Agricultural Conservation: USDA Should Improve Its Process for	 
Allocating Funds to States for the Environmental Quality	 
Incentives Program (22-SEP-06, GAO-06-969).			 
                                                                 
The Environmental Quality Incentives Program (EQIP) assists	 
agricultural producers who install conservation practices, such  
as planting vegetation along streams and installing waste storage
facilities, to address impairments to water, air, and soil caused
by agriculture or to conserve water. EQIP is a voluntary program 
managed by the U.S. Department of Agriculture's (USDA) Natural	 
Resources Conservation Service (NRCS). NRCS allocates about $1	 
billion in financial and technical assistance funds to states	 
annually. About $650 million of the funds are allocated through a
general financial assistance formula. As requested, GAO reviewed 
whether USDA's process for allocating EQIP funds to states is	 
consistent with the program's purposes and whether USDA has	 
developed outcome-based measures to monitor program performance. 
To address these issues, GAO, in part, examined the factors and  
weights in the general financial assistance formula		 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-06-969 					        
    ACCNO:   A61304						        
  TITLE:     Agricultural Conservation: USDA Should Improve Its       
Process for Allocating Funds to States for the Environmental	 
Quality Incentives Program					 
     DATE:   09/22/2006 
  SUBJECT:   Agricultural programs				 
	     Conservation practices				 
	     Conservation programs				 
	     Environmental protection				 
	     Financial management				 
	     Natural resources					 
	     Performance measures				 
	     Pollution control					 
	     Pollution prevention				 
	     Program evaluation 				 
	     Soil conservation					 
	     Water conservation 				 
	     Federal aid to states				 
	     Evaluation criteria				 
	     Data collection					 
	     Environmental monitoring				 
	     Program goals or objectives			 
	     USDA Environmental Quality Incentive		 
	     Program						 
                                                                 

******************************************************************
** 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-06-969

     

     * Report to the Ranking Democratic Member, Committee on Agriculture,
       Nutrition, and Forestry, U.S. Senate
          * September 2006
     * AGRICULTURAL CONSERVATION
          * USDA Should Improve Its Process for Allocating Funds to States
            for the Environmental Quality Incentives Program
     * Contents
          * Results in Brief
          * Background
          * NRCS's Process for Allocating EQIP Funds to the States Does Not
            Clearly Address the Program's Purpose of Optimizing Environmental
            Benefits
               * NRCS Does Not Have A Specific, Documented Rationale for
                 Formula Factors and Weights
               * Financial Assistance Formula Relies on Some Questionable and
                 Outdated Data
          * NRCS Has Begun to Develop More Outcome-Oriented Performance
            Measures
          * Conclusions
          * Recommendations for Executive Action
          * Agency Comments and Our Evaluation
     * Objectives, Scope, and Methodology
     * EQIP 2006 Funding Allocation Formulas
     * Statistical Techniques to Determine Influential Factors in the 2006
       EQIP Financial Allocation Formula
          * Principal Components Regression
               * Data Used
               * Results
          * Factor Analysis of EQIP Environmental Variables
               * Explanation of the Technique
               * Results
     * Initial EQIP Funding Provided to the States, Fiscal Year 2006
     * Historical EQIP Funding Levels, Fiscal Years 2001-2006
     * Fiscal Year 2005 EQIP Obligations by Conservation Practice
     * Comments from the U. S. Department of Agriculture
     * GAO Contact and Staff Acknowledgments

Report to the Ranking Democratic Member, Committee on Agriculture,
Nutrition, and Forestry, U.S. Senate

September 2006

AGRICULTURAL CONSERVATION

USDA Should Improve Its Process for Allocating Funds to States for the
Environmental Quality Incentives Program

Contents

Tables

Figure

September 22, 2006Letter

The Honorable Tom Harkin Ranking Democratic Member Committee on
Agriculture, Nutrition, and Forestry United States Senate

Dear Senator Harkin:

Approximately two-thirds of the continental U.S.'s land area is used as
range, forest, crop, or pasture land. The production of food and fiber on
these lands contributes to the health of the U.S. population and the
strength of the nation's economy. If not properly managed, however,
agricultural production on these lands can damage the environment and the
nation's natural resources, as when routine agricultural activities
produce sediment, fertilizer runoff, and animal waste that can impair the
nation's waterways. Improper management of natural resources can also
reduce the productive capacity of agricultural land; for example,
excessive soil erosion may lead to soil lacking in nutrients. Agriculture
is also a major user of both groundwater and surface water, contributing,
in part, to water scarcity in the western United States. Responsible
production management practices can mitigate many of these problems.

The Environmental Quality Incentives Program (EQIP) provides financial and
technical assistance to agricultural producers who enter into contracts
with the U.S. Department of Agriculture's (USDA) Natural Resources
Conservation Service (NRCS) to install conservation practices on their
land. A primary purpose of EQIP is to optimize the environmental benefits
achieved using program funds. Managed by NRCS, EQIP is a voluntary program
established in 1996 that currently provides about $1 billion annually in
cost-share and incentive payments to farmers and ranchers in all 50
states, as well as U.S. territories, whose production practices may put
soil, water, air, and related natural resources at risk for environmental
damage.1 The program provides funds to help implement conservation
practices, such as planting vegetation along rivers and streams-known as
riparian buffers-to prevent sediment and other materials from polluting
the waters, and constructing waste storage facilities to reduce the level
of nutrients from livestock production that enter neighboring bodies of
water. The Farm Security and Rural Investment Act of 2002 (the act)
reauthorized EQIP and increased annual authorized program funding from
about $200 million in 1997 to current levels of over $1 billion.2

NRCS allocates the majority of EQIP funds through a general financial
assistance formula with 31 factors related to the availability of natural
resources and the presence of environmental concerns or problems. NRCS
assigns each of the formula's factors a weight that determines the funds
to be allocated to states based on that factor. The agency also
periodically modifies factor weights. Additional funds are distributed
using a second technical assistance formula that considers ongoing and
expected future conservation work, as well as through a performance bonus
formula designed to reward states for optimizing environmental benefits
and efficient program management.3 States disburse EQIP funds to producers
to install conservation practices on their land.

As requested, we assessed the extent to which (1) USDA's process for
providing funds to the states is consistent with the program's purpose of
optimizing environmental benefits and (2) USDA has developed measures to
monitor program performance.

To address these issues, we reviewed relevant statutory provisions and
NRCS's regulations and guidelines for implementing EQIP and spoke with
officials in NRCS's national headquarters. To review NRCS's efforts to
allocate EQIP funding to the states, we analyzed documents accounting for
NRCS's disbursements of EQIP funds. We examined the factors and weights in
the formula for general financial assistance and discussed the role of the
data source for each factor in the formula with NRCS's EQIP officials. We
gathered comments from stakeholders about the strengths and weaknesses of
NRCS's EQIP funding approach, selecting stakeholders from environmental
and farm organizations to obtain a broad set of views on the effectiveness
of the formula in allocating funds. To evaluate the extent to which NRCS
has developed sufficient outcome-based measures to monitor program
performance, we spoke with representatives from the NRCS teams responsible
for strategic planning and oversight activities and representatives from
the EQIP program team. We examined documentation of EQIP performance
measures and reviewed NRCS's Performance Results System.

A more detailed description of our objectives, scope and methodology is
presented in appendix I. We performed our work between December 2005 and
August 2006 in accordance with generally accepted government auditing
standards.

Results in Brief

NRCS's funding process is not clearly linked to EQIP's purpose of
optimizing environmental benefits; as such, NRCS may not be directing EQIP
funds to states with the most significant environmental concerns arising
from agricultural production. NRCS's general financial assistance formula
has several weaknesses that raise questions about the formula's usefulness
for effectively directing funds to states. Specifically, while the 31
factors in the financial assistance formula, and the weights associated
with each factor, give the formula an appearance of precision, NRCS does
not have a specific, documented rationale for why it included each factor
in the formula or for how it assigns and periodically adjusts factor
weights. Factors and weights are important for ensuring that funds are
distributed to states to address the nation's most significant
environmental problems arising from agriculture. Small adjustments in the
weights of the factors can shift the amount of funding directed at a
particular resource concern and, ultimately, the amount of money each
state receives. For example, in 2006, a 1 percent increase in the weight
of any of the 31 factors would have resulted in $6.5 million more
allocated on the basis of that factor at the expense of other factors. In
addition, some data in the EQIP financial assistance formula is
questionable or outdated. First, 5 of the data sources-such as acres of
nonirrigated cropland and federal grazing land-were used in the formula
more than once. Using the same data for multiple factors may result in
factors being indirectly weighted more than intended and may make the
formula less reliable for allocating state funding. Second, NRCS could not
identify the source of the data used in 10 of the 31 factors in the
formula, such as livestock animal units and animal waste generation and,
therefore, we could not verify the accuracy of the data or the basis on
which the agency was allocating funding. Finally, the formula does not use
the most current data available for at least 6 of the 31 factors. For
example, the formula uses 1995 data to measure commercial fertilizer use
on cropland, but we identified 2005 data that would have made this factor
more current. Because it was not clear how NRCS originally calculated this
data, we could not quantify the effect of using more recent data. However,
using less recent data raises questions about whether the formula
allocates funds to areas of the country that currently have the greatest
environmental needs. When we brought our concerns to NRCS's attention,
officials agreed that the formula, including weights and data sources,
needed to be reviewed. NRCS subsequently announced plans to issue a
request for proposal soliciting comments and suggested revisions to NRCS's
formulas for allocating conservation funds, including the EQIP financial
assistance formula.

As part of its 2005 strategic planning effort, NRCS developed long-term,
outcome-based measures to assess changes to the environment resulting from
EQIP practices. NRCS has developed baselines for these measures and plans
to assess and report on them once computer models and other data
collection methods that estimate environmental change are completed. In
the meantime, NRCS will continue to use the results of its existing annual
measures to assess performance. As NRCS collects additional data about its
accomplishment of long-term performance measures, it may ultimately have
more complete information on which to gauge program performance. Such
information could help the agency refine its process for allocating funds
to the states via its financial assistance formula by directing funds
toward areas of the country that need the most improvement.

We are making recommendations to the Secretary of Agriculture to better
align NRCS's process for allocating EQIP funds with the program's stated
purpose of optimizing environmental benefits. In particular, we are
recommending that NRCS ensures that its rationale for the factors and
weights is documented and linked to program priorities, its data sources
are accurate and current, and it uses information about long-term program
performance to ensure funds are directed to areas of the highest priority.
We provided USDA with a draft of this report for review and comment. USDA
agreed that the EQIP allocation formula needs review. USDA did not agree
with our assessment that NRCS's funding process lacks a clear link to the
program's purpose of optimizing environmental benefits. The agency stated
that its use of factors related to the natural resource base and condition
of those resources shows the general financial assistance formula is tied
to the program's purpose of optimizing environmental benefits. USDA also
stated that, while some formula data sources and weights will be updated,
the types of factors used would be needed in any process that attempts to
inventory and optimize environmental benefits. While this may in fact be
the case, USDA needs to document this connection-that is, why factors were
chosen and weights assigned. USDA could make the connection between the
formula and the program's purpose of optimizing environmental benefits
more evident if it provided additional information describing its reasons
for including or excluding factors in the formula and its rationale for
assigning and modifying weights.

Background

The U.S. agricultural sector benefits our economy and the health of our
nation. However, if not properly managed, agricultural activities can
impair the nation's water, air, and soil; disrupt habitat for endangered
species; and constrain groundwater resources. For example, sediment
produced during routine agricultural activities may run off the land and
reach surface waters, including rivers and lakes. Sediment can destroy or
degrade aquatic habitat and can further impair water quality by
transporting into area waters both the pesticides applied to cropland and
the nutrients found in fertilizers and animal waste.4 These and other
water quality issues are of concern in a number of U.S.
agriculture-producing regions, including the Midwest and along the
Mississippi River. Agriculture is also a major user of groundwater and
surface water, which has led to water resource concerns across the
country, particularly in the West. In 2000, irrigation accounted for 65
percent of the nation's consumption of fresh water. Agricultural
production can also impair air quality, when wind carries eroded soil,
odors, and smoke, and may lead to the loss of wetlands, which provide
wildlife habitat, filter pollutants, retain sediment, and moderate
hydrologic extremes.

EQIP is one of a number of USDA conservation programs designed to mitigate
agriculture's potentially negative environmental effects. EQIP provides
cost-share funds and incentive payments for land used for agricultural
production and supports around 190 conservation practices, including
constructing facilities to temporarily store animal waste; planting rows
of trees or shrubs to reduce wind erosion and provide food for wildlife;
and planning the amount, form, placement, and timing of the application of
plant nutrients. EQIP is designed to fund conservation practices in a
manner that helps the program achieve the following national priorities
identified by NRCS:

o reducing nonpoint source pollution (nutrients, sediment, pesticides, or
excess salinity), groundwater contamination, and pollution from point
sources (such as concentrated animal feeding operations);

o conserving groundwater and surface water resources;

o reducing emissions that contribute to air quality impairment;

o reducing soil erosion from unacceptable levels on agricultural land; and

o promoting at-risk species habitat conservation.

The Federal Agriculture Improvement and Reform Act of 1996 created EQIP by
combining four existing conservation programs into a single program.5 The
Farm Security and Rural Investment Act of 2002, the farm bill,
reauthorized EQIP and increased its authorized funding from about $200
million in 1997 to current levels of over $1 billion.6 The 2002 act
required that at least 60 percent of EQIP funds be made available for
conservation practices relating to livestock production.7 In addition, it
authorized EQIP funds for specific conservation purposes-(1) funds for
producers to install water conservation practices to improve groundwater
and surface water conservation (the Ground and Surface Water Conservation
component of EQIP) and (2) funds for water conservation

practices in the Klamath Basin located on the California/Oregon border
(the Klamath Basin component of EQIP).8

Annually, NRCS headquarters officials determine the amount of funding each
state receives, while state and local NRCS officials decide what
conservation practices to fund in their state and local communities. The
total amount of EQIP funding a state receives can be derived by adding
together that state's funding for all categories. Table 1 describes the
different categories of funding that states received for fiscal year 2006
and NRCS's process for allocating that funding.

Table 1: Fiscal Year 2006 Categories of EQIP Funding

                                        

     EQIP funding       Funding purpose    Process for allocating  Percentage 
       category                                    funding           of total 
                                                                      funding 
General financial Cost-share and        Funds are divided among        65% 
assistance        incentive payments    states using a          
                     for installing        31-factor formula that  
                     conservation          considers the presence  
                     practices.            of available natural    
                                           resources and           
                                           environmental concerns  
                                           in each state. Each     
                                           factor is assigned a    
                                           weight, which           
                                           determines the amount   
                                           of money to be given to 
                                           states based on that    
                                           factor.                 
General technical Funds for technical   Technical assistance            19 
assistance        specialists' time.    dollars are divided     
                     Among other           among states based on   
                     activities,           the number of ongoing   
                     specialists process   EQIP contracts and      
                     EQIP administrative   expected future         
                     paperwork, advise     technical specialist    
                     farmers about the     needs.                  
                     installation of                               
                     practices, and                                
                     inspect installed                             
                     practices.                                    
Ground and        Funds for             Groundwater and surface          7 
Surface Water     conservation          water funds are         
Conservationa     practices that        allocated to eight High 
                     improve groundwater   Plains Aquifer states,  
                     and surface water     nine western drought    
                     conservation.         states, and other       
                     Practices must result states with             
                     in a net savings of   agricultural water      
                     groundwater or        needs using a formula   
                     surface water         based on groundwater,   
                     resources.            irrigation, and other   
                                           agricultural water      
                                           usage factors.          
Performance       Bonuses designed to   Performance bonuses are          4 
incentive         reward states that    divided among states    
bonusesa          achieve a high level  using a formula with    
                     of program efficiency seven factors.          
                     and optimize                                  
                     environmental                                 
                     benefits. States can                          
                     use bonuses as they                           
                     do other EQIP                                 
                     financial and                                 
                     technical assistance.                         
EQIP Colorado     Funds for salinity    Colorado Salinity                2 
Salinitya         control measures in   dollars are divided     
                     the Colorado River    between Colorado, Utah, 
                     Basin.                and Wyoming based on    
                                           the amount of land in   
                                           each state needing      
                                           salinity control        
                                           treatment.              
EQIP regional     Funds provided to     Regional equity funds            2 
equitya           states that receive   are provided to states  
                     less than $12 million that receive less than  
                     from NRCS             $12 million from NRCS   
                     conservation programs conservation programs   
                     (including EQIP) in a (including EQIP) in a   
                     given fiscal year.b   given fiscal year.b     
                     States can use funds  Headquarters officials  
                     as they do other EQIP determine the amount of 
                     financial and         funds to be provided to 
                     technical assistance. each state and from     
                                           which program the funds 
                                           will come.              
Klamath Basina    Funds to carry out    Klamath Basin funding           1% 
                     water conservation    is split evenly between 
                     activities in the     California and Oregon.  
                     Klamath Basin in                              
                     California and                                
                     Oregon.                                       

Source: GAO analysis of NRCS documentation.

Note: EQIP funds are also provided to producers through Conservation
Innovation Grants, funds competitively awarded for the development and
adoption of innovative conservation approaches and technologies. In fiscal
year 2006, around $20 million in grants was approved by NRCS. Conservation
Innovation Grants are awarded through national and state competitions to
producers demonstrating innovative approaches to conservation. Because the
grant money for national competitions is not provided to states along with
their initial EQIP allocations, it is not reflected in this table.

aNRCS provides these funds to the states through both financial and
technical assistance; the majority of the assistance is in the form of
financial assistance.

bIn fiscal year 2006, the threshold was lowered administratively to $11
million.

As the table shows, each category of EQIP funding is allocated to the
states using a different process. For the general financial assistance
formula, the availability of natural resources accounts for approximately
half of the funds allocated, and the presence of environmental concerns or
problems accounts for the remainder.  9 Table 2 shows the factors and
weights used in the financial assistance formula for fiscal year 2006.

Table 2: EQIP General Financial Assistance Formula Factors and Weights,
Fiscal Year 2006

Factora                                                             Weight 
Acres of nonirrigated cropland                                         3.2 
Acres of irrigated cropland                                            4.3 
Acres of federal grazing lands                                         0.5 
Acres of nonfederal grazing lands                                      4.3 
Acres of forestlands                                                   1.1 
Acres of specialty cropland                                            3.2 
Acres of wetlands and at-risk species habitat                          4.6 
Acres of bodies of water                                               3.2 
Livestock animal unitsb                                                5.8 
Animal waste generation                                                5.8 
Waste management capital cost                                          3.5 
Acres of American Indian tribal lands                                  3.3 
Number of limited resource producers                                   5.0 
Acres of grazing land lost to conversion                               0.8 
Air quality nonattainment areas                                        1.4 
Acres of pastureland needing treatment                                 5.5 
Acres of cropland eroding above Tc                                     6.2 
Acres of fair and poor rangeland                                       6.2 
Acres of forestlands eroding above Tc                                  1.4 
Acres of cropland and pastureland soils affected by saline and/or      2.6 
sodic conditionsd                                                   
Miles of impaired rivers and streams                                   3.6 
Potential for pesticide and nitrogen leaching                          1.3 
Potential for pesticide and nitrogen runoff                            1.7 
Ratio of livestock animal units to cropland                            1.7 
Number of concentrated animal feeding operations/animal feeding        2.8 
operationse                                                         
Ratio of commercial fertilizers to cropland                            0.9 
Wind erosion above Tc                                                  4.2 
Phosphorous runoff potential                                           3.9 
Riparian areas                                                         0.8 
Carbon sequestration                                                   3.6 
Coastal zone land                                                      3.6 

Source: NRCS.

aThe factor names in this chart are NRCS terminology. In certain cases,
they may not represent what is actually being measured. For example, the
factor for acres of cropland and pastureland soils affected by saline
and/or sodic conditions only measures the presence of salts on cropland
and pastureland and does not include data on the presence of sodium on
these lands.

bAnimal units are a standard way of quantifying livestock of different
types and sizes (e.g., cattle, dairy, poultry, etc.) One animal unit is
equivalent to 1,000 pounds of live animal weight.

cT is a term that refers to a tolerable rate of erosion. T is the maximum
rate of annual soil loss that will permit crop productivity to be
sustained economically and indefinitely on a given soil.

dSaline and sodic soils are soils that contain salts and sodium. Excess
amounts of salt and sodium in soils may adversely affect soil quality and
crop productivity.

eAnimal feeding operations are facilities where animals are raised in
confined or semiconfined situations usually with feed brought to the
animals. When large enough or when in environmentally sensitive locations,
these facilities are designated as concentrated animal feeding operations
and become subject to regulatory requirements to prevent point source
pollution.

In fiscal year 2006, approximately $652 million was divided among the
states through the general financial assistance formula.10 For example,
according to the formula, EQIP funding for nonirrigated cropland
(accounting for 3.2 percent of financial assistance) totaled $20.9
million. The state with the most acres of nonirrigated cropland received
$1.7 million of the funds associated with this factor, and the state with
the fewest acres of nonirrigated cropland received approximately $1,100. A
state's total allocation is composed of the funds it receives for each of
the 31 factors.

Although about 65 percent of EQIP funds are provided through the general
financial assistance formula, other categories of funding can have a
significant effect on the total amount of funds an individual state
receives. For example, 35 percent of Utah's fiscal year 2006 allocation
was from general financial assistance. The largest category of EQIP funds
Utah received-38 percent-was Colorado Salinity funds. Appendix II provides
additional information on the 2006 funding allocation formulas for general
financial assistance, Ground and Surface Water Conservation, performance
incentive bonuses and Klamath Basin funding categories.

Figure 1 shows the initial distribution of NRCS's fiscal year 2006 EQIP
allocations to the states in November 2005. States had to return any
unused funds by June 2006 for redistribution to states with a need for
additional funds. Appendix IV describes the amount of funding each state
initially received in fiscal year 2006.

Figure 1: Initial EQIP Funding to States, Fiscal Year 2006

NRCS's Process for Allocating EQIP Funds to the States Does Not Clearly
Address the Program's Purpose of Optimizing Environmental Benefits

NRCS's process for providing EQIP funds to the states is not clearly
linked to the program's purpose of optimizing environmental benefits. In
particular, NRCS's general financial assistance formula, which accounts
for approximately two-thirds of funding provided to the states, does not
have a specific, documented rationale for each of the formula's factors
and weights. In addition, the financial assistance formula relies on some
questionable and outdated data. As a result, NRCS may not be directing
EQIP funds to states with the most significant environmental concerns
arising from agricultural production.

NRCS Does Not Have A Specific, Documented Rationale for Formula Factors
and Weights

Although the 31 factors and weights used in the general financial
assistance formula give it an appearance of precision, NRCS does not have
a clearly documented rationale for including each factor in the formula
and assigning or modifying each weight. The original EQIP formula was
created in 1997 by an interagency task force that modified the formula
created for a different conservation program-the Conservation Technical
Assistance Program.11 The task force added and deleted factors and
adjusted factor weights so that the EQIP formula better corresponded to
the Federal Agriculture Improvement and Reform Act of 1996's requirement
that 50 percent of funds be targeted at funding livestock-related
practices.

Since the creation of the financial assistance formula, NRCS has
periodically modified factors and weights to emphasize different program
elements and national priorities, most recently in fiscal year 2004
following the passage of the 2002 Farm Security and Rural Investment Act.
Furthermore, NRCS officials stated that they meet annually to review the
allocation of funds to states. However, throughout this process, NRCS has
not documented the basis for its decisions to modify factors and weights
or documented how changes to its formula achieve the program's purpose of
optimizing environmental benefits. Thus, it is not always clear whether
the formula factors and weights guide funds to the states as effectively
as possible. For example, it is unclear why NRCS includes a factor in the
formula that addresses the waste management costs of small animal feeding
operations but not a factor that addresses such costs for large
operations-large operations can also damage the environment and are
eligible for EQIP funding.12 By not including the costs of the larger
operations in its financial assistance formula, some states may not be
receiving funds to address their specific environmental concerns. In
addition, NRCS has not demonstrated that it has the most appropriate water
quality factors in its formula. For example, the formula includes a factor
addressing river and stream impairment but no factor for impaired lakes
and other bodies of water. Moreover, it is not certain whether the
impaired rivers and streams factor results in funds being awarded on the
basis of general water quality concerns or water pollution specifically
caused by agricultural production. As a result, it was not certain whether
the formula allocates funds as effectively as possible to states with
water quality concerns arising from agricultural production.

While the factors in the EQIP general financial assistance formula
determine what resource and environmental characteristics are considered
when allocating funds, the weights associated with these factors directly
affect how much total funding is provided for each factor and, thus, the
amount of money each state receives. Factors and weights are key to
ensuring states with the greatest environmental problems receive funding
to address these problems. Small differences in the weights of the factors
can shift the amount of financial assistance directed at a particular
resource concern and, ultimately, the amount of money provided to a state.
In 2006, if the weight of any of the 31 factors had increased by 1
percent, $6.5 million would have been allocated on the basis of that
factor at the expense of one or more other factors. Such a shift could
impact the amount of financial assistance received by each state. For
example, a 1 percent increase in the weight of the specialty cropland
factor with a corresponding decrease of 1 percent in the American Indian
tribal land factor could result in large changes to the distribution of
EQIP general financial assistance. According to our analysis, the state
benefiting the most from such a change would receive $2.6 million more (a
7.2 percent increase in that state's level of general financial
assistance) and the state benefiting least from such a change would lose
$2.7 million (a 13.5 percent decrease in that state's level of general
financial assistance). The potential for the weights to significantly
affect the amount of funding a state receives underscores the importance
of having a well-founded rationale for assigning them. To date, NRCS has
not documented its rationale for choosing the weights.

Some stakeholders we spoke with questioned NRCS's assignment of weights to
certain factors in the financial assistance formula because they did not
believe NRCS's formula adequately reflected the states' environmental
priorities. For example, NRCS's general financial assistance formula
allocates 6.3 percent of EQIP funds to the states based on factors
specifically associated with animal feeding operations.13 However, states
spent more of their EQIP financial assistance on related practices, which
suggests that the weights in the financial assistance formula may not
reflect states' priorities. In fiscal year 2005, states spent a total of
11 percent of EQIP financial assistance, or $91.1 million, on one such
practice-the construction of waste storage facilities for animal feeding
operations. (App. VI outlines the practices funded in fiscal year 2005,
including other practices to control pollution from animal feeding
operations.) More generally, other stakeholders said that, as the program
develops, NRCS should give additional weight to factors related to the
presence of environmental concerns in a state and place less emphasis on
factors related to natural resources in a state. They believed this
reassignment of weights would better ensure that states contending with
the most significant environmental problems receive the most funding.
Currently, factors related to the presence of environmental concerns
account for approximately half of the total funding, while factors
relating to the availability of natural resources account for the
remainder. Factors related to the availability of natural resources
provide states that have significant amounts of a particular type of
land-such as grazing land or cropland-with more funds, regardless of
whether that land is impaired.

Although NRCS has stated that it meets annually to review its allocation
of funds to states, officials told us they had not conducted any
statistical analysis to examine the influence of factors on funding
outcomes. Statistical analyses can provide information on how the factors
in the allocation formula have affected the distribution of funds, thereby

providing information to improve program implementation.14 To better
understand the effect of the factors on the allocations to states, we used
two types of statistical analysis to assess the effects of the EQIP
financial assistance formula on state funding: (1) regression analysis to
show which factors are the most influential in determining funding levels
and (2) factor analysis to understand how factors can be grouped and
identified with program priorities.

Our regression analysis for the fiscal year 2006 funding allocation shows
that the factors that were the most important in explaining the
distribution of general financial assistance to states were acres of fair
and poor rangeland, acres of nonfederal grazing lands, livestock animal
units, acres of irrigated cropland, acres of American Indian tribal lands,
and wind erosion above T. This analysis suggests that regions of the
country with these types of characteristics are more likely to benefit
from the current formula. On the other hand, a few factors, such as acres
of forestlands, potential for pesticide and nitrogen leaching, and air
quality nonattainment areas were not significantly related to the
allocation, indicating that they had little or no impact on the formula.

Our factor analysis, which groups the data into a smaller number of
categories that actually drive the formula, found that the largest
grouping with the greatest amount of correlation, included acres of
nonfederal grazing land, acres of fair and poor rangeland, livestock
animal units, and wind erosion above T-all indicative of dryland
agriculture and livestock feeding and ranching. These results correspond
with those of our regression analysis and help to show how the current
national allocation formula prioritizes money to states. A complete
explanation of both analyses is included in appendix III.

Financial Assistance Formula Relies on Some Questionable and Outdated Data

Weaknesses in the financial assistance formula are compounded by NRCS's
use of questionable and outdated data. Accurate data are key to ensuring
that funds are distributed to states as intended. However, we identified
several methodological weaknesses in the data sources: (1) data that were
used more than once in the formula, (2) data sources whose accuracy could
not be verified, and (3) data that was not as recent as possible.

First, 5 of the 29 data sources behind the factors in the financial
assistance formula were used more than once, potentially causing NRCS to
overemphasize some environmental concerns at the expense of others.
Specifically:

o NRCS uses the same data to estimate pesticide and nitrogen runoff and
phosphorous runoff in its formula. According to NRCS, because data
measuring the potential for phosphorous runoff were unavailable, it
substituted data measuring the potential for pesticide and nitrogen
runoff. The agency did so believing that similar characteristics cause
both types of runoff. However, an NRCS official responsible for deriving
the runoff and leaching indicators commented that the substitution of one
type of runoff data for another was problematic because the mechanisms
through which pesticides and nitrogen are transported off-site to cause
environmental problems are different from those of phosphorous. A 2006
NRCS cropland report estimates that the intensity of nitrogen and
phosphorous losses may differ geographically.15 For example, nitrogen
dissolved in surface water runoff in the upper Midwest accounts for 28
percent of the national total, while phosphorous dissolved in surface
water runoff in the same region accounts for 45 percent of the national
total. This difference in the effect of these two pollutants in the same
region raises questions about the appropriateness of substituting one type
of data for the other. Until adequate data are available for a given
factor, it may not be appropriate to include that factor in the general
financial assistance formula.

o NRCS's formula uses nonirrigated cropland, federal grazing land,
nonfederal grazing land, and forestland once for estimating acreage and
then again for estimating carbon sequestration.16 According to NRCS, the
agency did not have good source data to measure potential areas where
management practices could improve levels of carbon sequestration so it
substituted these other data sources. While we could not fully assess the
soundness of NRCS's estimate of carbon sequestration, some academic
stakeholders we spoke with questioned whether NRCS had estimated carbon
sequestration as effectively as possible and noted that alternate data
sources were available. In discussing these alternate sources with NRCS,
the EQIP Manager said the agency had not previously considered using these
sources for the EQIP formula, but that they could prove relevant.

Using the same data for multiple factors may result in factors being
indirectly weighted higher than intended. For example, the effective
weight of the pesticide nitrogen runoff factor is 5.6 percent-the sum of
the original pesticide nitrogen runoff weight (1.7 percent) and the
phosphorous runoff weight (3.9 percent). Using data created for one factor
for a second factor also makes the formula less transparent and
potentially less reliable for allocating state funding.

Second, NRCS could not confirm the source of data used in 10 factors in
the formula; as such, we could not determine the accuracy of the data,
verify how NRCS generated the data, or fully understand the basis on which
the agency allocates funding. Specifically, we could not confirm the
source of data for acres of federal grazing land, livestock animal units,
animal waste generation, acres of cropland eroding above T, acres of
forestlands eroding above T, ratio of animal units to cropland, miles of
impaired rivers and streams, ratio of commercial fertilizers to cropland,
riparian areas, and coastal zone land.17 For example, we could not verify
how data for the livestock animal units and animal waste factors were
generated, and NRCS said it had not retained documentation of how the data
for these factors were calculated. As a result, it was uncertain whether
NRCS had chosen the most appropriate data as its basis for allocating
funds to states with pollution problems from livestock and animal waste or
whether the data were accurately calculated. EQIP officials told us that,
in most cases, the data sources had been chosen and incorporated into the
formula before they were involved with EQIP and that documentation had not
been kept to identify how data sources were used.

In addition, for one factor- the number of limited resource producers in a
state-we found that the data did not measure what its factor name
indicated. NRCS defines a limited resource producer as one who had, for
the last 2 years, (1) farm sales not more than $100,000 and (2) a
household income at or below the poverty level, or less than 50 percent of
the county median household income.18 However, the data NRCS uses in the
general financial assistance formula only captures farms with low sales,
which does not necessarily indicate whether producers on those farms have
limited means. As a result, NRCS may not be directing funds to states
having farmers with the most limited resources. A description of each
factor in the fiscal year 2006 general financial assistance formula can be
found in appendix II.

Third, NRCS does not use the most current data for six factors in the
formula-livestock animal units, animal waste generation, number of limited
resource producers, miles of impaired rivers and streams, ratio of
livestock animal units to cropland, and ratio of commercial fertilizers to
cropland.19 According to NRCS, the source of data on the ratio of
commercial fertilizers to cropland was a 1995 report by the Association of
American Plant Food Control Officials; we found a 2005 version of the same
report with more current data. In other cases, we identified more current,
alternate sources of data. For example, the formula currently uses 1996
EPA data for its waste management capital cost factor but could use 2003
NRCS data that estimates waste management costs.  20 Not using recent data
raises questions about whether the formula allocates funds to areas of the
country that currently have the greatest environmental needs, because
recent changes in a state's agricultural or environmental status may not
be reflected. According to our analysis, by using more current data for
the number of limited resource producers factor, one state would have
received approximately $151,000 more in fiscal year 2006 (a 0.2 percent
increase in that state's general financial assistance), and another state
would have received approximately $138,000 less (a 1.3 percent decrease in
that state's general financial assistance).21 Because we were unable to
determine how NRCS used the data for developing the remaining five
factors, we could not determine what impact using more current data for
those factors would have on financial assistance provided to states.
According to NRCS, the alternate sources we identified appeared to be
acceptable for use in the formula, and the agency is in the process of
updating the formula's livestock data.

In addition to these six factors, data used to measure acres of riparian
areas, fair and poor rangeland, and forestland eroding above T are about
20 years old and will likely become more inappropriate over time.

When we brought our concerns to NRCS's attention, officials agreed that
the formula, including weights and data sources, needed to be reexamined.
NRCS subsequently announced plans to issue a request for proposal
soliciting comments and suggested revisions to NRCS's formulas for
allocating conservation funds, including the EQIP financial assistance
formula. In addition, according to NRCS's EQIP Manager, the agency is in
the process of consolidating the data used in the financial assistance
formulas for its conservation programs into a single database. As a part
of this process, the agency plans to review its data sources for the
formula factors and update them with more relevant and current data when
possible.

NRCS Has Begun to Develop More Outcome-Oriented Performance Measures

NRCS has recently begun to develop program-specific, long-term measures to
monitor EQIP's outcomes. In 2000, we reported that performance measures
tied to outcomes would better communicate the results NRCS intended its
conservation programs to achieve.22 As part of its 2005 strategic planning
effort, NRCS developed outcome-based, long-term measures to assess changes
to the environment resulting from the installation of EQIP conservation
practices.23 These measures include such things as reduced sediment
delivery from farms, improved soil condition on working cropland, and
increased water conservation. Previously, in 2002, NRCS established annual
measures that primarily assess program outputs-the number and type of
conservation practices installed. Table 3 outlines NRCS's seven annual
performance measures for fiscal year 2006, and table 4 describes its seven
long-term EQIP performance measures approved in 2005.

Table 3: EQIP Annual Performance Measures, Fiscal Year 2006

                                        

          Performance measure        Measure unit Progress as of  Fiscal year 
                                                    September 1, target as of 
                                                            2006 September 1, 
                                                                        2006a 
Comprehensive nutrient management Number of             2,189        2,488 
plans applied                     plans                       
Comprehensive nutrient management Number of             2,231        2,435 
plans written                     plans                       
Grazing land with conservation    Acres            11,640,329   10,454,337 
practices to protect the resource                             
base                                                          
Improved irrigation efficiency    Acre-feet          641,158b     543,204c 
Nonfederal land managed to        Acres             1,163,850      381,124 
protect species with declining                                
populations                                                   
Reduction of cropland soils       Acres             1,345,101    1,360,622 
damaged by erosion                                            
Soil erosion reduced              Tons             16,230,336    9,912,788 

Source: NRCS.

aAccording to NRCS, performance targets may change as additional funds are
provided to the states and as states return unused funds to headquarters.

bThis figure represents combined progress for EQIP, Ground and Surface
Water Conservation, and Klamath Basin.

cThis figure represents a combined target for EQIP, Ground and Surface
Water Conservation, and Klamath Basin.

Table 4: EQIP Long-term Measures

Performance measure    Measure unit             Baseline year Proposed     
                                                                 target       
Improve soil condition Millions of acres moved  .5 in 2005    2.7 by 2010  
on working cropland    to a soil conditioning                 
                          index level > than 0.a                 
Reduce potential       Million tons per year    2.4 in 2004   18.5 by 2010 
sediment delivery from                                        
agricultural                                                  
operations                                                    
Reduce potential       Tons                     18,200 in     100,000 by   
nitrogen delivery from                          2005          2010         
agriculture                                                   
Reduce potential       Tons                     2,700 in 2005 14,000 by    
phosphorus delivery                                           2010         
from agriculture                                              
Increase water         Acre-feet                600,000 in    4,200,000 by 
conservation                                    2005          2010         
Improve grassland      Million acres            10.3 in 2005  52 by 2010   
condition, health, and                                        
productivity                                                  
Improve the quality of Million acres            .45 million   2.4 by 2010  
habitat for at-risk                             in 2005       
species                                                       

Source: NRCS.

aThe National Resources Inventory (NRI) includes data on soil type, soil
characteristics, and soil interpretations, in addition to historical
information on land use, management practices, and erosion. These data,
along with historical climate data, are being used to assess soil quality
by deriving a Soil Conditioning Index value for each NRI sample site. This
index quantifies the effects of cropping sequences, tillage, and other
management inputs on soil organic matter content, which serves as an
indicator of soil quality.

According to NRCS, it has developed baselines for its long-term,
outcome-based performance measures and plans to assess and report on them
once computer models and other data collection methods that estimate
environmental change are completed. The Director of the NRCS Strategic
Planning and Performance Division said NRCS expects to assess and report
on the status of all measures by 2010 but will be able to assess the
results of some measures, such as improved soil condition on working land,
sooner. In the meantime, the agency will continue to utilize its existing
annual measures to assess performance. The Director of NRCS's Strategic
Planning and Performance Division acknowledged that the long-term measures
were not as comprehensive as needed but represented measures NRCS could
reasonably assess using modeling and data collection methods that would
soon become available. NRCS plans to continue to improve its performance
measures going forward.

Although we did not assess the comprehensiveness of the EQIP performance
measures, the additional information they provide about the results of
EQIP outcomes should allow NRCS to better gauge program performance. Such
information could also help the agency refine its process for allocating
funds to the states via its financial assistance formula by directing
funds toward practices that address unrealized performance measures and
areas of the country that need the most improvement. The Chief of NRCS's
Environmental Improvement Programs Branch agreed that information about
program performance might eventually be linked back to the EQIP funding
allocation process. However, the agency does not yet have plans to do so.

Conclusions

As a key NRCS conservation program with over $1 billion in annual funding,
EQIP was designed to help producers mitigate the potentially negative
environmental impacts of agricultural production. However, the program may
not be fully optimizing the environmental benefits resulting from
practices installed using EQIP dollars because of weaknesses in NRCS's
process for allocating funds to the states. Moreover, outdated and
duplicate formula data sources may further compromise EQIP's effectiveness
in allocating funds. Currently, it is not clear that factors, weights, and
data sources in the general financial assistance formula help the agency
direct funding to the areas of the nation with the greatest environmental
threats arising from agricultural production. NRCS has an opportunity to
address this issue as it moves forward on its plans to reexamine its
conservation funding formulas. Furthermore, the agency may be able to use
information gathered from the results of its outcome-based performance
measures to refine the financial assistance formula, making it easier for
NRCS to direct EQIP funds at the most pressing environmental problems
related to agriculture production.

Recommendations for Executive Action

To achieve EQIP's purpose of optimizing environmental benefits, we
recommend that the Secretary of Agriculture direct the Chief of the
Natural Resources Conservation Service to take the following two actions:

o ensure that the rationale for the factors and weights used in the
general financial assistance formula are documented and linked to program
priorities, and data sources used in the formula are accurate and current;
and

o continue to analyze current and newly developed long-term performance
measures for the EQIP program and use this information to make any further
revisions to the financial assistance formula to ensure funds are directed
to areas of highest priority.

Agency Comments and Our Evaluation

We provided USDA with a draft of this report for review and comment. USDA
agreed that the EQIP allocation formula needs review. USDA did not agree
with our assessment that NRCS's funding process lacks a clear link to the
program's purpose of optimizing environmental benefits. The agency stated
that its use of factors related to the natural resource base and condition
of those resources shows the general financial assistance formula is tied
to the program's purpose of optimizing environmental benefits. USDA stated
that, while some formula data sources and weights will be updated, the
types of factors used would be needed in any process that attempts to
inventory and optimize environmental benefits. While this may in fact be
the case, USDA needs to document this connection-that is, why factors were
chosen and weights assigned. USDA could make the connection between the
formula and the program's purpose of optimizing environmental benefits
more evident if it provided additional information describing its reasons
for including or excluding factors in the formula and its rationale for
assigning and modifying weights.

Appendix VII presents USDA's comments.

We are sending copies of this report to interested congressional
committees, the Secretary of Agriculture, the Director of the Office of
Management and Budget, and other interested parties. We also 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 or your staff have any questions about this report, please contact
me at (202) 512-3841 or [email protected] . Contact points for our Offices
of Congressional Relations and of Public Affairs may be found on the last
page of this report. GAO staff who made major contributions to this report
are listed in appendix VIII.

Sincerely yours,

Daniel Bertoni Acting Director, Natural Resources     and Environment

Appendix I

Objectives, Scope, and Methodology

At the request of the Ranking Democratic Member, Senate Committee on
Agriculture, Nutrition, and Forestry, we reviewed the extent to which (1)
the U.S. Department of Agriculture's (USDA) process for allocating
Environmental Quality Incentives Program (EQIP) funds to states is
consistent with the program's purpose of optimizing environmental benefits
and (2) USDA has developed measures to monitor program performance.

To review the Natural Resources Conservation Service's (NRCS) process for
allocating EQIP funding to the states, we examined EQIP funding documents
and spoke with NRCS officials from the Financial Assistance Program
Division, Budget Planning and Analysis Division, and Financial Management
Division. Our analysis considered each of the different categories of EQIP
funding, including EQIP general financial assistance, EQIP technical
assistance, regional equity funds, performance bonuses, Conservation
Innovation Grants, Colorado Salinity funds, Ground and Surface Water
Conservation funds, and Klamath Basin funds. We gathered comments from
stakeholders about the strengths and weaknesses of NRCS's EQIP funding
approach. We selected stakeholders from environmental and farm
organizations to get a broad set of views on the effectiveness of the
formula in allocating funds. Specifically, we spoke with representatives
from environmental organizations, including Environmental Defense, the
National Association of Conservation Districts, the Soil and Water
Conservation Society, and the Sustainable Agriculture Coalition, as well
as farm organizations, including the American Farm Bureau and the National
Pork Producers Council. We also discussed the EQIP funding allocation
process with selected participants on state technical committees-the Iowa
Department of Natural Resources, Iowa Farm Bureau, and Nebraska Department
of Environmental Quality; academic stakeholders; and former NRCS employees
who participated in the development of the original formula.1 We examined
the factors and weights in the financial assistance formula and discussed
their purpose with EQIP program officials. We performed statistical
analysis of the financial assistance formula to determine what impact the
different factors had on overall funding. A discussion of the analysis we
performed can be found in appendix III. We searched for information about
the source of data for each factor in the formula in order to formulate an
understanding of what each factor measured and verify the accuracy of the
data being used by NRCS. NRCS did not retain documentation of the source
data for 10 factors and, as a result, we were unable to verify all data
used in the financial assistance formula. To estimate the number of
factors using outdated data, we searched for more updated versions of the
same data sources NRCS said it used in its formula. We did not include
more updated, but different, sources of data in our count.

To understand Congress's and NRCS's goals for EQIP, we reviewed the
Federal Agriculture Improvement and Reform Act of 1996, Farm Security and
Rural Investment Act of 2002, associated regulations, and related
appropriations laws. We reviewed program documentation describing the
purpose and priorities of EQIP and discussed the documentation with EQIP
officials. To understand agency conservation priorities, we analyzed a
2005 database of conservation practices funded using EQIP, Ground and
Surface Water Conservation, and Klamath Basin funds.

To determine how the factors and weights in the formula aligned with
resource concerns across the nation, we conducted research on the impact
agricultural production has on the environment. We spoke with NRCS
officials from selected states-Iowa, Maryland, Mississippi, Missouri,
Montana, Nebraska, New Mexico, Rhode Island, and Texas-to better
understand resource concerns important to their state and how they used
funds received from headquarters to address those concerns. We also spoke
with officials from three county offices within these states. This
geographically diverse group included states that received varying amounts
of EQIP funding and engaged in a range of types of agricultural
production.

To review what measures are in place to monitor EQIP program performance,
we spoke with representatives from the NRCS teams responsible for
strategic planning and oversight activities-the Operations Management and
Oversight Division, Oversight and Evaluation staff, and Strategic and
Performance Planning Division-and representatives from the Financial
Assistance Program Division. We examined agency strategic planning and
performance documents. We reviewed documentation of agency and EQIP goals
and performance measures and reviewed the Web-based NRCS Performance
Results System.2 We also spoke with representatives from NRCS and
nongovernmental organizations working on the Conservation Effects
Assessment Project and reviewed related documentation to determine how
that initiative might influence the development of future EQIP goals. Our
analysis did not include an independent verification of NRCS's compliance
with internal controls.3

We performed our work between December 2005 and August 2006 in accordance
with generally accepted government auditing standards.

Appendix II

EQIP 2006 Funding Allocation Formulas

Tables 5, 6, 7, and 8, respectively, describe the formulas for allocating
general financial assistance, Ground and Surface Water Conservation funds,
performance bonuses, and Klamath Basin funds. In the case of the general
financial assistance formula, we have identified the source of data for
each factor and described what each factor measures.

Table 5: Factors, Data Sources, and Weights in the EQIP General Financial
Assistance Formula for Allocating Funding to the States in Fiscal Year
2006

                                        

          Factor               Source              Description         Weight 
Acres of nonirrigated 1997 National      Nonirrigated cultivated      3.2% 
cropland              Resources          and noncultivated cropland 
                         Inventory (Revised acres                      
                         December 2000)                                
Acres of irrigated    1997 National      Irrigated cultivated and      4.3 
cropland              Resources          noncultivated cropland     
                         Inventory (Revised acres                      
                         December 2000)                                
Acres of federal      a                  b                             0.5 
grazing lands                                                       
Acres of nonfederal   1997 National      Nonfederal, rural acres of    4.3 
grazing lands         Resources          pastureland, rangeland,    
                         Inventory (Revised and grazed forestland      
                         December 2000)                                
Acres of forestlands  1997 National      Nonfederal, rural acres of    1.1 
                         Resources          forestland                 
                         Inventory (Revised                            
                         December 2000)                                
Acres of specialty    1997 National      Acres of land used as         3.2 
cropland              Resources          vineyards or to grow       
                         Inventory (Revised fruits, nuts, berries,     
                         December 2000)     bush fruit, or other       
                                            specialty crops            
Acres of wetlands and 1997 National      Acres of wetlands and         4.6 
at-risk species       Resources          deepwater habitats on      
habitat               Inventory (Revised water areas and nonfederal 
                         December 2000)     land                       
Acres of bodies of    1997 National      Surface area (in acres) of    3.2 
water                 Resources          water areas                
                         Inventory (Revised                            
                         December 2000)                                
Livestock animal      1997 NRCS          b                             5.8 
units                 calculation based                             
                         on data gathered                              
                         prior to 1997                                 
                         (exact year                                   
                         unknown)c                                     
Animal waste          NRCS calculation   b                             5.8 
generation            based on 1987                                 
                         Census of                                     
                         Agriculture and                               
                         other datac                                   
Waste management      1996 Environmental Modeled estimates of state    3.5 
capital cost          Protection Agency  needs for controlling      
                         Clean Water Needs  nonpoint source pollution  
                         Survey Report to   from confined animal       
                         Congress           facilities with fewer than 
                                            1,000 animal units         
Acres of American     1997 Bureau of     Acres of American Indian      3.3 
Indian tribal lands   Indian Affairs     reservations and Tribal    
                         data               Trust Land                 
Number of limited     1997 Census of     Number of farms with sales    5.0 
resource producers    Agriculture        under $100,000             
Acres of grazing land 1997 National      Acres of grazing and          0.8 
lost to conversion    Resources          pastureland converted to   
                         Inventory (Revised another form of land or    
                         December 2000)     development between 1982   
                                            and 1997                   
Air quality           NRCS analysis of   Measure of air quality        1.4 
nonattainment areas   2005 Environmental nonattainment based on the 
                         Protection Agency  percent of a state         
                         air quality data   affected by certain air    
                                            quality pollutants and the 
                                            number of air quality      
                                            standards not met by that  
                                            state                      
Acres of pastureland  1992 National      Acres of pastureland          5.5 
needing treatment     Resources          needing conservation       
                         Inventory          treatment                  
Acres of cropland     1992 National      d                             6.2 
eroding above T       Resources                                     
                         Inventoryc                                    
Acres of fair and     1987 National      Acres of rangeland in fair    6.2 
poor rangeland        Resources          and poor condition         
                         Inventory                                     
Acres of forestlands  1987 National      f                             1.4 
eroding above T       Resources                                     
                         Inventorye                                    
Acres of cropland and 1997 National      Acres of cultivated and       2.6 
pastureland soils     Resources          noncultivated cropland and 
affected by saline    Inventory (Revised pastureland with the       
and/or sodic          December 2000)     presence of salts          
conditions                                                          
Miles of impaired     Environmental      b                             3.6 
rivers and streams    Protection Agency                             
                         1994 National                                 
                         Water Quality                                 
                         Inventoryc                                    
Potential for         1997 NRCS          NRCS formula based on data    1.3 
pesticide and         analysisg          about land vulnerability   
nitrogen leaching                        to manure nitrogen,        
                                            commercial nitrogen, and   
                                            pesticide leaching         
Potential for         1997 NRCS          NRCS formula based on data    1.7 
pesticide and         analysisg          about land vulnerability   
nitrogen runoff                          to manure nitrogen,        
                                            commercial nitrogen, and   
                                            pesticide runoff           
Ratio of livestock    a                  b                             1.7 
animal units to                                                     
cropland                                                            
Number of             2003 NRCS report   Number of farms needing a     2.8 
concentrated animal   based on 1997      comprehensive nutrient     
feeding               Census of          management plan            
operations/animal     Agriculture datah                             
feeding operations                                                  
Ratio of commercial   NRCS calculation   b                             0.9 
fertilizers to        based on 1995 data                            
cropland              from the                                      
                         Association of                                
                         American Plant                                
                         Food Control                                  
                         Officials and 1997                            
                         NRI cropland datac                            
Wind erosion above T  1997 National      Cultivated and                4.2 
                         Resources          noncultivated cropland     
                         Inventory (Revised with a 4-year average rate 
                         December 2000)     of estimated soil loss due 
                                            to wind erosion greater    
                                            than T-a tolerable rate of 
                                            erosion above which soil   
                                            productivity is believed   
                                            to decrease                
Phosphorous runoff    1997 NRCS          Same data used for factor     3.9 
potential             analysisg          measuring potential for    
                                            pesticide and nitrogen     
                                            runoff                     
Riparian areas        1982 National      i                             0.8 
                         Resources                                     
                         Inventoryc                                    
Carbon sequestration  1997 National      Sum of data from other        3.6 
                         Resources          factors in the financial   
                         Inventory (Revised assistance                 
                         December 2000) and formula-nonirrigated       
                         unknown data       cropland, federal grazing  
                         source             lands, nonfederal grazing  
                                            lands and forestlands      
Coastal zone land     NRCS calculation   j                            3.6% 
                         based on 1992                                 
                         National Oceanic                              
                         and Atmospheric                               
                         Administration and                            
                         unknown datac                                 

Sources: GAO analysis of NRCS and USDA data.

Note: We used NRCS's own terminology for the factor names in this chart.
In some instances, names do not precisely capture what is being measured.

aWe were unable to verify the source of data for this factor.

bBecause we could not verify certain data sources, we were unable to
provide an accurate description of what each factor measured. Blank cells
indicate that we were unable to accurately describe what the factor
measured.

cData source as reported by NRCS. We were unable to verify the source of
data for this factor.

dAccording to an NRI official, cropland eroding above T could have been
estimated in one of two ways-(1) acres of cropland where the total wind,
sheet and rill erosion rates exceeded T or (2) acres of cropland where
either wind erosion, sheet and rill erosion, or both, exceeded T. We were
not able to confirm how the data was estimated.

eNRCS could not confirm the source or date of this data. The National
Resources Inventory believed this data was from work NRI performed in
1987.

fAccording to an NRI official, this factor measures acres of nonfederal,
rural forestland with estimated average annual sheet and rill erosion
above T. We were not able to obtain documentation to confirm this
definition.

g"Potential Priority Watersheds for Protection of Water Quality from
Nonpoint Sources Related to Agriculture." Poster Presentation at the 52nd
Annual SWCS Conference Toronto, Ontario, Canada, July 22-25, 1997 (Revised
October 7, 1997).

hCosts Associated with Development and Implementation of Comprehensive
Nutrient Management Plans Part I-Nutrient Management, Land Treatment,
Manure and Wastewater Handling and Storage, and Recordkeeping (NRCS, June
2003).

iAccording to NRCS, the definition for riparian areas in the 1982 National
Resources Inventory was acres of riparian areas-the banks, shorelines, or
edges of the rising ground bordering a natural or manmade watercourse or
water area (riparian areas are not limited to natural areas).

jAccording to NRCS, this factor considers data on square miles of
coastlines.

Table 6: Fiscal Year 2006 Formula for Allocating Ground and Surface Water
Conservation Financial Assistance

Targeted area                           Allocation methodology      Weight 
High Plains Aquifer states-Colorado,    Percentage of state's        40.6% 
Kansas, Nebraska, New Mexico, Oklahoma, acreage in the High Plains  
South Dakota, Texas, and Wyoming        Aquifer                     
Western drought states-Arizona,         Amount of irrigated acreage  41.5% 
California, Idaho, Montana, Nevada,     in each state               
North Dakota, Oregon, Utah, and                                     
Washington                                                          
Additional states with agricultural     Proportional comparison of   17.9% 
water needs-Alabama, Arkansas,          agriculture to              
Delaware, Florida, Georgia, Hawaii,     nonagricultural use of      
Iowa, Louisiana, Maine, Minnesota,      water                       
Mississippi, Missouri, North Carolina,                              
Puerto Rico, Wisconsin                                              

Source: NRCS.

Table 7: Factors Used in the Fiscal Year 2006 Formula for Allocating EQIP
Performance Bonuses

Factor                      Description                             Weight 
Ratio of technical          Ratio of obligated EQIP funds used for     25% 
assistance obligations to   technical assistance in fiscal year     
total obligations           2005 to total obligated funds           
Livestock-related contracts Ratio between the number of EQIP            15 
                               contracts issued for Comprehensive      
                               Nutrient Management Plans to the number 
                               of farms needing such plansa            
Cost-share obligations      Ratio of cost-share dollars obligated       15 
versus payments             to cost-share dollars paid in fiscal    
                               years 2004 and 2005                     
Technical service provider  Ratio of disbursements to obligations       15 
obligations and             in fiscal years 2004 and 2005 to        
disbursements               technical service providers-contractors 
                               that help producers install practices   
Weighted cost-share         Average cost-share rate by state,           10 
percentage                  excluding limited resource farmer       
                               cost-share and incentive payments       
Limited resource farmer     Percentage of total EQIP contracts          10 
                               entered into with limited resource      
                               farmers                                 
Program national priorities Ratio between acres treated with           10% 
                               conservation practices that address the 
                               national priorities to the total        
                               agricultural base                       

Source: NRCS.

aComprehensive nutrient management plans are conservation plans unique to
livestock operations. These plans document practices and strategies
adopted by livestock operations to address natural resource concerns
related to manure and organic by-products and their potential impacts on
water quality.

Table 8: Fiscal Year 2006 Formula for Allocating Klamath Basin Program
Financial Assistance

State                                                               Weight 
California                                                             50% 
Oregon                                                                 50% 

Source: NRCS.

Appendix III

Statistical Techniques to Determine Influential Factors in the 2006 EQIP
Financial Allocation Formula

Using statistical techniques-that is, principal components regression and
factor analysis-we analyzed the Environmental Quality Incentives Program
(EQIP) formula used to allocate fiscal year 2006 financial assistance to
the states to identify the environmental factors that most influenced the
allocations. Sixty-five percent of the total EQIP funds for 2006 were
based on the allocation formula for financial assistance.

Principal Components Regression

In order to determine the relationships between the allocation and the
environmental factors (variables), we typically would apply regression
techniques to a model, expressed as

(1) (i = 1,..., 48)

In equation (1), the dependent variable is the funding allocation for
state i, the x's are the j factors in the allocation formula, b0,
b1,...,bj  are the regression coefficients, and ei is the model error for
the ith state.

When we used this model, however, standard regression techniques were not
possible because many of the environmental factors used in the allocation
formula were highly collinear.1 Collinearity occurs when variables are so
highly correlated that it is difficult to distinguish their independent
influences on the dependent variable-in this case, state allocation
funding. In a regression analysis, highly correlated independent variables
cause the following effects: (1) regression coefficients change, depending
on which variables are included or excluded in the model, (2) standard
errors are large, (3) regression coefficients are large with random signs,
and (4) achieving statistical significance of the collinear parameters is
difficult. Moreover, multicollinearity poses a problem if the purpose of
the model is to estimate, or explain, rather than predict, the individual
contributions of variables. Following Fekedulegn et al., (2002), Norton

(1984), and others, we used principal components regression analysis since
this technique is recommended when there is multicollinearity in the
data.2

Before running the regression analysis, we performed the principal
components analysis.3 This procedure generates a set of latent variables,
called principal components-uncorrelated linear transformations of the
original variables.4 At this stage, even though the new variables are not
collinear, the same magnitude of variance is retained. Therefore, the
elimination of small principal components reduces the total variance and
substantially improves the diagnostic capability of the model. In order to
eliminate these small principal components, various selection procedures
are used. Following Fekedulegn (2002), we chose the cumulative eigenvalue
product rule, which keeps the first principal components whose combined
product is greater than 1.00 (Guiot et al., 1982).5 The principal
components themselves are expressed as

(2) Z = X*V.

In equation (2), Z is an (i x j) matrix of principal components, X is an
(i x j) matrix of standardized environmental factors, and V is a (j x j)
matrix of eigenvectors.76,

After the principal components analysis and the elimination of smaller
principal components as described above, we used the data in a
cross-sectional multivariate regression expressed as

(3) .

In equation (3), A is an (i x 1) vector for the allocation of funding for
the states (the dependent variable in the regression), b01 is an (i x 1)
vector of the intercept terms, Z is an (i x j) matrix of principal
components, and a  is a (j x 1) vector of new coefficients of the
principal components. However, this procedure will usually leave some
principal components that are not statistically significant. Therefore, to
further eliminate the nonsignificant principal components, we used the SAS
stepwise regression procedure.8 Specifically, we eliminated "r" principal
components in the analysis, which consisted of the (1) number eliminated
using the eigenvalue product rule and (2) number eliminated from the
stepwise regression. We were then left with (j - r) principal components
estimators or coefficients and the reduced form in equation 3 becomes

(4) .

In equation (4), a is the vector of coefficients associated with the
reduced set of (j-r) principal components and Z is an (i x (j-r)) matrix
of principal components. With the r components eliminated, the principal
components estimators-in terms of the standardized environmental factors
of the allocation model-are obtained by multiplying the new vector of
coefficients by the associated vectors in the matrix of eigenvectors:

(5) bspc= Vj x (j-r) a(j-r) x 1 .

In equation (5), bspc (subscript pc stands for principal components) is
the vector of j standardized principal component estimators of the
regression coefficients of the environmental factors, V is the (j x ( j -
r)) matrix of eigenvectors, and a  is the reduced vector of ((j - r) x 1)
estimated coefficients as in equation 4. Once we have the standardized
coefficients of the principal components estimators of the factors, we can
transform them back into the coefficients of the original environmental
factors. For the standardized estimators, the method for this
transformation is expressed as

(6)

In equation (6), Sxj is the standard deviation of the original jth
environmental factor, xj, bsj,pc is the jth standardized estimator, and
bj,pc is the coefficient of the original environmental factor.

While we can obtain the regression coefficients of the original
environmental factors (the bj,pc's) that have been corrected for
multicollinearity, we cannot directly compare them because most have
different units. For instance, some environmental and resource factors
used in the formula are measured in acres, while others may be measured in
terms of animal units. In other words, the largest coefficient may not be
the most influential in the regression. Therefore, when comparing the
relative importance of the factors (variables) in the regression, we
mainly

discuss the standardized estimators of the environmental factors used in
the allocation formula.9

Data Used

For the 48 contiguous states, we used a cross-section of data for the
dependent variable-the allocation variable-and the independent
variables-the environmental variables (factors). We could not incorporate
Alaska or Hawaii because we lacked complete data. We excluded two
factors-independent variables-from the regression analysis because they
were linear combinations of factors already included in the data. For
instance, we could not include the carbon sequestration factor because it
is the sum of four factors already included in the formula allocation
model: acres of nonirrigated cropland, forestland, federal grazing land,
and nonfederal grazing land. We also excluded the factor for pesticide and
nitrogen runoff because it contains the same data as the phosphorous
runoff potential factor. Although the U.S. Department of Agriculture
(USDA) weights these factors differently, they are still linear
combinations and, for regression analysis, must be excluded. In all, we
ran the regression using the 2006 state allocations for the 48 states as
our dependent variable and the 29 environmental and resource factors in
the formula for our independent variables.

Results

After reducing the components from the eigenvalues product rule and the
stepwise regression, we were left with 13 principal components from the
original 29. We then transformed the parameter estimates of the stepwise
regression,, back into the coefficients of the standardized principal
components of the environmental factors, the bspc. The results for these
standardized coefficients-bspc, the t-values, and the probability values
of t-sorted by the size of the standardized coefficient are shown in table
9. Specifically, a standardized coefficient of a factor measures the
expected change in the dependent variable for a one unit change in the
standardized independent variable, in this case the ith factor, all other
things being equal. Those variables that had the largest standardized
coefficient as well as being highly statistically significant were acres
of fair and poor rangeland, acres of nonfederal grazing land, acres of
irrigated cropland, acres of American Indian tribal lands, wind erosion
above T, and livestock animal units. As table 9 shows, as one would expect
with a formula, most of the factors in the regression were highly
significant and positively related to the allocation, except the four
factors, acres of forestlands, potential for pesticide and nitrogen
leaching, air quality nonattainment areas, and acres of federal grazing
lands.

Table 9: Standardized Principal Components Estimators of the Original
Variables and Statistical Significance

                                        

                      Factor                    Standardized  t-value p-value 
                                                 coefficient          
Acres of fair and poor rangeland                  1399095   35.322 <0.0001 
Acres of nonfederal grazing lands                 1389052  35.7784 <0.0001 
Acres of irrigated cropland                       1372591   13.448 <0.0001 
Acres of American Indian tribal lands             1313695  13.2325 <0.0001 
Wind erosion above T                              1210688  18.4507 <0.0001 
Livestock animal units                            1197842  37.3831 <0.0001 
Riparian areas                                     935709  12.5955 <0.0001 
Number of limited resource producers               776918  17.9933 <0.0001 
Acres of cropland eroding above T                  748625  11.1347 <0.0001 
Acres of bodies of water                           699697  8.91583 <0.0001 
Acres of cropland and pastureland soils            654109  6.14821 <0.0001 
affected by saline and/or sodic conditions                         
Acres of specialty cropland                        648891  10.1944 <0.0001 
Acres of pastureland needing treatment             625264  11.5436 <0.0001 
Animal waste generation                            537929  6.19323 <0.0001 
Acres of wetlands and at-risk species              528679  10.9819 <0.0001 
habitat                                                            
Waste management capital cost                      504716  7.58475 <0.0001 
Coastal zone land                                  501769   9.1963 <0.0001 
Acres of grazing land lost to conversion           449376  6.32035 <0.0001 
Miles of impaired rivers and streams               446096  3.99622  0.0008 
Ratio of commercial fertilizers to cropland        409840  4.10749  0.0007 
Acres of nonirrigated cropland                     403724  7.66134 <0.0001 
Acres of forestlands eroding above T               393498  5.97645 <0.0001 
Phosphorous runoff potential                       306870  5.30147 <0.0001 
Number of concentrated animal feeding              251359  4.95367  0.0001 
operations/animal feeding operations                               
Ratio of livestock animal units to cropland        213299  2.08889  0.0512 
Acres of forestlands                                89181  1.19149  0.2489 
Potential for pesticide and nitrogen                45721     0.77  0.4513 
leaching                                                           
Air quality nonattainment areas                    -33022 -0.64754  0.5255 
Acres of federal grazing lands                    -280851 -2.53757  0.0206 

Source: GAO analysis of USDA data.

Note: All variables above the bolded line are positive and statistically
significant at the 0.05 percent level. Our analysis does not include an
adjustment made to the variable "acres of American Indian tribal lands"
affecting two states. We do not expect that this adjustment would have a
material affect on the results.

Factor Analysis of EQIP Environmental Variables

We used the factor analysis technique to reduce the original set of
variables (environmental factors) in the EQIP formula to a smaller set of
underlying factors that actually drive the variables and the relationships
among these variables.10 Factor analysis has been used previously by
researchers to identify, group, and interpret various environmental
concerns, such as soil quality, that cannot be measured directly, but must
be inferred by measuring other attributes that serve as indicators.11 For
this formula, the underlying factors should mimic, in some sense, the
underlying environmental concerns, such as water quality and quantity,
soil productivity, and wildlife habitat preservation.

Explanation of the Technique

Factor analysis is a technique used to explain the correlations between
variables and to derive a new set of underlying variables, called
"factors," that give a better understanding of the data being analyzed.
Using this technique allows us to determine what smaller number of factors
accounts for the correlation in the larger set of variables in the
formula.

In factor analysis, each observed variable, x, can be expressed as a
weighted composite of a set of underlying, latent variables (f's) such
that

(7) .

In equation (7), the correlation between the observed variables, the x's,
can be explained in terms of the underlying (latent) factors. These latent
factors explain the common variance between the variables. For example,
given a set of observed variables, factor analysis forms a set of factors
that are as independent from each other as possible, while the observed
variables within each factor are as highly correlated as possible.

To perform the factor analysis, we used the SAS PROC FACTOR procedure,
choosing the principal factors method to extract the factors. One part of
the analysis was to determine the number of factors to extract.
Hypothetically, there can be one factor for every variable, but the goal
is to reduce this number to a subset of factors that drive, or control,
the values of the variables being measured. We postulated that the
underlying factors should mimic, in some sense, the underlying
environmental concerns, such as water quality and quantity, soil
productivity, and wildlife habitat. However, since the data contain
certain variables such as acres of nonirrigated cropland, acres of
nonfederal grazing land, or acres of American Indian tribal lands, the
latent factors may be different in character. To determine the number of
factors, there are several computational methods and more subjective
methods such as ease of interpretability of factors. We used both the ease
of interpretability of the factors, as well as the "scree test."12 As is
typically done to achieve a more meaningful and interpretable solution, we
applied a rotation technique to the initial factor pattern matrix.13

Results

We used the rotated factor pattern matrix to interpret the meaning of the
latent factors, which we identified through their correlations with the
environmental factors (variables), as shown in table 10. The factor
loadings that have an absolute value equal to or greater than 0.4 are
shaded, and several variables are significantly correlated with more than
one factor-called a "split loading."14

Table 10: Rotated Factor Pattern Matrix

                                     Latent factor 
    Factor or variable from formula              1        2        3        4 
Acres of nonfederal grazing land        0.95432 -0.14193  0.00329  0.10117 
Acres of fair and poor rangeland        0.89044 -0.18858  0.04876 -0.03961 
Livestock animal units                  0.77845  0.51746  0.07331  0.19868 
Wind erosion above T                    0.76534  0.12833  0.11974 -0.29702 
Acres of cropland eroding above T       0.69604    0.613 -0.01102 -0.21171 
Riparian areas                          0.68226  0.16742  0.37957  0.23914 
Acres of grazing land lost to           0.64431  0.26539  -0.0365  0.11935 
conversion                                                        
Acres of irrigated cropland             0.61829   0.0418  0.14401   0.4548 
Acres of pastureland needing            0.60112  0.28467  0.24348  0.11009 
treatment                                                         
Number of limited resource              0.59529  0.56896  0.23444  0.20013 
producers                                                         
Acres of federal grazing land           0.46349  -0.4339 -0.32474  0.01709 
Acres of cropland and pastureland       0.39952 -0.17105 -0.18815  0.39389 
soils affected by saline and/or                                   
sodic conditions                                                  
Number of concentrated animal           0.05818  0.87513 -0.03339  -0.0465 
feeding operations/animal feeding                                 
operations                                                        
Phosphorous runoff potential            0.12202  0.86976  0.28758  0.13109 
Waste management capital cost            0.1018  0.76694  0.04769  0.07004 
Acres of nonirrigated cropland          0.46277  0.71639 -0.15479 -0.23632 
Potential for pesticide and             0.01166  0.69366  0.46219  0.29581 
nitrogen leaching                                                 
Animal waste generation                  0.2898  0.57515  0.26865  0.19698 
Acres of American Indian tribal         0.24861 -0.37768 -0.26536  0.01991 
lands                                                             
Ratio of livestock animal units         0.08217 -0.47811 -0.05783  0.29948 
to cropland                                                       
Acres of wetlands and at-risk           0.14074  0.18228  0.86993  -0.1914 
species habitat                                                   
Acres of bodies of water                0.35908  0.10564  0.81371 -0.01889 
Coastal zone land                        0.1042 -0.05594  0.79583  0.29895 
Acres of forestlands                    0.13012  0.24312  0.69724  0.20541 
Acres of specialty cropland             0.09679 -0.10928  0.30919  0.78195 
Acres of forestlands eroding            0.15413 -0.00078 -0.04538  0.77575 
above T                                                           
Miles of impaired rivers and            0.20763  0.12325   0.1149  0.60384 
streams                                                           
Ratio of commercial fertilizer to      -0.15704  0.21558  0.05047  0.52823 
cropland                                                          
Air quality nonattainment areas        -0.19264  -0.2472 -0.00932  0.32038 

Source: GAO analysis of USDA data.

Note: Variable loadings with a significant correlation are shaded.

The factor analysis technique also calculates the amount of common
variance explained by each latent factor. For these data, the variances
are: factor 1- 6.44, factor 2-5.49, factor 3-3.56, and factor 4-3.00,
accounting for about 71 percent of the common variance in the data.

Overall, the four factors (1) all relate to environmental concerns, as
well as agricultural resources, and (2) each latent factor contributes a
decreasing amount of common variance to the total variation among all of
the variables. We interpreted the EQIP data that went into the factor
analysis to represent (1) dryland agriculture and cattle feeding, (2)
water quality concerns relating to concentrated livestock feeding
operations and nonirrigated cropland, (3) wildlife habitat preservation,
and (4) specialty crops/intensive agriculture and water quality/quantity
concerns. Specifics of the factor analysis follow:

Factor 1: This factor contributes the most variation to the factor
analysis and seems to be associated with dryland agriculture and cattle
grazing and feeding. The variables-acres of nonfederal grazing lands,
acres of fair and poor rangeland, wind erosion above T, acres of cropland
eroding above T, and acres of irrigated cropland-are all descriptors of
this type of agriculture. In addition, factor 1 is also strongly
correlated with the livestock animal units variable, although it has a
split loading with factor 2. While the number of limited resource
producers variable has a split loading between this factor and factor 2,
it is most heavily loaded with this factor.

Factor 2: This factor, like factor 1, has to do with livestock operations,
as well as with other important livestock-related variables that affect
water quality. Here, the highest loading is with the variable, number of
concentrated animal feeding operations/animal feeding operations, (CAFOs)
(0.88), although it has the split loading with livestock animal units
(0.52). In addition, factor 2 showed high loadings for phosphorous runoff
potential and potential for pesticide and nitrogen leaching, which may be
related to sediment losses from both animal and cropland agriculture.
Moreover, as cropland and CAFOs are usually in the same location, one
would expect the variable for acres of nonirrigated cropland to also have
a high loading, which it does (0.72).

Factor 3: This factor seems to be related to environmental concerns about
wildlife habitat, with the highest loading going to acres of wetland and
at-risk species habitat (0.87), as well as to acres of bodies of water,
(0.81) coastal zone land (0.80) and acres of forestlands (0.70). Potential
for pesticide and nitrogen leaching (0.46) showed a split loading with
factor 2.

Factor 4: This factor seems to represent variables relating to specialty
crop and intensive agriculture, with high loadings for acres of specialty
crops, ratio of commercial fertilizer to cropland, and acres of irrigated
cropland, (which had a split loading with factor 1). Also, acres of
cropland and pastureland affected by saline and/or sodic conditions, a
soil condition that often accompanies irrigated soils, is almost
significantly correlated to Factor 4 (0.39). This factor also highly loads
with miles of impaired rivers and streams, which may be an indication of
water quality and quantity concerns associated with soils that require
irrigation. Factor 4 is also highly associated with acres of forestlands
eroding above T, many of which are found in the same areas that contain
acres of irrigated cropland.

The two variables-air quality nonattainment areas and acres of American
Indian tribal lands-did not load onto any of the latent factors. When this
happens, the variable has a unique variance that is not explained by the
common factors.

Appendix IV

Initial EQIP Funding Provided to the States, Fiscal Year 2006
Source: GAO analysis of NRCS data.

Note: Dollars allocated at the national level to producers through
Conservation Innovation Grants are not included.

aNRCS provides these funds to the states through both financial and
technical assistance, the majority of which are financial assistance.

bThe source for data on total EQIP funding, except for Mississippi, is
NRCS at h
ttp://www.nrcs.usda.gov/PROGRAMS/2006_allocations/2006Allocationstostatesbyprog/FY2006program_allocations_by_states.html
. Due to rounding, totals may not equal the sum of funding from all
categories.

cMississippi's funding total is approximately $189,000 more than what was
reported by NRCS. The $189,000 represents a payment transfer made from the
Mississippi state office to headquarters for training. In order to
consistently represent the initial amount of funding each NRCS state
office received from headquarters, we included this $189,000 in
Mississippi's funding total.

dTotals for Puerto Rico also include funding provided to the U.S. Virgin
Islands.

eTotal funding may not equal the sum of state funding due to rounding.

Appendix V

Historical EQIP Funding Levels, Fiscal Years 2001-2006

Source: GAO analysis of NRCS data.

aThe data source for fiscal year 2006 total EQIP funding, except for
Mississippi, was NRCS's Web site:
http://www.nrcs.usda.gov/PROGRAMS/2006_allocations/2006Allocationstostatesbyprog/FY2006program_allocations_by_states.html
.

bMississippi's funding total for 2006 is approximately $189,000 more than
what was reported by NRCS. The $189,000 represents a payment transfer made
from the Mississippi state office to headquarters for training.

cTotals for Puerto Rico also include funding provided to the U.S. Virgin
Islands.

Appendix VI

Fiscal Year 2005 EQIP Obligations by Conservation Practice

Source: GAO analysis of NRCS data.

Note: This table only provides data on financial assistance obligations.
It does not contain data on technical assistance obligations. The data
used were generated in March 2006 and represent obligations as of that
date. NRCS said the database from which these data were generated is
continually modified as contracts are altered or cancelled.

aIn fiscal year 2005, NRCS entered into 49,406 contracts for the EQIP
program. Each contract included one or more practices. This column
represents the total number of practices for which EQIP, Ground and
Surface Water Conservation, and Klamath Basin funds were obligated.

bThis represents an interim state practice, rather than a national
approved practice. Interim state practices are tested by NRCS for 2 years,
after which they are approved for national use, extended for further
testing, added to an existing state standard, or cancelled.

cTotals may not add due to rounding.

Appendix VII

Comments from the U. S. Department of Agriculture

Appendix VIII

GAO Contact and Staff Acknowledgments

GAO Contact

Daniel Bertoni (202) 512-3841

Staff Acknowledgments

In addition to the individual named above, Ronald E. Maxon, Jr., Assistant
Director; William Bates; Thomas Cook; Barbara El Osta; Paige Gilbreath;
Lynn Musser; Omari Norman; and Carol Herrnstadt Shulman made key
contributions to this report.

(360644)

www.gao.gov/cgi-bin/getrpt? GAO-06-969 .

To view the full product, including the scope
and methodology, click on the link above.

For more information, contact Daniel Bertoni at (202) 512-3841 or
[email protected].

Highlights of GAO-06-969 , a report to the Ranking Democratic Member,
Committee on Agriculture, Nutrition, and Forestry, U.S. Senate

September 2006

AGRICULTURAL CONSERVATION

USDA Should Improve Its Process for Allocating Funds to States for the
Environmental Quality Incentives Program

The Environmental Quality Incentives Program (EQIP) assists agricultural
producers who install conservation practices, such as planting vegetation
along streams and installing waste storage facilities, to address
impairments to water, air, and soil caused by agriculture or to conserve
water. EQIP is a voluntary program managed by the U.S. Department of
Agriculture's (USDA) Natural Resources Conservation Service (NRCS). NRCS
allocates about $1 billion in financial and technical assistance funds to
states annually. About $650 million of the funds are allocated through a
general financial assistance formula.

As requested, GAO reviewed whether USDA's process for allocating EQIP
funds to states is consistent with the program's purposes and whether USDA
has developed outcome-based measures to monitor program performance. To
address these issues, GAO, in part, examined the factors and weights in
the general financial assistance formula.

What GAO Recommends

GAO recommends, among other things, that NRCS document its rationale for
the factors and weights in its general financial assistance formula and
use current and accurate data. USDA agreed with GAO that the formula
needed review. USDA did not agree with GAO's view that NRCS's funding
process does not clearly link to EQIP's purpose of optimizing
environmental benefits. It believes that the funding process clearly links
to EQIP's purpose, but it has not documented the link.

NRCS's process for providing EQIP funds to states is not clearly linked to
the program's purpose of optimizing environmental benefits; as such, NRCS
may not be directing funds to states with the most significant
environmental concerns arising from agricultural production. To allocate
most EQIP funds, NRCS uses a general financial assistance formula that
consists of 31 factors, including such measures as acres of cropland,
miles of impaired rivers and streams, and acres of specialty cropland.
However, this formula has several weaknesses. In particular, while the 31
factors in the financial assistance formula and the weights associated
with each factor give the formula an appearance of precision, NRCS does
not have a specific, documented rationale for (1) why it included each
factor in the formula, (2) how it assigns and adjusts the weight for each
factor, and (3) how each factor contributes to accomplishing the program's
purpose of optimizing environmental benefits. Factors and weights are
important because a small adjustment can shift the amount of funding
allocated to each state on the basis of that factor and, ultimately, the
amount of money each state receives. For example, in 2006, a 1 percent
increase in the weight of any factor would have resulted in $6.5 million
more allocated on the basis of that factor and a reduction of 1 percent in
money allocated for other factors. In addition to weaknesses in
documenting the design of the formula, some data NRCS uses in the formula
to make financial decisions are questionable or outdated. For example, the
formula does not use the most recent data available for 6 of the 31
factors, including commercial fertilizers applied to cropland. As a
result, any recent changes in a state's agricultural or environmental
status are not reflected in the funding for these factors. During the
course of GAO's review, NRCS announced plans to reassess its EQIP
financial assistance formula.

NRCS recently developed a set of long-term, outcome-based performance
measures to assess changes to the environment resulting from EQIP
practices. The agency is also in the process of developing computer models
and other data collection methods that will allow it to assess these
measures. Thus, over time, NRCS should ultimately have more complete
information on which to gauge program performance and better direct EQIP
funds to areas of the country that need the most improvement.
*** End of document. ***