographic-based fields of membership.

    1. Text has been added to reflect the average and median population size
       of community charter conversions approved from 2000 to 2005.
    2. NCUA stated that we inaccurately attributed its change in chartering
       policy as being triggered partly by concerns about competing with
       states having more expansive credit union chartering rules. As we
       reported in 2003, NCUA stated to us at that time that a major reason
       for its regulatory changes was to maintain the competitiveness of the
       federal charter in a dual (federal and state) chartering system. In
       subsequent discussions with NCUA they indicated that it would be more
       accurate to attribute changes in chartering policy to factors such as
       the continued viability of federal credit unions in changing economic
       and financial industry developments. We have modified the text of our
       report to reflect the influence of these factors.
    3. Text has been added to reflect that credit unions historically have
       had the ability to offer real estate and business loans.
    4. Text has been added to the report to recognize that interest rates
       during the period of our credit union and bank rate analysis were at
       historic lows.
    5. Text has been added to the background section of the report based on
       the information provided by NCUA in its comment letter regarding the
       proportion size of the credit union industry in comparison with other
       federally insured depository institutions and the relatively small
       size of most federally chartered credit unions. However, it is
       important to note that the disparity in size between the credit union
       and banking industries does not affect our rate analysis methodology
       or our conclusions since that analysis is broken out by asset
       groupings, starting with institutions with assets of $100 million or
       less.
    6. NCUA stated that it was inaccurate and inappropriate to use its
       Low-Income Credit Union program and underserved area expansion program
       to define and assess service to people of modest means. As noted
       previously, we used low- and moderate-income as a proxy for modest
       means due to a lack of a legislative or regulatory definition or

Appendix V Comments from the National Credit Union Administration

other criteria. Moreover, we note that NCUA's regulations for its
underserved program includes criteria (area in a metropolitan area where
the median family income is at or below 80 percent of the metropolitan
area median family income or the national metropolitan area median family
income) that is roughly similar to that used to define low- and
moderate-income for CRA purposes (less than 80 percent of the median
family income for the Metropolitan Statistical Area).

    1. We clarified in the report that both single-bond and community credit
       unions are currently not permitted to include underserved areas in
       their fields of membership. As noted in the report, the American
       Bankers Association contended that the Federal Credit Union Act allows
       multiple-bond credit unions, but it does not specifically identify
       single or community credit unions to add underserved areas to their
       field of membership.
    2. We added additional information in the report on NCUA's criteria for
       federal credit unions applying to include underserved areas in the
       credit union's field of membership. However, we disagree with NCUA's
       assertion that the example we provided in our report is misleading.
    3. We clarified in the report that NCUA examiners assess executive
       compensation during the examination process primarily to determine its
       reasonableness as it relates to safety and soundness, but that since
       it has not found a systemwide issue with executive compensation, NCUA
       has not considered it necessary to collect or aggregate executive
       compensation data.
    4. NCUA noted that our characterization of NCUA's Member Service
       Assessment Pilot implies that the pilot is deficient because it does
       not collect executive compensation information for banks; thereby,
       preventing a direct comparison between federal credit unions and
       banks. It also noted that it is not within NCUA's authority to collect
       data from banks or thrifts and that its authority to collect executive
       compensation data extends only to federal credit unions in the context
       of credit union safety and soundness issues. We do not intend to imply
       that collecting compensation data from banks is the responsibility of
       NCUA but point out the lack of available data that would allow a
       direct comparison of credit union and bank executive compensation.

Appendix V Comments from the National Credit Union Administration

    1. NCUA indicated that comparing executive compensation of federal credit
       unions and banks was not a stated objective for our study and that
       attempting to make a direct comparison is impossible, given the
       differences in the forms of compensation available to federal credit
       unions versus bank executives. We acknowledge that comparing executive
       compensation of federal credit unions and banks was not a stated
       objective for this study. Our report text merely points out that due
       to the lack of consistent, available, and transparent compensation
       data for credit unions, any overall comparison is difficult. For this
       reason, we did not provide bank executive compensation data in the
       main body of the report or make any direct comparisons between credit
       union and bank executive compensation. However, we believe that
       inclusion of bank executive compensation data in the appendix provides
       a useful benchmark on selected executive positions.
    2. NCUA noted that the report neither discusses nor includes any data
       regarding the compensation paid to directors of banks and that
       including such data and discussion would make a more thorough and
       accurate comparison of executive compensation. We acknowledge this
       point and added some additional discussion on bank director
       compensation for context.
    3. Our original characterization of NCUA's Member Service Assessment
       Pilot was based on a discussion with NCUA officials. We have revised
       the text of the report to reflect that compensation data that NCUA
       obtained can be stratified into two statistically valid subsets based
       on the asset size of the credit unions surveyed.

Appendix VI

                     GAO Contact and Staff Acknowledgments

  GAO Contact

Yvonne D. Jones, (202) 512-8678 or [email protected]

  Staff

In addition to the above contact, Harry Medina, Assistant Director; Janet
Fong; May Lee; John Lord; Donald Marples; Edward Nannenhorn; Jasminee
Persaud; Carl Ramirez; Barbara Roesmann; Paul Thompson; and Richard
Vagnoni made key contributions to this report.

  GAO's Mission

The Government Accountability Office, the audit, evaluation and
investigative arm of Congress, exists to support Congress in meeting its
constitutional responsibilities and to help improve the performance and
accountability of the federal government for the American people. GAO
examines the use of public funds; evaluates federal programs and policies;
and provides analyses, recommendations, and other assistance to help
Congress make informed oversight, policy, and funding decisions. GAO's
commitment to good government is reflected in its core values of
accountability, integrity, and reliability.

Obtaining Copies of GAO Reports and Testimony

The fastest and easiest way to obtain copies of GAO documents at no cost
is through GAO's Web site ( www.gao.gov ). Each weekday, GAO posts
GAO Reports and newly released reports, testimony, and correspondence on
its Web site. To
have GAO e-mail you a list of newly posted products every afternoon, go to
www.gao.gov and select "Subscribe to Updates."

                             Order by Mail or Phone

The first copy of each printed report is free. Additional copies are $2
each. A check or money order should be made out to the Superintendent of
Documents. GAO also accepts VISA and Mastercard. Orders for 100 or more
copies mailed to a single address are discounted 25 percent. Orders should
be sent to:

U.S. Government Accountability Office 441 G Street NW, Room LM Washington,
D.C. 20548

To order by Phone: Voice: (202) 512-6000 TDD: (202) 512-2537 Fax: (202)
512-6061

To Report Fraud, Waste, and Abuse in Federal ProgramsContact:

Web site: www.gao.gov/fraudnet/fraudnet.htm

  E-mail: [email protected]

Automated answering system: (800) 424-5454 or (202) 512-7470

Congressional Relations

Gloria Jarmon, Managing Director, [email protected] (202) 512-4400 U.S.
Government Accountability Office,  441 G  Street NW,  Room 7125  Relations
Washington, D.C. 20548

  Public Affairs

Paul Anderson, Managing Director, [email protected] (202) 512-4800
U.S.  Government  Accountability  Office,  441  G  Street  NW,  Room  7149
Washington, D.C. 20548

References

Visible links
  39. http://www.gao.gov/cgi-bin/getrpt?GAO-07-29
 137. http://www.gao.gov/cgi-bin/getrpt?GAO-06-220T
 138. http://www.gao.gov/cgi-bin/getrpt?GAO-04-91
 139. http://www.gao.gov/cgi-bin/getrpt?GAO-04-91
 140. http://www.gao.gov/cgi-bin/getrpt?GAO-04-91
 141. http://www.gao.gov/cgi-bin/getrpt?GAO-04-91
 145. http://www.gao.gov/cgi-bin/getrpt?GAO-04-91
*** End of document. ***