Long-Term Care Insurance: Partnership Programs Include Benefits  
That Protect Policyholders and Are Unlikely to Result in Medicaid
Savings (11-MAY-07, GAO-07-231).				 
                                                                 
Partnership programs allow individuals who purchase Partnership  
long-term care insurance policies to exempt at least some of	 
their personal assets from Medicaid eligibility requirements. In 
response to a congressional request, GAO examined (1) the	 
benefits and premium requirements of Partnership policies as	 
compared with those of traditional long-term care insurance	 
policies; (2) the demographics of Partnership policyholders,	 
traditional long-term care insurance policyholders, and people	 
without long-term care insurance; and (3) whether the Partnership
programs are likely to result in savings for Medicaid. To examine
benefits, premiums, and demographics, GAO used 2002 through 2005 
data from the four states with Partnership programs--California, 
Connecticut, Indiana, and New York--and other data sources. To	 
assess the likely impact on Medicaid savings, GAO (1) used data  
from surveys of Partnership policyholders to estimate how they	 
would have financed their long-term care without the Partnership 
program, (2) constructed three scenarios illustrative of the	 
options for financing long-term care to compare how long it would
take for an individual to spend his or her assets on long-term	 
care and become eligible for Medicaid, and (3) estimated the	 
likelihood that Partnership policyholders would become eligible  
for Medicaid based on their wealth and insurance benefits.	 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-07-231 					        
    ACCNO:   A69491						        
  TITLE:     Long-Term Care Insurance: Partnership Programs Include   
Benefits That Protect Policyholders and Are Unlikely to Result in
Medicaid Savings						 
     DATE:   05/11/2007 
  SUBJECT:   Comparative analysis				 
	     Financial analysis 				 
	     Health care programs				 
	     Health policy					 
	     Long-term care					 
	     Long-term care insurance				 
	     Medicaid						 
	     Population statistics				 
	     Program evaluation 				 
	     Surveys						 
	     California 					 
	     Connecticut					 
	     Indiana						 
	     Long-Term Care Partnership Program 		 
	     New York						 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Product.                                                 **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-07-231

   

     * [1]Results in Brief
     * [2]Background

          * [3]Long-Term Care Insurance
          * [4]Long-Term Care Insurance Regulation
          * [5]Medicaid
          * [6]Long-Term Care Partnership Programs

     * [7]Partnership Policies Must Include Certain Benefits Not Requi

          * [8]The Four States Require Partnership Policies to Include Cert
          * [9]Partnership Policyholders Purchased Policies with Benefits T
          * [10]Insurance Companies Cannot Charge Partnership Policyholders

     * [11]Compared with Traditional Long-Term Care Insurance Policies,
     * [12]Long-Term Care Insurance Policyholders Are Generally Wealthi

          * [13]Long-Term Care Insurance Policyholders Generally Have Higher
          * [14]Partnership Policyholders Are Younger on Average than Tradit

     * [15]Partnership Programs Unlikely to Result in Savings for Medic

          * [16]Most Partnership Policyholders Would Have Purchased Traditio
          * [17]Few Partnership Policyholders Are Likely to Become Eligible

     * [18]Concluding Observations
     * [19]Agency and State Comments and Our Evaluation

          * [20]Impact of Asset Transfers on Medicaid

               * [21]Methodology for Assessing Medicaid Savings
               * [22]Impact on Medicaid Savings of Purchasing Different
                 Levels of

     * [23]Appendix I: Data and Methods for Analysis of Long-Term Care

          * [24]Examining Long-Term Care Insurance Benefits Purchased by Par

               * [25]Examining Income and Asset Distributions Among
                 Partnership P
               * [26]Examining Demographic Characteristics--Age, Gender, and
                 Marit
               * [27]Data Reliability

     * [28]Appendix II: Explanation of the Simplifying Assumptions Used

          * [29]Illustrative Scenarios for Time Taken to Become Eligible for

               * [30]Evaluating the Effects of Adjusting the Assumptions
                 Underlyi

     * [31]Appendix III: Comments from the Department of Health & Human
     * [32]Appendix IV: Comments from the Four States with Partnership
     * [33]Appendix V: GAO Contact and Staff Acknowledgments

          * [34]GAO Contact
          * [35]Acknowledgments

               * [36]Order by Mail or Phone

Report to Congressional Requesters

United States Government Accountability Office

GAO

May 2007

LONG-TERM CARE INSURANCE

Partnership Programs Include Benefits That Protect Policyholders and Are
Unlikely to Result in Medicaid Savings

GAO-07-231

Contents

Letter 1

Results in Brief 7
Background 11
Partnership Policies Must Include Certain Benefits Not Required of
Traditional Policies, and Insurance Companies Cannot Charge Higher
Premiums for Asset Protection in Partnership Policies 21
Compared with Traditional Long-Term Care Insurance Policies, Two of Four
States Subject Partnership Policies to Additional Review, and All Four
States Require Additional Agent Training 28
Long-Term Care Insurance Policyholders Are Generally Wealthier than Those
Without Such Insurance, and Partnership Policyholders Are Typically
Younger than Traditional Long-Term Care Insurance Policyholders 30
Partnership Programs Unlikely to Result in Savings for Medicaid Largely
Because of the Asset Protection Benefit of Partnership Policies 34
Concluding Observations 42
Agency and State Comments and Our Evaluation 43
Appendix I Data and Methods for Analysis of Long-Term Care Insurance
Benefits and Demographics 51
Appendix II Explanation of the Simplifying Assumptions Used in the
Illustrative Scenarios 55
Appendix III Comments from the Department of Health & Human Services 62
Appendix IV Comments from the Four States with Partnership Programs,
California, Connecticut, Indiana, and New York 63
Appendix V GAO Contact and Staff Acknowledgments 79

Tables

Table 1: Percentage of Partnership and Traditional Long-Term Care
Insurance Policyholders Purchasing Benefits from 2002 through 2005 26
Table 2: Household Income and Household Asset Distribution among
Partnership Policyholders and Comparison Populations in Partnership States
and Nationally 32
Table 3: Demographic Characteristics of Partnership Policyholders and
Comparison Populations in Partnership States and Nationally 33

Figure

Figure 1: Financing of Long-Term Care Nursing Facility Stays Under Three
Scenarios 37

Abbreviations

ADL activities of daily living
ACS American Community Survey
CBO Congressional Budget Office
CMS Centers for Medicare & Medicaid Services
DOI Department of Insurance
DRA Deficit Reduction Act of 2005
HHS Department of Health and Human Services
HIPAA Health Insurance Portability
and Accountability Act of 1996
HRS Health and Retirement Study
IADL instrumental activities of daily living
NAIC National Association of Insurance Commissioners
OBRA `93 Omnibus Budget Reconciliation Act of 1993
UDS Uniform Data Set

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

United States Government Accountability Office
Washington, DC 20548

May 11, 2007

The Honorable Max Baucus
Chairman
The Honorable Charles E. Grassley
Ranking Member
Committee on Finance
United States Senate

The Honorable John D. Rockefeller, IV
United States Senate

In 2004, national spending on long-term care, which includes care provided
in nursing facilities, totaled $193 billion and nearly half of that was
paid for by Medicaid, the joint federal-state program that finances
medical services for certain low-income adults and children. In contrast,
private insurance paid for about $14 billion worth of long-term
care--about 7 percent of the total cost. The demand for this type of care
is likely to increase as the proportion of those in the population age 65
and older--those most likely to need long-term care--increases. With
Medicaid financing nearly half of the long-term care costs nationwide,
policymakers are concerned that, without changes in how long-term care is
financed, the growing demand for this type of care will continue to strain
the resources of federal and state governments. In 2004, national spending
on long-term care, which includes care provided in nursing facilities,
totaled $193 billion and nearly half of that was paid for by Medicaid, the
joint federal-state program that finances medical services for certain
low-income adults and children. In contrast, private insurance paid for
about $14 billion worth of long-term care--about 7 percent of the total
cost. The demand for this type of care is likely to increase as the
proportion of those in the population age 65 and older--those most likely
to need long-term care--increases. With Medicaid financing nearly half of
the long-term care costs nationwide, policymakers are concerned that,
without changes in how long-term care is financed, the growing demand for
this type of care will continue to strain the resources of federal and
state governments.

In the late 1980s the Robert Wood Johnson Foundation provided start-up
funds for programs in eight states--California, Connecticut, Indiana,
Massachusetts, New Jersey, New York, Oregon, and Wisconsin--aimed at
helping to shift some of the responsibility for financing long-term care
from Medicaid to private long-term care insurance. Four of the states that
received funds--California, Connecticut, Indiana, and New
York--established the programs. These four state-run long-term care
programs, which are known as Partnership programs, encourage individuals
to purchase long-term care insurance by providing an
incentive--specifically, allowing those who purchase long-term care
insurance policies through the program to exempt some or all of their
personal assets from Medicaid eligibility requirements should the
policyholders exhaust their long-term care insurance benefits and need to
continue financing their long-term care. Without the exemption, before
individuals could receive Medicaid benefits they would typically have to
spend their assets on their long-term care until the assets met or fell
below certain Medicaid thresholds. In the late 1980s the Robert Wood
Johnson Foundation provided start-up funds for programs in eight
states--California, Connecticut, Indiana, Massachusetts, New Jersey, New
York, Oregon, and Wisconsin--aimed at helping to shift some of the
responsibility for financing long-term care from Medicaid to private
long-term care insurance. Four of the states that received
funds--California, Connecticut, Indiana, and New York--established the
programs. These four state-run long-term care programs, which are known as
Partnership programs, encourage individuals to purchase long-term care
insurance by providing an incentive--specifically, allowing those who
purchase long-term care insurance policies through the program to exempt
some or all of their personal assets from Medicaid eligibility
requirements should the policyholders exhaust their long-term care
insurance benefits and need to continue financing their long-term care.
Without the exemption, before individuals could receive Medicaid benefits
they would typically have to spend their assets on their long-term care
until the assets met or fell below certain Medicaid thresholds. Medicaid
does not allow for asset protection for long-term care insurance policies
purchased outside of Partnership programs.^1 In order to implement their
Partnership programs, the four states with Partnership programs had to
obtain approval from the Centers for Medicare & Medicaid Services (CMS),
the agency within the Department of Health and Human Services (HHS) that
oversees Medicaid, and amend their state Medicaid plans to allow them to
exempt the assets of Partnership program participants from Medicaid
eligibility requirements.^2,3

Since the early 1990s, the treatment of Partnership programs under federal
law has changed. Although a number of states established, or were
authorized to establish, programs prior to the enactment of the Omnibus
Budget Reconciliation Act of 1993 (OBRA `93), OBRA `93 prohibited
additional states from establishing similar programs. The legislation was
enacted, in part, because of concerns about potential costs to Medicaid,
but allowed California, Connecticut, Indiana, and New York to maintain
their programs.^4,5 More recently, the Deficit Reduction Act of 2005 (DRA)
authorized all states to establish Partnership programs that meet certain
criteria and required the original 4 participating states to maintain the
existing consumer protections in their Medicaid plans. DRA provisions are
intended, in part, to allow states to provide an incentive for individuals
to take responsibility for their own long-term care needs rather than
relying on Medicaid. According to the National Association of Health
Underwriters, prior to the enactment of DRA, there was legislative
activity in 19 additional states to begin development of a Partnership
program. As of October 2006, the only states with active Partnership
programs were the original 4 states: California, Connecticut, Indiana, and
New York.^6 However, HHS indicated that as of February 2007, CMS had
approved Partnership program state plan amendments in 6 states: Florida,
Georgia, Idaho, Minnesota, Nebraska, and Virginia. Although the program
appears to be expanding beyond the original 4 states, concerns about the
potential cost to Medicaid of expanding the program remain an issue. In
2005, the Congressional Budget Office (CBO) estimated that repealing the
moratorium on new Partnership programs could increase Medicaid spending by
$86 million between 2006 and 2015.^7

^1For the purposes of this report, we use the term "Partnership policies"
to refer to long-term care insurance policies purchased through
Partnership programs and the term "traditional long-term care insurance"
to refer to long-term care insurance policies that are not purchased
through these programs. To refer to both Partnership and traditional
long-term care insurance policies, we use the term "long-term care
insurance."

^2A state plan describes the state's Medicaid program and establishes
guidelines for how the state's Medicaid program will function.

^3For our purposes we use the Partnership program's definition of
"assets," that is, when we refer to assets, we mean savings and
investments, while excluding income. For eligibility purposes, the
Medicaid program considers both income--which is anything received during
a calendar month that is used or could be used to meet food or shelter
needs--and resources, which are cash or anything owned, such as savings
accounts, stocks, or property that can be converted to cash.

^4Another objective of OBRA `93, as expressed in the accompanying House of
Representatives Budget Committee report, was to close a loophole
permitting wealthy individuals to qualify for Medicaid. H.R. Rep. No.
103-111, at 536.

^5Prior to the enactment of OBRA `93, California, Connecticut, Indiana,
and New York established Partnership programs. Iowa and Massachusetts also
received permission from the Health Care Financing Administration (now
CMS) to establish a Partnership program, but had not implemented one as of
October 2006.

States are responsible for overseeing Partnership programs and regulating
the Partnership programs as well as the traditional long-term care
insurance policies sold in their states. As more states consider
establishing Partnership programs, there is interest, on the part of
Congress and others, in understanding how the four states with Partnership
programs designed and regulate their Partnership programs, who purchases
Partnership policies, and how these programs will impact Medicaid
financially.

You asked us to analyze the experience of the four states with Partnership
programs. In August 2005, we provided you with a briefing, which
summarized aspects of the design of these Partnership programs and
included demographic information on Partnership policyholders.^8 In this
report, we updated our briefing information and provided a more detailed
analysis of the Partnership programs. Specifically, we examined (1) the
benefits and premium requirements of Partnership policies as compared with
those of traditional long-term care insurance policies, including
information on benefits purchased by policyholders; (2) the extent to
which states oversee Partnership policies as compared with their oversight
of traditional long-term care insurance policies; (3) the demographics,
including asset and income levels, of Partnership policyholders,
traditional long-term care insurance policyholders, and people without
long-term care insurance; and (4) whether the Partnership programs are
likely to result in savings for Medicaid.

^6When we refer to the four states with Partnership programs or the four
states, we are referring to California, Connecticut, Indiana, and New
York. According to CMS officials, as of October 2006, no other states had
active Partnership programs; that is, no insurance companies were issuing
Partnership policies in any other states.

^7Congressional Budget Office Cost Estimate: S. 1932 Deficit Reduction Act
of 2005, at 36-39, January 27, 2006.

^8GAO, Overview of the Long-Term Care Partnership Program,
[37]GAO-05-1021R (Washington, D.C.: Sept. 9, 2005).

To compare the benefits and premium requirements of Partnership and
traditional long-term care insurance policies, we reviewed state
regulations, and interviewed Partnership program officials and department
of insurance (DOI) officials in each of the four states with Partnership
programs-California, Connecticut, Indiana, and New York. To compare the
benefits purchased by Partnership policyholders and traditional long-term
care insurance policyholders, we obtained data from 2002 through 2005 from
two sources. Our data source for benefits purchased by Partnership
policyholders was the Uniform Data Set (UDS)--a data set with information
on Partnership policyholders compiled by officials in each of the four
states with Partnership programs from data provided by participating
insurers.^9 Our data source for benefits purchased by traditional
long-term care insurance policyholders was from a survey we conducted of
five of the largest long-term care insurance companies in the individual
long-term care insurance market.^10

To examine the extent to which states oversaw Partnership policies
compared with state oversight of traditional long-term care insurance
policies, we reviewed state regulations and Partnership program documents,
and interviewed officials from Partnership programs, long-term care
insurance companies, and each Partnership state's DOI, the entities that
are responsible for regulating insurance policies, including long-term
care insurance policies, that are sold in the states. We reviewed state
regulations, Partnership program documents, and conducted interviews about
how training requirements for insurance agents who sell Partnership
policies compared with training requirements for agents who sell
traditional long-term care insurance policies.

^9The UDS is a data set developed by the four states with Partnership
programs; participating insurers; the National Program Office at the
Center on Aging, University of Maryland; and the Program Evaluator, Laguna
Research Associates. Data in the UDS are submitted by insurers to the
Partnership program in the state in which they are participating and
contain information on Partnership policyholders.

^10We selected the five insurance companies on the basis of the total
number of policies and amount of annualized premiums in effect in the
individual market as of December 31, 2004.

To examine the demographics, including income and assets levels, of
Partnership policyholders, traditional long-term care insurance
policyholders, and individuals without long-term care insurance, we used
data from three sources. First, to calculate the household income and
assets of Partnership policyholders, we used available survey data from a
sample of Partnership policyholders in California and Connecticut. We
restricted our analysis to the income and asset data from these two states
because Indiana's data were not sufficiently detailed to include in our
analysis, and New York was not able to provide us with data from recent
years. We combined multiple years of these data in order to increase the
sample size.^11 To estimate the household income of individuals without
insurance in California and Connecticut, we used data from the American
Community Survey (ACS) for 2004 published by the U.S. Census Bureau.
Finally, we used national data from the Health and Retirement Study (HRS)
for 2004, to compare household income and household assets for those
individuals with traditional long-term care insurance and those without
long-term care insurance.^12 The HRS is a national survey sponsored by the
National Institute on Aging and conducted by the University of Michigan of
individuals over the age of 50.^13 The HRS collected information about
retirement, health insurance, savings, and other issues confronting the
elderly. To examine the age, marital status, and gender of Partnership
policyholders, traditional long-term care insurance policyholders, and
individuals without long-term care insurance, we used data from the UDS
and the HRS.

To examine whether the Partnership programs in the four states are likely
to result in savings for Medicaid, we assessed (1) available state survey
data of Partnership policyholders and (2) the options an individual has
for financing long-term care and the time it would take for the individual
to become eligible for Medicaid under three illustrative scenarios. We
used the illustrative scenarios because the Partnership programs in the
four states have only been operating since the early 1990s, and as yet
there are no available data describing when or if Partnership
policyholders would have accessed Medicaid. As a result, there are
insufficient data available to directly measure whether the Partnership
programs are associated with increased or decreased Medicaid spending. We
used available survey data in California, Connecticut, and Indiana to
determine what Partnership policyholders report they would have done to
finance their long-term care needs if there had not been a Partnership
program in their state.^14 In addition, we assessed three scenarios that
represent the three main options an individual has for financing long-term
care: financing using a Partnership policy, financing using a traditional
long-term care policy, and self-financing without any long-term care
insurance. The latter two scenarios describe the financing options that a
Partnership policyholder could use if the Partnership programs did not
exist. We used the three scenarios to explore how long it would likely
take before the individual depicted in our scenarios would become eligible
for Medicaid with a Partnership policy and--in the scenarios in which
Partnership programs did not exist--with the other two financing options.
In the scenarios, if, in the absence of a Partnership program, an
individual using a traditional long-term care insurance policy or relying
on self-financing is likely to become eligible for Medicaid sooner than
the same individual would have using a Partnership policy, we consider the
Partnership programs to be a potential source of savings for Medicaid. In
contrast, if the same individual delays Medicaid eligibility using a
traditional long-term care insurance policy or self-financing, when
compared with the time it would take the individual to become eligible for
Medicaid using a Partnership policy, we consider the Partnership program
to be a potential source of increased spending for Medicaid. To develop
our scenarios, we made several simplifying assumptions. These include the
following:

^11For income and asset data in California we combined data for 2003 and
2004, and for Connecticut, we combined data for 2002 through 2005.

^12To make our income analysis consistent across the different data
sources, we restricted our calculations of household income to individuals
aged 55 and over.

^13The Health and Retirement Study (HRS) is a longitudinal national panel
survey that collects information over time on individuals over age 50. The
first survey was conducted in 1992, and subsequent surveys were conducted
every 2 years. The most recent survey for which data were available was
2004.

           o The individual depicted in the scenarios has $300,000 in assets,
           and in two of our scenarios a long-term care insurance policy
           worth $210,000--assets and benefits that are typical of many
           individuals with long-term care insurance--and the individual
           receives long-term care in a nursing facility with costs for a
           year of care of $70,000, about equal to average nursing facility
           costs nationwide in 2004.

           o The individual has assets that are no less than the value of the
           individual's Partnership policy--that is, the individual does not
           overinsure his or her assets.
		   
		   ^14New York State survey data were unavailable.
		   
           o The individual is unmarried. While most Partnership
           policyholders are married at the time they purchase a Partnership
           policy, they are unlikely to require long-term care for many
           years, and their marital status can change. Most individuals who
           are admitted to a nursing facility are unmarried.

           Where possible, we use data from surveys of Partnership
           policyholders to support our assumptions. We also explored whether
           adjusting the assumptions changed the conclusions we could draw.
           Although our scenarios represent the choices facing a single
           individual, the results of this analysis are applicable beyond
           this individual. For example, the relative impact on Medicaid
           spending across the scenarios is independent of the amount of
           assets owned by the individual or the level of the individual's
           insurance coverage.

           As part of our efforts to examine whether the Partnership programs
           are likely to result in savings for Medicaid, we also examined the
           likelihood that the population of Partnership policyholders will
           ever become eligible for Medicaid. To assess this likelihood, we
           examined the long-term care insurance benefits and income of
           Partnership policyholders. We also assessed the number of people
           with Partnership policies who accessed Medicaid as of October
           2006.

           Based on discussions with state officials and reviewing
           documentation on uniformly collected insurer data and surveys of
           policyholders, we determined that the information we used was
           sufficiently reliable for our purposes. We also examined reports
           on the Partnership program from the CBO, the Congressional
           Research Service, and other research organizations. Appendix I
           provides information on the data and methods used for our analyses
           of long-term care insurance benefits, policyholder income, assets,
           age, gender, and marital status. Appendix II provides more
           information about the illustrative scenarios, the simplifying
           assumptions underlying the scenarios, and the effect on our
           analysis of adjusting these assumptions. We conducted our work
           from September 2005 through May 2007 in accordance with generally
           accepted government auditing standards.
		   
		   Results in Brief

           In the four states with Partnership programs, Partnership policies
           must include certain benefits not generally required of
           traditional long-term care insurance policies, and insurance
           companies cannot charge higher premiums for asset protection in
           Partnership policies. Partnership policies must include certain
           benefits, such as inflation protection and minimum daily benefit
           amounts, while traditional long-term care insurance policies may
           include these benefits but are generally not required to do so.
           Partnership policies include these benefits in order to increase
           the likelihood that Partnership policyholders will have sufficient
           long-term care insurance coverage to pay for a significant portion
           of their long-term care. For example, Partnership policies must
           include inflation protection, which increases the amount a policy
           pays over time to account for increases in the cost of care, and
           minimum daily benefit amounts, which are set at levels designed to
           cover a significant portion of the costs of an average day in a
           nursing facility. Though traditional long-term care insurance
           policyholders are able to purchase most of the same benefits as
           Partnership policyholders, in comparing these two groups we found
           that a higher percentage of Partnership policyholders purchased
           policies from 2002 through 2005 with more extensive coverage--for
           example, higher levels of inflation protection and coverage for
           care in both nursing facility and home and community-based care
           settings. Officials in states with Partnership programs told us
           that companies selling long-term care insurance are not permitted
           to charge Partnership policyholders higher premiums for the asset
           protection benefit--Partnership and traditional long-term care
           insurance policies with otherwise comparable benefits must have
           equivalent premiums. However, Partnership policies are likely to
           have higher premiums because they are required to have inflation
           protection and other benefits that are not required for
           traditional long-term care insurance policies.

           According to state officials, compared with traditional long-term
           care insurance policies, Partnership policies in two Partnership
           states are subject to additional review, and in all four
           Partnership states, insurance agents who sell Partnership policies
           are subject to additional state training requirements compared
           with agents who sell only traditional long-term care policies.
           While all long-term care insurance policies are reviewed by the
           DOI in each state, Partnership policies in California and
           Connecticut are also reviewed by Partnership program offices. This
           additional review is designed by these states to ensure that the
           Partnership policies that are issued meet all specific Partnership
           regulatory requirements, and the insurance companies issuing these
           policies meet the data reporting and other administrative
           requirements. Before they can sell Partnership policies, long-term
           care insurance agents are required by each of the four Partnership
           states to undergo training specific to the state's Partnership
           program, in addition to the training that is required for those
           who sell traditional long-term care insurance. This specialized
           training typically provides information on long-term care
           planning, Medicaid, Medicare, the specific benefits required by
           the state's Partnership program, and how policies sold through the
           program differ from traditional long-term care insurance policies.

           Partnership and traditional long-term care insurance policyholders
           tend to be relatively wealthy with higher incomes and more assets,
           compared with those without insurance. At the time they purchased
           their Partnership policies, more than half of Partnership
           policyholders over 55 in California and Connecticut had monthly
           household incomes of $5,000 or greater, and more than half of all
           households had assets of $350,000 or greater. Partnership
           policyholders in the four states with Partnership programs are
           also younger on average than traditional long-term care insurance
           policyholders. In addition, a higher percentage of Partnership and
           traditional long-term care insurance policyholders are married
           rather than unmarried, and female rather than male.

           Available survey data from three states with Partnership programs
           and our three illustrative financing scenarios together suggest
           that the Partnership programs are unlikely to result in savings
           for Medicaid and may result in increased Medicaid spending. Based
           on surveys of Partnership policyholders in California,
           Connecticut, and Indiana, we estimate that, in the absence of a
           Partnership program in their state, 80 percent of Partnership
           policyholders would have purchased a traditional long-term care
           insurance policy. Our long-term care financing scenarios suggest
           that it takes longer for an individual with a traditional
           long-term care insurance policy to become eligible for Medicaid
           than it would take the same individual to become eligible for
           Medicaid if he or she owned a Partnership policy. Therefore, the
           80 percent of surveyed Partnership policyholders may represent a
           potential source of increased spending for Medicaid, as Medicaid
           is likely to begin paying for the long-term care of these
           policyholders sooner than if they had held traditional long-term
           care insurance policies instead. The survey data also indicate
           that the remaining 20 percent of surveyed policyholders would not
           have purchased any long-term care insurance if Partnership
           programs did not exist. Data are not yet available to directly
           measure when or if these individuals will access Medicaid had they
           not purchased a Partnership policy. However, our scenarios suggest
           that an individual who self-finances his or her long-term care
           without any long-term care insurance is likely to become eligible
           for Medicaid at about the same time as the individual would using
           a Partnership policy, though there were some circumstances that
           could accelerate or delay the individual's time to Medicaid
           eligibility. While the majority of Partnership policyholders have
           the potential to increase spending, we also anticipate that the
           impact of these programs is likely to be small because few
           policyholders will become eligible. Partnership policyholders tend
           to have incomes that exceed Medicaid eligibility thresholds and
           insurance benefits that cover most of their long-term care needs.

           With DRA authorizing all states to implement Partnership programs,
           information on the Partnership policies and policyholders from the
           four states with Partnership programs may prove useful to other
           states considering implementing such programs. States may want to
           consider the benefits to Partnership policyholders, the likely
           impact on Medicaid expenditures, and the incomes and assets of
           those likely to be able to afford long-term care insurance.

           We received comments on a draft of this report from HHS and state
           officials from California, Connecticut, Indiana, and New York. HHS
           commented that our study results should not be considered
           conclusive and the simplified scenarios were flawed because they
           did not adequately account for the effect of asset transfers. HHS
           also noted that our data sources were unlikely to yield accurate
           data on asset transfers and criticized the report for not
           incorporating a review of the literature on this issue and the
           analyses conducted by the four states with Partnership programs.
           The four states disagreed with our conclusion that the Partnership
           programs are unlikely to result in Medicaid savings, and like HHS,
           commented that our scenarios did not adequately account for the
           impact of asset transfers. California, Connecticut, and New York
           objected to our methodology for estimating the financial impact of
           the program on Medicaid. California and Connecticut suggested that
           our analysis should have included results from two Partnership
           policyholder survey questions that they consider in their own
           analysis of the Partnership program.

           We maintain that the evidence suggests that the Partnership
           program is unlikely to result in savings for Medicaid, despite
           limited data and program experience. As discussed in our draft
           report, some savings to Medicaid could be associated with
           individuals who would have transferred their assets and become
           eligible for Medicaid sooner in the absence of the Partnership
           program. However, we noted that these savings are unlikely to
           offset the potential costs associated with policyholders who would
           have purchased traditional long-term care insurance in the absence
           of the Partnership programs. We did not provide a review of the
           literature on asset transfers because--as we previously noted in
           our March 2007 report on the subject--the evidence on the transfer
           of assets to become eligible for Medicaid coverage for long-term
           care is generally limited.^15 However, in response to HHS'
           comments, we have amended our draft report to make the discussion
           of asset transfers more prominent in the body of our report and to
           include reference to the 2007 GAO study. We also maintain that our
           methodology for estimating the financial impact of the program on
           Medicaid is sound and disagree with California and Connecticut
           regarding the appropriateness of using the two survey questions.
           Specifically, by relying on the responses from these questions,
           the method California, Connecticut, and Indiana use to evaluate
           Medicaid costs underestimates the percentage of people who would
           have purchased traditional policies in the absence of the
           Partnership program, while their method of evaluating Medicaid
           savings overestimates the percentage of people who would transfer
           assets.
		   
		   Background

           Long-term care comprises services provided to individuals who,
           because of illness or disability, are generally unable to perform
           activities of daily living (ADL)--such as bathing, dressing, and
           getting around the house--for an extended period of time.^16 These
           services can be provided in various settings, such as nursing
           facilities, an individual's own home, or the community.^17 The
           typical 65-year-old has about a 70 percent chance of needing
           long-term care services in his or her life.^18 Long-term care can
           be expensive, especially when provided in nursing facilities. In
           2005, the average cost of a year of nursing facility care was
           about $70,000.^19 In 1999, the most recent year for which data
           were available, the average length of stay in a nursing facility
           was between 2 and 3 years.^20
		   
^15GAO, Medicaid Long-Term Care: Few Transferred Assets before Applying
for Nursing Home Coverage; Impact of Deficit Reduction Act on Eligibility
Is Uncertain, [38]GAO-07-280 (Washington, D.C.: Mar. 26, 2007).

^16As people age, they typically experience a decline in their ability to
perform basic physical functions, increasing the likelihood that they will
need long-term care services. Individuals qualify for Medicaid coverage
for long-term care services if they meet certain functional criteria that
in general involve a degree of impairment measured by the level of
assistance an individual needs to perform six activities of daily living
(ADL): eating, bathing, dressing, using the toilet, getting in and out of
bed, and getting around the house, as well as the instrumental activities
of daily living (IADL), which include preparing meals, shopping for
groceries, and venturing outside of a home or facility.

^17Long-term care services, such as personal care, homemaker services, and
respite care, are known as home care. Home care can also include services
provided outside of policyholders' homes, such as services provided in
adult day care centers. Long-term care services provided in
community-based facilities are generally designed to help people receive
long-term care and remain living in their own homes. Known as
community-based services, these long-term care services can be supplied in
settings such as policyholders' homes, adult day care facilities, or
during visits to a physician's office.

^18P. Kemper, H.L. Komisar, and L. Alecxih, Long-Term Care Over an
Uncertain Future: What Can Current Retirees Expect? Inquiry, vol. 42, no.
4 (Winter 2005/2006) pp. 335-350.

^19MetLife Mature Market Institute, The MetLife Market Survey of Nursing
Home & Home Care Costs (September 2005).

^20Department of Health and Human Services, Centers for Disease Control
and Prevention, National Center for Health Statistics, The National
Nursing Home Survey: 1999 Summary, Series 13, No. 152 (June 2002).

           Long-Term Care Insurance		   

           Long-term care insurance is used to help cover the cost associated
           with long-term care. Individuals can purchase long-term care
           insurance policies directly from insurance companies, or through
           employers or other groups. The number of long-term care insurance
           policies sold has been small-about 9 million as of 2002, the most
           recent year for which data were available. About 80 percent of
           these policies were sold through the individual insurance market
           and the remaining 20 percent were sold through the group market.

           Long-term care insurance companies generally structure their
           long-term care insurance policies around certain types of benefits
           and related options.

           o A policy with comprehensive coverage pays for long-term care in
           nursing facilities as well as for care in home and community
           settings, while a policy with coverage for home and
           community-based settings pays for care only in these settings.

           o A daily benefit amount specifies the amount a policy will pay on
           a daily basis toward the cost of care, while a benefit period
           specifies the overall length of time a policy will pay for care.
           Data from 2002 through 2005 show that the maximum daily benefit
           amounts can range from less than $100 to several hundred dollars
           per day, while benefit periods can range from 1 year to lifetime
           coverage.^21
		   
^21GAO analysis of data provided by five insurance companies selling
traditional long-term care insurance: see GAO, Long-Term Care Insurance:
Federal Program Compared Favorably with Other Products, and Analysis of
Claims Trend Could Inform Future Decisions, [39]GAO-06-401 (Washington,
D.C.: March 2006).

           o A policy's elimination period establishes the length of time a
           policyholder who has begun to receive long-term care has to wait
           before his or her insurance will begin making payments towards the
           cost of care. According to data from 2002 through 2005,
           elimination periods can range from 0 to at least 730 days.^22

           o Inflation protection increases the maximum daily benefit amount
           covered by a policy, and helps ensure that over time the daily
           benefit remains commensurate with the costs of care.

           There can be a substantial gap between the time a long-term care
           insurance policy is purchased and the time when policyholders
           begin using their benefits, and the costs associated with
           long-term care can increase significantly during this time. A
           typical gap between the time of purchase and the use of benefits
           is 15 to 20 years: the average age of all long-term care insurance
           policyholders at the time of purchase is 63, and in general
           policyholders begin using their benefits when they are in their
           mid-70s to mid-80s. Usually, automatic inflation protection
           increases the benefit amount by 5 percent annually on a compounded
           basis. A policy with automatic 5 percent compound inflation
           protection and a $150 per day maximum daily benefit in 2006 would
           be worth approximately $400 per day 20 years later. Another means
           to protect against inflation is a future purchase option. This
           option allows the consumer to increase the dollar amount of
           coverage every few years at an extra cost. Some future purchase
           options do not allow consumers to purchase extra coverage once
           they begin receiving their insurance benefit and the opportunity
           to purchase extra coverage may be withdrawn should the consumer
           decline a predetermined number of premium increases. A policy with
           a future purchase option may be less expensive initially than a
           policy with compound inflation protection. However, over time the
           policy with a future purchase option may become more expensive
           than a policy with compound inflation.

           Without inflation protection, policyholders might purchase a
           policy that covers the current cost of long-term care but find,
           many years later, when they are most likely to need long-term care
           services, that the purchasing power of their coverage has been
           reduced by inflation and that their coverage is less than the cost
           of their care. For example, if the cost of a day in a nursing
           facility increases by 5 percent every year for 20 years, a nursing
           facility that costs $150 per day in 2006 would cost about $400 per
           day 20 years later in 2026. A policy purchased in 2006 with a
           daily benefit of $150 without inflation protection would pay $150
           per day--or 38 percent--of the daily cost of about $400 in 2026.
           The remaining $250 of the daily cost of the nursing facility care
           would have to be paid by the policyholder.
		   
^22See [40]GAO-06-401 .		   

           Long-term care insurance policies may also include other benefits
           or options. For example, policies can offer coverage for home care
           at varying percentages of the maximum daily benefit amount. Some
           policies include features in which the policy returns a portion of
           the premium payments to a designated third party if the
           policyholder dies. Some policies provide coverage for long-term
           care provided outside of the United States or offer
           care-coordination services that, among other things, provide
           information about long-term care services to the policyholder and
           monitor the delivery of long-term care services.

           Many factors impact the premiums individuals pay for long-term
           care insurance. Notably, long-term care insurance companies
           typically charge higher premiums for policies with more extensive
           benefits. In general, policies with comprehensive coverage have
           higher premiums than policies without such coverage, and
           policyholders pay higher premiums the higher their maximum daily
           benefit amounts, the longer their benefit periods, the greater
           their inflation protection, and the shorter their elimination
           periods. For example, in Connecticut, if a 55-year-old man decided
           to buy a 1-year, $200 per day comprehensive coverage policy, in
           2005 it would have cost him about $1,000 less per year than a
           comparable 3-year policy. Similarly, the age of an applicant also
           impacts the premium, as premiums typically are more expensive the
           older the policyholder at the time of purchase. For example, in
           Connecticut, a 55-year-old purchasing a 3-year, $200 per day
           comprehensive coverage policy in 2005 would pay about $2,500 per
           year, whereas a 70-year-old purchasing the same policy would pay
           about $5,900 per year. Health status may also affect premiums.
           Insurance companies take into account the health status of an
           applicant to evaluate the risk that he or she will access
           long-term care services. If an applicant has a medical condition
           that increases the likelihood of the applicant using long-term
           care services, but does not automatically disqualify the applicant
           from purchasing insurance, the applicant may receive a substandard
           rating from an insurance company, which may result in a higher
           premium.^23
		   
		   Long-Term Care Insurance Regulation

           Regulation of the insurance industry, including those companies
           selling long-term care insurance, is a state function. Those who
           sell long-term care insurance must be licensed by each state in
           which they sell policies, and the policies sold must be in
           compliance with state insurance laws and regulations. These laws
           and regulations can vary but their fundamental purpose is to
           establish consumer protections that are designed to ensure that
           the policies' provisions comply with state law, are reasonable and
           fair, and do not contain major gaps in coverage that might be
           misunderstood by consumers and leave them unprotected.

           The Health Insurance Portability and Accountability Act of 1996
           (HIPAA) specified conditions under which long-term care insurance
           benefits and premiums would receive favorable federal income tax
           treatment.^24 Individuals who purchase policies that comply with
           HIPAA requirements, which are therefore "tax-qualified," can
           itemize their long-term care insurance premiums as deductions from
           their taxable income along with other medical expenses, and can
           exclude from gross income insurance company proceeds used to pay
           for long-term care expenses. Under HIPAA, tax-qualified plans must
           begin coverage when a person is certified as: needing substantial
           assistance with at least two of the six ADLs for at least 90 days
           due to a loss of functional capacity, having a similar level of
           disability, or requiring substantial supervision because of a
           severe cognitive impairment. HIPAA also requires that a policy
           comply with certain provisions of the National Association of
           Insurance Commissioners' (NAIC) Long-Term Care Insurance Model Act
           and Regulation adopted in January 1993. This model act and
           regulation established certain consumer protections that are
           designed to prevent insurance companies from (1) not renewing a
           long-term care insurance policy because of a policyholder's age or
           deteriorating health, and (2) increasing the premium of an
           existing policy because of a policyholder's age or claims history.
           In addition, in order for a long-term care insurance policy to be
           tax-qualified, HIPAA requires that a policy offer inflation
           protection. The NAIC, which represents insurance regulators from
           all states, reported in 2005 that 41 states based their long-term
           care insurance regulations on the NAIC model, 7 based their
           regulations partially on the model, and 3 did not follow the
           model.
		   
^23The process of reviewing medical and health-related information
furnished by an applicant to determine if the applicant presents an
acceptable level of risk and is insurable is known as underwriting.
Examples of medical conditions that may not disqualify an individual from
obtaining insurance but that can result in a substandard rating during the
underwriting process include osteoporosis, emphysema, and diabetes.
However, the severity and the ability to control and treat the medical
condition are all factors that can also impact how a nondisqualifying
medical condition impacts an underwriting rating.

^24Pub. L. No. 104-191, SS 321-327, 110 Stat. 1936, 2054-2067.

           Medicaid

           Medicaid is the primary source of financing for long-term care
           services in the United States. In 2004, almost one-third of the
           total $296 billion in Medicaid spending was for long-term care.
           Some health care services, such as nursing facility care, must be
           covered in any state that participates in Medicaid. States may
           choose to offer other optional services in their Medicaid plans,
           such as personal care.^25

           Medicaid coverage for long-term care services is most often
           provided to individuals who are aged or disabled. To qualify for
           Medicaid coverage for long-term care, these individuals must meet
           both functional and financial eligibility criteria. Functional
           eligibility criteria are established by each state and are
           generally based on an individual's degree of impairment, which is
           measured in terms of the level of difficulty in performing the
           ADLs and IADLs. To meet the financial eligibility criteria, an
           individual cannot have assets or income that exceed thresholds
           established by the states and that are within standards set by the
           federal government. Generally, the value of an individual's
           primary residence and car, as well as a few other personal items,
           are not considered assets for the purpose of determining Medicaid
           eligibility.^26 Individuals with high medical costs and whose
           income exceeds state thresholds can "spend down" their income on
           their long-term care, which may bring their income below the
           state-determined income eligibility limit. In all four states with
           Partnership programs, for the purpose of obtaining Medicaid
           eligibility, individuals are allowed to deduct medical expenses,
           including those for long-term care, in order to bring their
           incomes below the state-determined thresholds.
		   
^25Personal care includes long-term care services that help people meet
personal needs such as assistance with personal hygiene, nutritional or
support functions, and health-related tasks.

^26Under DRA, certain individuals with an equity interest in their home of
greater than $500,000 are not eligible for Medicaid coverage for nursing
facility services or other long-term care services. However, states have
the option to increase the home equity interest level to an amount that
does not exceed $750,000. This home equity limitation does not apply to
individuals if they have a spouse, a child under age 21, or a child who is
blind or disabled living in the home.

           In order to meet Medicaid's eligibility requirements, some
           individuals may choose to divest themselves of their assets--for
           example, by transferring assets to their spouses or other family
           members.^27 However, those who transfer assets for less than fair
           market value during a specified "look-back" period--a period of
           time before an individual applies for Medicaid during which the
           program reviews asset transfers--may incur a penalty, that is, a
           period during which they are ineligible for Medicaid coverage for
           long-term care services. Evidence of the extent to which
           individuals transfer assets for less than fair market value to
           become financially eligible for Medicaid coverage for long-term
           care is generally limited and often based on anecdote. However,
           our March 2007 report on asset transfers suggests that the
           incidence of asset transfers is low among nursing home residents
           covered by Medicaid.^28 Nationwide, about 12 percent of
           Medicaid-covered elderly nursing home residents reported
           transferring cash during the 4 years prior to nursing home entry,
           and the median amount transferred was very small ($1,239). The
           percentage of nursing home residents not covered by Medicaid who
           transferred cash was about twice that of Medicaid-covered nursing
           home residents. However, the median amount of cash transferred as
           reported by non-Medicaid covered residents and Medicaid-covered
           residents did not vary greatly.^29 In addition to the nationwide
           analysis, our report summarized an analysis of a sample of
           approved Medicaid nursing home applicants in three states who
           generally applied to Medicaid in 2005 or before, and found that
           about 10 percent of applicants had transferred assets for less
           than the fair market value during the 3-year look-back period
           before Medicaid eligibility began. The median amount transferred
           was about $15,000. DRA tightened the requirements on Medicaid
           applicants transferring assets by extending the look-back period
           for all asset transfers from 3 to 5 years. In addition, DRA
           changed the beginning date of the penalty period. Prior to
           enactment of DRA, the penalty period started on the first day of
           the month during or after which assets were transferred. DRA
           changed this so that the penalty period now begins on the first
           day of the month when the asset transfer occurred, or the date on
           which the individual is eligible for medical assistance under the
           state plan, and is receiving institutional care services that
           would be covered by Medicaid were it not for the imposition of the
           penalty period, whichever is later. The extension of the look-back
           period and the redefinition of the penalty period may reduce
           transfers of assets.
		   
^27For asset transfer purposes, Medicaid defines the term "assets" to
include income and resources, such as bank accounts.

^28See [41]GAO-07-280 .

^29The median amount of cash transferred by non-Medicaid-covered residents
during the 4 years prior to nursing home entry was $1,859. During the 2
years prior to nursing home entry, the median amount transferred for both
non-Medicaid-covered residents and Medicaid-covered residents was $2,194.

           Long-Term Care Partnership Programs

           The Partnership programs are public-private partnerships between
           states and private long-term care insurance companies. Established
           in 1987 as programs funded through the Robert Wood Johnson
           Foundation, the programs are designed to encourage individuals,
           especially moderate income individuals, to purchase private
           long-term care insurance in an effort to reduce future reliance on
           Medicaid for the financing of long-term care. As of October 2006,
           the original four Partnership programs in California, Connecticut,
           Indiana, and New York remained the only active Partnership
           programs.^30

           Partnership programs attempt to encourage individuals to purchase
           private long-term care insurance by offering them the option to
           exempt some or all of their assets from Medicaid spend-down
           requirements. However, Partnership policyholders are still
           required to meet Medicaid income eligibility thresholds before
           they may receive Medicaid benefits. In the four states with
           Partnership programs, those who purchase long-term care insurance
           Partnership policies generally must first use those benefits to
           cover the costs of their long-term care before they begin
           accessing Medicaid.^31 In 2006, there were about 190,000 active
           Partnership policies, out of the approximately 218,000 Partnership
           policies that had been sold since the inception of the Partnership
           programs.^32,33 Between September 2005 when we last reported on
           the Partnership programs, and August 2006, the number of
           Partnership policies in the four states combined increased by
           about 10 percent.^34
		   
^30Iowa and Massachusetts received approval from the Health Care Financing
Administration (now CMS) to establish Partnership programs, but programs
were not functioning in these states as of October 2006. Since enactment
of DRA, a Partnership program in Idaho was approved by CMS, though the
program was not functioning as of October 2006. Also, as of that date,
amendments to state Medicaid plans allowing Partnership programs in
Florida, Georgia, Minnesota, and Nebraska were under review at CMS.

^31For the purposes of this report, we use the term "accessing Medicaid"
to describe the point at which long-term care policyholders first begin
receiving Medicaid payments for their long-term care.

^32Partnership program offices reported that about 235,000 Partnership
policies had been sold since the four Partnership programs began, but that
number included people who subsequently dropped their policies within 30
days of purchasing the product. The four states with Partnership programs
give Partnership policy purchasers a 30-day "free look" period during
which they can decide whether to keep their policy or drop it and receive
a full refund.

^33By state, the number of Partnership policies, excluding those that were
dropped, was 73,811 in California and 33,040 in Connecticut, through March
2006; 31,750 in Indiana through June 2006; and 51,262 in New York through
December 2005.

^34This rate of increase varied across the states: the sales of
Partnership policies in California increased by 14 percent--the largest
percentage increase among the Partnership states--compared with increases
of 7, 9, and 8 percent in Connecticut, Indiana, and New York,
respectively.	   

           The four states with Partnership programs vary in how they protect
           policyholders' assets. The Partnership programs in California,
           Connecticut, Indiana, and New York have dollar-for-dollar models,
           in which the dollar amount of protected assets is equivalent to
           the dollar value of the benefits paid by the long-term care
           insurance policy. For example, a person purchasing a long-term
           care dollar-for-dollar insurance policy with $300,000 in coverage
           would have $300,000 of assets protected if he or she were to
           exhaust the long-term care insurance benefits and apply for
           Medicaid. However, New York's program also offers total
           protection. That is, those who purchase a comprehensive long-term
           care insurance policy, covering a minimum of 3 years of nursing
           facility care or 6 years of home care, or some combination of the
           two, can protect all their assets at the time of Medicaid
           eligibility determination. In Indiana, in addition to the
           dollar-for-dollar models, the Partnership program offers a hybrid
           model that allows purchasers to obtain dollar-for-dollar
           protection up to a certain benefit level as defined by the state;
           all policies with benefits above that threshold provide total
           asset protection for the purchaser.

           Under DRA, any state that implements a Partnership program must
           ensure that the policies sold through that program contain certain
           benefits, such as inflation protection.^35,36 DRA also requires
           that Partnership policies provide dollar-for-dollar asset
           protection. Insurers are not allowed to offer Partnership policies
           that provide the total asset protection feature found in
           Partnership policies in New York and Indiana.^37 DRA also requires
           Partnership policies to include consumer protections contained in
           the NAIC Long-Term Care Insurance Model Act and Regulation as
           updated in October 2000. DRA established specific requirements for
           Partnership policies that do not apply to traditional long-term
           care insurance policies sold in the Partnership states, such as
           inflation protection and dollar-for-dollar asset protection. DRA
           prohibits states from creating other requirements for Partnership
           policies that do not also apply to traditional long-term care
           insurance policies in the four states with Partnership policies.
           The Partnership programs in California, Connecticut, Indiana, and
           New York, which were implemented before DRA, are not subject to
           these specific requirements, but in order for those programs to
           continue, they must maintain consumer protection standards that
           are no less stringent than those that applied as of December 31,
           2005.
		   
^35DRA requires Partnership policies to provide compound inflation
protection for individuals younger than 61. For individuals younger than
76, Partnership policies must provide policyholders with some level of
inflation protection, although not necessarily compound inflation
protection, while inflation protection is an optional feature for
Partnership policyholders aged 76 or older. Pub. L. No. 109-171, S
6021(a)(1), 120 Stat. 68 (codified at 42 U.S.C. S 1396
p(b)(1)(c)(iii)(IV)).

^36Some of the states that passed legislation prior to the passage of DRA
to enable the creation of a Partnership program may need to make
additional changes to meet DRA requirements.

^37According to CMS officials, policies in New York and Indiana may
continue to provide this type of coverage.	

           Partnership Policies Must Include Certain Benefits Not Required
		   of Traditional Policies, and Insurance Companies Cannot Charge
		   Higher Premiums for Asset Protection in Partnership Policies
   
           The four states with Partnership programs require that Partnership
           policies include certain benefits--such as inflation protection
           and minimum daily benefit amounts--while traditional long-term
           care insurance policies may include these benefits but are not
           generally required to do so. Compared with policyholders of
           traditional long-term care insurance policies, a higher percentage
           of Partnership policyholders purchased policies with more
           extensive coverage. In the four states, insurance companies are
           not allowed to charge policyholders higher premiums for policies
           with asset protection, and Partnership and traditional long-term
           care insurance policies with comparable benefits are required to
           have equivalent premiums.
		   
		   The Four States Require Partnership Policies to Include Certain
		   Benefits Not Required for Traditional Long-Term Care Insurance 
		   Policies

           In general, the four states with Partnership programs require that
           Partnership policies sold in their states include certain benefits
           that are not required for those states' traditional long-term care
           insurance policies. A state DOI official told us that they have
           these benefit requirements for Partnership policies in order to
           protect policyholders by helping to ensure that benefits are
           sufficient to cover a significant portion of their anticipated
           long-term care costs and to protect the Medicaid program by
           reducing the likelihood that policyholders will exhaust their
           benefits and become eligible for Medicaid.

           In addition to asset protection, which by definition Partnership
           policies include, all four states require Partnership policies to
           include inflation protection.^38 Three of the four Partnership
           states--California, Connecticut, and New York--require that
           Partnership policies include inflation protection that
           automatically increases benefit amounts by 5 percent annually on a
           compounded basis.^39 The four states do not require traditional
           long-term care insurance policies to include inflation protection,
           though insurance companies in these states are required to offer
           inflation protection as an optional benefit. While policies with
           inflation protection may include coverage that is more
           commensurate with expected future costs of care, these policies
           can be two or three times as expensive as policies without
           inflation protection. For example, in 2005 a long-term care
           insurance policy with a $200 daily benefit, a 3-year benefit
           period, and inflation protection cost about $3,000 per year for a
           60-year-old male; the same policy cost about $1,350 per year
           without inflation protection. An insurance company official told
           us that the additional cost of inflation protection is the primary
           reason individuals do not buy a Partnership policy.
		   
^38There are some exceptions to the inflation protection requirement. For
example, in New York, insurance companies are allowed to sell Partnership
policies to policyholders 80 years of age or older without inflation
protection.

39In Indiana, Partnership policies are required to include either
automatic compound inflation protection at 5 percent annually or in
accordance with the consumer price index, or an inflation protection
option that covers at least 75 percent of the average daily private pay
rate.

           The four states with Partnership programs also require minimum
           daily benefit amounts for all Partnership policies, while in three
           of the Partnership states, traditional long-term care insurance
           policies are not subject to this requirement.^40 According to
           Partnership and DOI officials in California and Connecticut,
           minimum daily benefit amounts are required for Partnership
           policies in order to prevent consumers from purchasing coverage
           that would be insufficient to cover a substantial portion of the
           cost of their care. According to Partnership program materials
           from New York, for example, the average daily cost of long-term
           care in a nursing facility in New York was about $263 per day in
           2004. Anything less than New York's 2004 minimum daily benefit
           amount of $171 for nursing facility care would therefore have
           required out-of-pocket payments for policyholders of more than
           one-third of the cost of their nursing facility care. In 2006, the
           required minimum daily benefit amounts for nursing facility care
           in Partnership policies ranged from $110 per day in Indiana to
           $189 per day in New York.

           In the four states with Partnership programs, Partnership policies
           are subject to minimum nursing facility benefit period
           requirements established by the states, but some traditional
           long-term care insurance policies are not subject to these same
           requirements. In California and Indiana, Partnership policies are
           required to have dollar coverage that provides for at least 1 year
           of care in a nursing facility, while traditional long-term care
           insurance policies are not subject to a minimum benefit period
           requirement.^41 In New York, Partnership policies are required to
           have minimum nursing facility benefit periods ranging from 18
           months to 4 years, depending on the type of coverage an individual
           purchases, while certain traditional long-term care insurance
           policies are required to have 1-year minimum nursing facility
           benefit periods. In Connecticut, Partnership and traditional
           long-term care insurance policies are both required to have 1-year
           minimum benefit periods for care provided in nursing facilities.
		   
40New York requires minimum daily benefit amounts for traditional
long-term care insurance policies.

41The minimum amount paid under a Partnership policy for this dollar
coverage can be no less than 70 and 75 percent of the average daily
private pay rate for nursing facilities in California and Indiana,
respectively.

           Partnership and traditional long-term care insurance policies both
           typically include elimination periods, which establish the length
           of time a policyholder who has begun to receive long-term care has
           to wait before receiving long-term care insurance benefits. The
           four states with Partnership programs limit the length of the
           elimination periods that can be included in Partnership policies.
           Two of the four states, Connecticut and New York, also generally
           limit the elimination period included in traditional long-term
           care insurance policies. In 2006, the elimination period for
           Partnership policies in California was no more than 90 days, while
           New York had a 100-day limit^42 and Indiana had a 180-day limit.
           In Connecticut, the elimination period limit for both Partnership
           and traditional long-term care insurance policies was 100 days.
           According to a New York Partnership program staff member, in New
           York the elimination period for traditional policies was generally
           no more than 180 days. The effect of increasing the elimination
           period is to increase the out-of-pocket costs policyholders incur
           in paying for their long-term care. One official from an insurance
           company that sells long-term care insurance policies told us that
           having long elimination periods could quickly deplete an
           individual's assets, which might make the asset protection under
           the Partnership program less valuable.

           Unlike traditional long-term care insurance policies, Partnership
           policies in the four states must cover or offer case management
           services.^43 Case management services can include providing
           individual assessments of policyholders' long-term care needs,
           approving the beginning of an episode of long-term care,
           developing plans of care, and monitoring policyholders' medical
           needs. According to a Partnership program official, by helping
           policyholders assess their medical needs and develop a plan of
           care, case management services can help policyholders use their
           benefit dollars efficiently. Partnership program officials in
           California, Connecticut, and Indiana explained that their states
           require that Partnership policies cover case management services
           provided through state-approved intermediaries that are
           independent of insurance company control. Partnership program
           officials in New York told us that Partnership policyholders have
           the option to seek case management services from independent case
           management service providers, but they can also elect to receive
           case management services from their own insurance company.
           Traditional long-term care insurance policies are not required to
           cover case management services, though some may offer them as an
           optional benefit. In addition, some insurance companies that sell
           traditional long-term care insurance policies may directly provide
           case management services.
		   
42This was for New York's total asset protection policies. The maximum
elimination period for New York's dollar-for-dollar policies was 60 days.

43In Connecticut and Indiana, the case management provision for
Partnership policies is specific to home and community-based services.

           Insurance companies in the four states with Partnership programs
           are subject to restrictions on the types of coverage they can
           offer in Partnership policies, while they are allowed to offer
           traditional long-term care insurance policies with more coverage
           options. In California, Connecticut, and Indiana, insurance
           companies can only offer Partnership policies with two types of
           coverage: an option that covers only nursing facility care, and a
           comprehensive option that covers nursing facility care as well as
           care provided in the home and in community-based facilities.^44 In
           New York, insurance companies may only offer Partnership policies
           that cover comprehensive care. The four states do not allow
           insurance companies to offer Partnership policies in their state
           that exclusively cover care provided in the home and in
           community-based facilities. However, in the four states, insurance
           companies can offer traditional long-term care insurance policies
           with nursing facility care only, home and community-based facility
           only, and comprehensive coverage options.
		   
44In California, Indiana, and New York, nursing facility coverage also
includes other settings that are similar to nursing facilities.
          
		   Partnership Policyholders Purchased Policies with Benefits That
		   Were More Extensive Than Those Purchased by Traditional
		   Policyholders Nationwide

           In the four states with Partnership programs, traditional
           long-term care insurance policies can include--and individuals can
           therefore choose to purchase--generally the same benefits found in
           Partnership policies.^45 However, Partnership policyholders tended
           to purchase benefits that are more extensive than those purchased
           by traditional long-term care insurance policyholders. We found
           that from 2002 through 2005, a higher percentage of Partnership
           policyholders purchased policies with more extensive coverage
           compared with policyholders who purchased traditional long-term
           care insurance nationally. Specifically, more Partnership
           policyholders purchased policies with higher levels of inflation
           protection and coverage that includes care in both nursing
           facility and home and community-based care settings. See table 1
           for a summary of the benefits purchased by Partnership and
           traditional long-term care insurance policyholders. For example,
           while all Partnership policyholders had policies from 2002 through
           2005 with the required inflation protection that generally
           increases daily benefit amounts by 5 percent annually, about 76
           percent of traditional long-term care insurance policyholders had
           policies with some form of inflation protection. Similarly, during
           this period, 64 percent of all Partnership policyholders had
           policies that included daily benefit amounts of $150 or greater,
           while 36 percent of traditional long-term care insurance
           policyholders nationwide had policies that provided daily benefit
           amounts at this level or greater. While these differences may
           reflect the benefit requirements found in Partnership policies,
           they may also reflect the incentive offered by the asset
           protection benefit of Partnership policies, which may influence
           consumers deciding whether to buy a Partnership or traditional
           long-term care insurance policy. The differences may also reflect
           the demographic and financial characteristics of the people living
           in the four states with Partnership programs relative to other
           states.
		   
45Traditional long-term care insurance policyholders cannot obtain asset
protection through their policies.		   

           Table 1: Percentage of Partnership and Traditional Long-Term Care
           Insurance Policyholders Purchasing Benefits from 2002 through 2005
		   
                                                   Traditional long-term care 
                               Partnership^a        insurance policyholders^b 
Inflation protection                                                       
Yes                                  100%                              76% 
No                                      0                               16 
Other^c                                 0                                8 
Daily benefit amount                                                       
Less than $100                          0                               11 
$100 to $149                           35                               53 
$150 to $199                           40                               25 
$200 and greater                       24                               11 
Benefit period                                                             
1 year                                  3                                3 
More than 1 and less than 3            13                               11 
years                                                                      
3 years                                37                               23 
More than 3 years but not              30                               37 
unlimited                                                                  
Lifetime/unlimited benefit             19                               26 
Elimination period                                                         
Less than 30 days                       3                                8 
30 days to 89 days                     23                               21 
90 days                                47                               60 
More than 90 days                      27                               11 
Coverage type                                                              
Comprehensive^d                        99                               91 
Nursing facility-only                   1                                3 
Other^e                                 0                                6 

           Sources: GAO analysis of the four states' UDS Partnership data and
           data provided by five insurance companies selling traditional
           long-term care insurance.

           Note: Percentages may not add to 100 due to rounding.

           aReported values for daily benefit amount, benefit period, and
           elimination period include nursing facility data, but not home
           care data.

           bApproximately 2 percent of people nationwide with long-term care
           policies have Partnership policies. Thus, although the data may
           include a number of Partnership policyholders, about 98 percent of
           these people are likely to have traditional long-term care
           insurance. Because this is only 2 percent, we consider this as a
           reasonable proxy for traditional long-term care policyholders.

           cIncludes policies with a future purchase option (7 percent) and
           policies with a deferred inflation option (1 percent). Enrollees
           who select a deferred inflation option may increase benefits at a
           later date that they choose.

           dComprehensive coverage insurance policies provide benefits for
           both nursing facility-only and home care services.

           eIncludes home care coverage.		   
		   
           Insurance Companies Cannot Charge Partnership Policyholders Higher
		   Premiums for Asset Protection, and Premiums for Partnership Policies
		   Must Be Equivalent to Premiums of Traditional Policies That Have
		   Comparable Benefits

           According to state officials, the four states with Partnership
           programs require Partnership and traditional long-term care
           insurance policies to have equivalent premiums if the benefits
           offered--except for asset protection--are otherwise comparable.
           According to information from one state's Partnership program, one
           reason for this requirement is that, unlike other insurance
           company benefits, insurance companies do not provide asset
           protection to Partnership policyholders. Instead, the four states
           with Partnership programs provide the asset protection benefit by
           allowing Partnership policyholders to protect some or all of their
           assets from Medicaid spend-down requirements. However, because
           Partnership policies are required to have inflation protection and
           other benefits that traditional long-term care insurance policies
           are not required to have, Partnership policies are likely to have
           higher premiums. According to a Connecticut state official, in
           1996, before the state required that Partnership and traditional
           long-term care insurance policies have equivalent premiums for the
           same benefits, Partnership policies were 25 to 30 percent more
           expensive than traditional long-term care insurance policies with
           comparable benefits. The official further explained that after the
           requirement was established, sales of Partnership policies in
           Connecticut more than tripled.
		   
		   Compared with Traditional Long-Term Care Insurance Policies, Two of
		   Four States Subject Partnership Policies to Additional Review, and
		   All Four States Require Additional Agent Training

           State officials told us that, while both Partnership and
           traditional long-term care insurance policies undergo reviews by
           the DOI in each of the four states with Partnership programs,
           Partnership policies in California and Connecticut also undergo
           another review by state Partnership program officials.^46,47
           California and Connecticut Partnership program staff review
           Partnership policies to determine whether the policies include the
           benefits mandated by Partnership regulations, and whether the
           insurance companies can meet additional data reporting and other
           administrative requirements. The programs' staff also try to
           ensure that the policies can be easily understood and contain all
           of the required language. The Partnership program offices in
           California and Connecticut perform their review of policies first,
           and then pass the application on to the DOI for further review.

           DOI officials in California and Connecticut told us that the
           Partnership office review of Partnership policies tends to be
           lengthier for insurance companies than the DOI review. A DOI
           official explained that when insurance companies add new benefit
           options to policies, the Partnership review can take longer. Other
           factors that may slow the Partnership review process include the
           time spent coordinating between the Partnership program and the
           state DOI, and the time it takes for insurance companies to learn
           how to complete the Partnership review process for the first time.
           State officials in Indiana and New York--where reviews of new
           Partnership policies are conducted by the DOI and not a separate
           Partnership program office--told us that it generally takes the
           same amount of time for Partnership and traditional long-term care
           insurance policies to pass through the review process.

           Before they can sell Partnership policies, insurance agents are
           subject to additional state training requirements compared with
           agents who sell only traditional long-term care insurance
           policies. Although each of the four states with Partnership
           programs has somewhat different requirements, in general the
           states require Partnership agents to undergo about a day of
           training specific to the Partnership program in addition to the
           training that the states require for those who sell traditional
           long-term care insurance.^48,49 Partnership program training
           typically includes information on topics such as long-term care
           planning, Medicaid, Medicare, the specific benefits required by
           the Partnership program, and how Partnership policies differ from
           traditional long-term care insurance policies. According to some
           state officials, agents need training on the Partnership program
           and Medicaid in order to understand the program and provide
           appropriate advice to their clients. In 2006, in three of the four
           states all Partnership program training was conducted in person,
           rather than via correspondence or on the internet; however, in New
           York agents completed an online internet-based course as well as
           classroom training as part of the Partnership program training.
           According to state officials, all four Partnership states require
           that the provider of this specialized Partnership training be
           approved by the state DOI, and in Connecticut, the training is
           provided exclusively by Partnership program staff.
		   
46The New York Partnership program does not conduct a review of
Partnership policies. The New York DOI reviews all Partnership and
traditional long-term care insurance policies.

47Until recently, the Indiana Partnership program was housed in the
Medicaid office and conducted an initial review of Partnership policies
prior to the DOI review. As of September 2006, the Indiana Partnership
program was housed in, and administered by, the DOI and there was only one
review of Partnership policies, which was conducted by the DOI.	   

           Despite the complexity of long-term care insurance products, DOI
           officials in three states with Partnership programs reported that
           long-term care insurance policies, including Partnership policies,
           garner few complaints from policyholders. For example, from 1998
           to 2005 the New York Insurance Department received an average of
           two to three complaints about Partnership policies each year
           (there were 51,262 active Partnership policies in the fourth
           quarter of 2005 in New York). During this time period, according
           to data from the New York state DOI, complaints about all
           long-term care insurance policies in New York related to issues
           such as the interpretation of policy provisions, premium amounts,
           and refusals to issue policies.
		   
48In order to continue to sell long-term care insurance in the four
Partnership states, insurance agents must receive several hours of
continuing education every 2 years. The required hours ranged from 5 hours
every 2 years in Indiana to 24 hours every 2 years in Connecticut.

49In New York, the continuing education credits from the required
Partnership policy training can be used to meet the DOI requirements for
agent recertification for traditional long-term care policies.

           Long-Term Care Insurance Policyholders Are Generally Wealthier
		   than Those Without Such Insurance, and Partnership Policyholders
		   Are Typically Younger than Traditional Long-Term Care Insurance
		   Policyholders

           Long-term care insurance policyholders--that is, both Partnership
           policyholders and traditional long-term care insurance
           policyholders--are more likely to have higher incomes and more
           assets than people without long-term care insurance. On average,
           Partnership policyholders are younger than traditional long-term
           care insurance policyholders. Those with long-term care insurance
           policies are also more likely to be female rather than male, and
           married than unmarried.
		   
           Long-Term Care Insurance Policyholders Generally Have Higher
		   Incomes and More Assets than Those Without Long-Term Care Insurance

           In examining Partnership policyholders in two states, traditional
           long-term care insurance policyholders nationwide, and those
           without long-term care insurance nationwide, we found that
           Partnership and traditional long-term care policyholders are more
           likely to have higher incomes than those without such
           insurance.^50 In California and Connecticut--the two states with
           Partnership programs for which we had data--at the time they
           purchased a policy, 55 percent of Partnership policyholders over
           age 55 had monthly household incomes of $5,000 or greater. In
           comparison, 43 percent of all households with people over age 55
           in these states had monthly household incomes at this level at the
           time they were surveyed.^51,52 Similarly, at the national level,
           when surveyed, 46 percent of traditional long-term care
           policyholders over age 55 had monthly household income of $5000 or
           greater, whereas 29 percent of those individuals over age 55
           without long-term care insurance had such incomes.^53 We also
           found that more than half (53 percent) of Partnership
           policyholders had household assets of $350,000 or more in
           California and Connecticut. Data on the asset levels of all
           households in those states were not available for our comparison.
           Nationwide, 36 percent of traditional long-term care insurance
           policyholders and 17 percent of people without long-term care
           insurance had household assets exceeding $350,000 (see table 2).
		   
50Data from Indiana and New York are excluded from our income and asset
comparisons. New York did not collect income or asset data for its Partnership
program, while Indiana income and asset data were not detailed enough to make
comparisons with other states.

51Income data for Partnership policyholders in Connecticut were from 2002
through 2005. Income data for Partnership policyholders in California were
from 2003 to 2004. Data for all households in those two states were from
2004. We combined multiple years of these data in order to increase the
sample size.

52Because we did not have a direct measure of the population without
long-term care insurance, we used the general population of all households
as a proxy. Nationally, about 12 percent of the population over age 55 has
long-term care insurance. Therefore we assume that the income information
from all households in two states with Partnership programs-California and
Connecticut-largely reflects the income and asset patterns of people
without long-term care insurance.

53The national-level data are from 2004.
		   
Table 2: Household Income and Household Asset Distribution among
Partnership Policyholders and Comparison Populations in Partnership States
and Nationally

                       Partnership states: CA and                    
                                  CT^a                     All states^b
                                                                         Those 
                                                         Traditional   without 
                          Partnership           All   long-term care long-term 
                     policyholders^c, households^e,        insurance      care 
                                    d             f  policyholders^g insurance 
Monthly household income ranges^h                                           
<$1000                          1%            8%               4%       15% 
$1000-$4999                     45            49               50        56 
$5000 or greater                55            43               46        29 
Household asset                                                             
ranges^i, j                                                                 
<$100,000                      16% Not Available              36%       62% 
$100,000-$199,999               14 Not Available               14        12 
$200,000-$349,999               17 Not Available               14         9 
$350,000 or                     53 Not Available               36        17 
greater                                                                     

Sources: GAO analysis of Partnership program purchaser surveys, American
Community Survey (ACS), and the HRS.

Note: Percentages may not add to 100 due to rounding.

aDoes not include data from New York and Indiana.

bData for all states are from the HRS, 2004.

cConnecticut values are based on survey data from 2002 through 2005.

dCalifornia values are based on survey data from 2003 and 2004.

eData are from the ACS, 2004.

fWe use the All Households category as a proxy for those without long-term
care insurance in California and Connecticut. Approximately 12 percent of
people nationwide over 55 have long-term care insurance so our measure is
likely to contain approximately 88 percent without long-term care
insurance.

gApproximately 2 percent of people nationwide with long-term care policies
have Partnership policies. Thus, although the HRS data may include a small
number of Partnership policyholders, about 98 percent of these people
likely have traditional long-term care insurance.

hData for monthly income ranges are for survey respondents aged 55 and
over.

iData for asset ranges are for all survey respondents regardless of age.

jIn the policyholder surveys, California and Connecticut instructed
policyholders to exclude the value of homes and cars when reporting their
assets. The HRS data for assets also exclude homes and vehicles.

Partnership Policyholders Are Younger on Average than Traditional Long-Term Care
Insurance Policyholders and People Without Long-Term Care Insurance

In our analyses, we found that Partnership policyholders in California,
Connecticut, Indiana, and New York are younger on average than traditional
long-term care insurance policyholders nationally and those without
long-term care insurance nationally (see table 3). We also found that
those who purchase long-term insurance policies--both traditional and
Partnership--are more likely to be women than men, and married than
unmarried.^54

Table 3: Demographic Characteristics of Partnership Policyholders and
Comparison Populations in Partnership States and Nationally

                                                               People without 
                  All partnership   Traditional long-term care long-term care 
               policyholders^c, e insurance policyholders^a, d    insurance^d 
Average age                 59                         63^c           64^b 
Age range               18-104                      30-95^c       24-107^b 
Age                                                                        
categories                                                                 
Under 55                   28%                          20%            21% 
years                                                                      
55-64 years                 49                           37             36 
65-74 years                 19                           30             22 
75 years &                   3                           13             20 
over                                                                       
Sex                                                                        
Female                     58%                          56%            54% 
Male                        42                           44             46 
Marital                                                                    
status                                                                     
Married                    75%                          72%            62% 
Not married                 24                           28             38 
Unknown                      1                            0              0 

Source: GAO analysis of the UDS and the HRS.

Note: Percentages may not add to 100 due to rounding.

aApproximately 2 percent of people nationwide with long-term care policies
have Partnership policies. Thus, although the HRS data may include a small
number of Partnership policyholders, about 98 percent of these people are
likely to have traditional long-term care insurance.

bDenotes average age at time of survey.

cData are as of time of purchase.

dData are from the 2004 HRS, which is a longitudinal national panel survey
of individuals over age 50.

54To make this comparison, we used cumulative data from the 2002 through
2005 UDS data sets on Partnership policyholders and data from the 2004 HRS
survey.

eData are from the UDS for 2002 through 2005.

Partnership Programs Unlikely to Result in Savings for Medicaid Largely Because
of the Asset Protection Benefit of Partnership Policies

Surveys conducted in some states with Partnership programs and our
illustrative financing scenarios together suggest that in the four states
with Partnership programs, the programs are unlikely to result in Medicaid
savings and could result in increased Medicaid spending. Survey data show
that in the absence of a Partnership program in their state, 80 percent of
Partnership policyholders would have purchased a traditional long-term
care insurance policy and may represent a potential source of increased
spending for Medicaid. Data are not yet available to determine the extent
to which the 20 percent of individuals who would have self-financed their
care will access Medicaid in the absence of a Partnership program.
However, our scenarios suggest that an individual could self-finance care
and delay Medicaid eligibility for about the same amount of time as he or
she would have with a Partnership policy, although we identify some
circumstances that could delay or accelerate the time to Medicaid
eligibility. Because of the amount of insurance Partnership policyholders
generally purchase and their typical income and assets, few Partnership
policyholders are likely to ever become eligible for Medicaid, which
suggests that the Partnership programs are likely to have a small impact
on Medicaid spending.

Most Partnership Policyholders Would Have Purchased Traditional Long-Term Care
Insurance in Absence of Partnership Program, Suggesting an Increase in Medicaid
Spending

The four Partnership programs are unlikely to result in savings for their
state Medicaid programs and may result in increased Medicaid spending.^55
Based on surveys of Partnership policyholders conducted by state
Partnership programs in California, Connecticut, and Indiana, we estimate
that, in the absence of a Partnership program in their state, 80 percent
of Partnership policyholders would have purchased traditional long-term
care insurance policies instead, while the other 20 percent would have
self-financed their care.^56 To assess the impact Partnership programs may
have on Medicaid savings in the four states with Partnership programs, we
explored, under three different illustrative financing scenarios and using
certain assumptions, how long it would take before an individual using a
Partnership policy would become eligible for Medicaid and how long--in the
absence of a Partnership program--it would take for the same individual to
become eligible for Medicaid using the other two financing options
depicted in the scenarios. Our financing scenarios indicate that with a
Partnership policy, an individual with assets and benefits typical of many
policyholders becomes eligible for Medicaid sooner than if the individual
financed his or her long-term care with a traditional long-term care
policy. Because a Partnership policy, unlike a traditional long-term care
insurance policy, exempts the individual in the scenario from spending his
or her protected assets on long-term care before the individual becomes
eligible for Medicaid, the individual with a Partnership policy becomes
eligible for Medicaid sooner than if the individual had a traditional
policy, which is likely to increase the amount of time Medicaid finances
the individual's long-term care. The scenarios also suggest that if the
individual would have self-financed his or her long-term care in the
absence of the Partnership program, the individual would become eligible
for Medicaid at about the same time as he or she would have with a
Partnership policy.

55This is consistent with CBO's estimate that repealing the moratorium on
new Partnership programs could increase Medicaid spending.

56The results for the individual states were 84 percent, 76 percent, and

57 percent for California, Connecticut, and Indiana, respectively. Using
the number of respondents in each state to weight the calculation, the
average for the three states combined was approximately 80 percent.

The three financing scenarios we compared were

           o financing using a Partnership policy,

           o financing using a traditional long-term care insurance policy,
           and

           o self-financing without any long-term care insurance.

           For illustrative purposes, our scenarios are based on an
           individual with assets that are typical of many of those who have
           long-term care insurance--that is, an individual who holds assets
           of $300,000.^57 In two of our scenarios, the individual holds
           long-term care insurance benefits of $210,000, which will cover a
           nursing facility stay of about 3 years--the average nursing
           facility stay is between 2 and 3 years. We also make several
           simplifying assumptions, such as that the individual is not
           overinsured (i.e., does not have insurance that exceeds the value
           of the individual's assets) and is unmarried at the time long-term
           care is required.

           Specifically, scenario A (see fig. 1) depicts a Partnership
           policyholder with $300,000 in assets who purchases a policy valued
           at $210,000 (worth about 3 years of nursing facility coverage),
           automatically receiving $210,000 in asset protection. When the
           individual requires long-term care, the Partnership policy will
           pay for the first $210,000 worth of care--the total amount of his
           or her insurance benefits. After these Partnership benefits have
           been exhausted, the individual will have to spend down the $90,000
           of unprotected assets on long-term care and then, assuming the
           individual meets state Medicaid income eligibility requirements,
           Medicaid will begin to finance the individual's long-term care. As
           depicted by scenario B, if this same individual purchases a
           traditional long-term care insurance policy worth $210,000 instead
           of the Partnership policy, insurance will pay for the first
           $210,000, and the individual will then have to spend down the
           unprotected assets--all $300,000--before he or she is eligible for
           Medicaid.^58,59 Scenario C describes how this same individual
           would finance his or her long-term care without any long-term care
           insurance. As scenario C shows, if the individual had $300,000 in
           assets, these would have to be spent before the individual would
           be eligible for Medicaid.^60 In both this scenario and in the
           scenario in which the individual owns a Partnership policy,
           Medicaid begins paying for the individual's long-term care at
           about the same time, with the difference being whether long-term
           care costs prior to Medicaid eligibility are paid by long-term
           care insurance or by the individual.

57For example, 53 percent of Partnership policyholders in California and
Connecticut had household assets of $350,000 or more. Approximately 37
percent of Partnership policyholders purchased policies with a 3-year
benefit period.

58To qualify for Medicaid, individuals must meet a number of requirements,
including their state's allowable asset limitation, excluding the amount
of protected assets due to the Partnership policy. For 2006, these were
$2,000 in California, $1,600 in Connecticut, $1,500 in Indiana, and $4,150
in New York. The situation is more complicated when the person has a
spouse. For instance, regarding assets, when someone in an institution
applies for Medicaid and they are married, Medicaid looks at all the
assets of the couple, regardless of ownership (certain items such as the
couple's home, personal and household property, one vehicle, and a small
amount set aside for burial, are excluded). One half of the remaining
countable assets, up to a maximum of approximately $100,000 in 2007, are
then protected for the community spouse. Any remaining assets are then
used to determine Medicaid eligibility for the spouse in the institution.

59Individuals may choose a different set of benefits, depending on whether
they select a traditional or Partnership program policy. In order to
simplify our comparison of scenarios A and B, we assume that the benefits
of a Partnership policy and a program traditional long-term care policy
are the same, except for the asset protection benefit of the Partnership
policy.

60More than half of all Partnership policyholders in California and
Connecticut combined reported assets of at least $350,000, which is more
than the national average of about $210,000 that would be needed to pay
for a 3-year stay in a nursing facility--the average stay being between 2
and 3 years.

Figure 1: Financing of Long-Term Care Nursing Facility Stays Under Three
Scenarios

Note: To simplify our scenarios, we made some simplifying assumptions,
such as, the individual depicted in the scenarios has assets and benefits
that are typical of many individuals with long-term care insurance; the
individual is unmarried; and the individual has assets that are greater
than or equal to the value of the individual's Partnership policy. Our
results do not depend on the level of assets or the amount of insurance
dollars, provided the amount of insurance dollars does not exceed the
amount of assets. Appendix II further discusses the effects of changing
these assumptions.

We found some circumstances when adjusting the assumptions underlying our
scenarios resulted in delaying or accelerating Medicaid eligibility, but
most did not change the outcomes related to Medicaid savings. For example,
to construct our scenarios, we assumed an individual who had $300,000 in
assets, $210,000 in insurance coverage, and who used this coverage for
long-term care that cost about $70,000 per year. When we changed these
amounts--as long as the amount of insurance coverage did not exceed the
amount of assets--the scenarios still showed that the individual became
eligible for Medicaid sooner with a Partnership policy than with a
traditional policy, and became eligible for Medicaid at the same time with
a Partnership policy and self-financing.

Our scenarios also assumed that the individual with a Partnership policy
or a traditional long-term insurance policy was not overinsured--that is,
had more insurance coverage than the value of his or her assets. When we
modified this assumption, we found that one portion of our finding still
held true--the individual in the scenarios using the Partnership policy
still became eligible for Medicaid sooner than he or she did using a
traditional long-term care insurance policy. However, the individual also
became eligible for Medicaid later using the Partnership policy than when
the individual self-financed his or her own long-term care. This suggests
that if individuals overinsure their assets, those who finance their
long-term care using Partnership policies could represent a source of
savings for Medicaid when compared with those who self-finance their care.
However, the number of policyholders that this applies to is unlikely to
be large enough to offset the number of Partnership policyholders who
represent a potential source of increased Medicaid spending. While we do
not have information about the amount of assets that Partnership
policyholders have at the time they use their benefits, survey data from
California and Connecticut indicate that when Partnership policyholders
purchased their policies, they tended to purchase policies that were equal
to or lower than the value of their household assets. This suggests that
most individuals are unlikely to overinsure their assets at the time of
purchase, though their status could change over time.^61 In California and
Connecticut combined, in 2004, 53 percent of Partnership policyholders had
at least $350,000 worth of household assets at the time of purchase, while
only about 32 percent of these Partnership policyholders have more than 5
years of coverage equal to about $350,000.

Our scenarios also depicted an unmarried individual. While most
Partnership policyholders are married when they purchase a Partnership
policy, by the time most individuals require long-term care services, they
are unmarried. Our analysis of 2004 HRS data of individuals entering a
nursing facility who are age 65 or older showed that about 66 percent are
widowed, and more than 75 percent are not married. However, there are
likely some individuals who will be married when they require long-term
care services. In general, after applying the Medicaid spousal exemption,
if the individual's assets remain higher than the value of his or her
insurance, being married does not change the result that compared with a
Partnership policy, the individual's time to attain Medicaid eligibility
is accelerated with a traditional policy and is the same as with
self-financing.^62 However, if the amount of the Medicaid spousal
exemption brings the individual's eligible assets below the value of the
insurance policy, then the individual would fall into an overinsured
category. Being overinsured means the individual would become a source of
savings for Medicaid; however, this only applies to the 20 percent of
individuals who would have self-financed their care in the absence of a
Partnership program. The 80 percent of individuals who would have
purchased a traditional policy still represent potential increased
spending, whether they are overinsured or not.

61Household assets may be jointly owned by a couple. In order for assets
to be fully protected for the couple, both individuals need to have their
own Partnership policies that insure all eligible assets because many
individuals are no longer married by the time they require long-term care
services, and the asset protection associated with a Partnership policy is
not transferable. We do not consider a married couple to be overinsured
when both individuals have long-term care insurance policies that are
worth the value of their estate.

We also explored what would occur if we modified our assumption that an
individual is equally likely to transfer assets in all three of our
scenarios. We found that if the individual who would have self-financed
care transfers his or her assets, it would likely take less time for the
individual to become eligible for Medicaid than it would with a
Partnership policy. This could result in some savings to Medicaid for
those individuals who purchase Partnership policies instead of
transferring assets. We also found that for an individual who would have
purchased traditional insurance, the amount of assets transferred would
have to be at least as much as the value of the insurance policy purchased
in order for the Partnership program to result in Medicaid savings. While
we do not know how many individuals would have transferred assets in the
absence of the Partnership program, one of our recent reports suggests
that asset transfers may not be that prevalent. In March 2007, we reported
that few applicants who were approved for Medicaid coverage ultimately
transferred assets.^63 In addition, the asset transfer standards
established under DRA increased the look-back period to 5 years, which
reduces the opportunity for individuals to transfer assets to establish
Medicaid eligibility.

62The Medicaid spousal exemption, also known as the community spouse
resource allowance, permits the spouse remaining in the community to
retain an amount equal to one-half of the couple's combined countable
assets, up to a state-specified maximum level. In 2007, the federal
maximum was $101,640; that is, states were allowed to set their community
spouse resource allowance equal to a value no greater than this amount.
Medicaid eligibility of the institutionalized spouse is determined using
the remaining assets.

63This finding was based on our review of the prevalence of asset
transfers among 465 approved applicants in three states. See
[42]GAO-07-280 .

Overall, our scenarios suggest that in the aggregate the savings potential
from the Partnership programs of the 20 percent of individuals who would
have self-financed their care is outweighed by the 80 percent of
individuals who will likely result in increased Medicaid spending. For
more information on our simplifying assumptions and the impact of
adjusting these assumptions on our findings, see appendix II.

Few Partnership Policyholders Are Likely to Become Eligible for Medicaid,
Limiting the Impact on Medicaid Expenditures

Although our survey data and scenarios show that about 80 percent of
Partnership policyholders who become eligible for Medicaid are likely to
do so sooner than they otherwise would have without a Partnership program,
we also expect that few Partnership policyholders will actually become
eligible for Medicaid and turn to the program to finance their long-term
care. There are two reasons for this expectation. First, most Partnership
policyholders purchase policies that are likely to cover all or most of
their long-term care expenses during their lifetimes, thereby reducing the
likelihood that the policyholders will require financing from Medicaid for
their long-term care. We found that 86 percent of Partnership
policyholders had benefits covering 3 or more years, while the average
nursing facility stay lasts between 2 and 3 years. One study of
traditional long-term care insurance policyholders with lifetime benefits
found that only about 14 percent of policyholders used their benefits for
more than 3 years, and fewer than 5 percent of all policyholders used
their benefits for more than 5 years. These data suggest that if
Partnership policyholders continue to purchase policies with benefit
periods that cover their long-term care needs, the percentage of
Partnership policyholders who exhaust their benefits and then become
eligible for Medicaid is likely to be limited. While some experts have
reported that there is a recent trend for traditional long-term care
insurance policies to be sold with shorter benefit periods, the minimum
benefit requirements that applied to Partnership policies could result in
Partnership benefits remaining more stable over time.

The second reason we estimate that few Partnership policyholders are
likely to turn to Medicaid for their long-term care financing is that, in
general, Partnership policyholders have incomes that exceed Medicaid
income eligibility thresholds. Although Partnership policyholders can
purchase varying amounts of asset protection, they must still meet state
Medicaid income thresholds in order to become eligible for Medicaid. In
2006, the monthly income eligibility thresholds for all states were
required to be no higher than 300 percent of the Supplemental Security
Income standard, which was $1,809 in 2006.^64 However, only 1 percent of
the Partnership policyholders in California and Connecticut had household
incomes less than $1,000 per month at the time they purchased their
long-term care insurance policies. Our analysis of HRS data also indicates
that wealthy individuals continue to have a high level of assets^65 at the
time they are admitted to a nursing facility, which suggests that many
Partnership policyholders will continue to be relatively wealthy and
unlikely to meet Medicaid eligibility thresholds, even at the time they
enter a nursing facility. For example, of all people who entered a nursing
facility in 2004, the average asset value for the 25 percent of people
with the highest assets was over $334,000 in 1992, and by 2004, 12 years
later, their assets had grown to almost $430,000. Similarly, the average
monthly income for the 25 percent of people with the highest incomes who
were admitted to a nursing facility in 2004 was about $5,600 in 1992, and
about $3,700 in 2004--more than double the threshold for Medicaid
eligibility in any of the four states with Partnership programs.

The income levels of Partnership policyholders may reflect the fact that
the cost of purchasing a long-term care insurance policy--including a
Partnership policy--may exceed what most elderly households can afford.
According to guidelines published by the NAIC, a person should spend no
more than 7 percent of his or her income on long-term care insurance. A
traditional long-term care insurance policy covering 3 years of care, with
inflation protection, a $200 daily benefit allowance, and comprehensive
coverage, costs about $3,000. In order to afford such a policy, an
individual would need an annual income of about $43,000. However, data
from the 2004 HRS show that about half of elderly households nationwide
had annual incomes below $43,000. A survey of Connecticut Partnership
policyholders suggested that cost was the most important factor in
policyholders' decision to let their policies lapse. Sixty-two percent of
surveyed individuals in Connecticut who let their Partnership policy lapse
said that they dropped their Partnership policy because it was too
costly.^66

64Specifically, the Medicaid eligibility thresholds in the four states
with Partnership programs were $600 in California, $619 in Indiana, $1,809
in Connecticut, and $692 in New York in 2006. If individuals were in a
nursing facility, they were permitted to keep a personal allowance amount
to cover incidental purchases in the nursing facility. The personal
allowances for individuals in nursing facilities in 2006 were $35 in
California, $52 in Indiana, $61 in Connecticut, and $50 in New York.

65In this particular example, our criterion for being among the wealthiest
people is those people whose assets are in the highest 25 percent.

As of 2006, few Partnership policyholders in the four states with
Partnership programs had accessed Medicaid to finance their long-term
care. Of the approximately 218,000 Partnership policies sold since the
program was first introduced in the late 1980s, approximately 190,000 were
still active as of August 2006. In addition, as of that same date, a total
of 3,454 Partnership policyholders--less than 2 percent of all Partnership
policyholders--have accessed long-term care benefits since the Partnership
programs began. Of that group, 292 Partnership policyholders exhausted
their long-term care insurance benefits, and 159
policyholders--approximately 54 percent of those who exhausted their
benefits--subsequently went on to access Medicaid benefits. The number of
Partnership policyholders who access benefits and also access Medicaid is
likely to grow, because people typically use long-term care services 15 to
20 years after they purchase a policy, and the first Partnership policies
were established less than 20 years ago. We do not know why some of the
292 individuals who exhausted their long-term care insurance benefits did
not access Medicaid. It is possible that their income was higher than
Medicaid eligibility thresholds, or they may have had unprotected assets
that they had to spend down. Alternatively, they may have preferred to
self-finance their care, they may have died, or they may have stopped
using long-term care services.

Concluding Observations

With DRA authorizing all states to implement Partnership programs,
information on the Partnership policies and policyholders from the four
states with Partnership programs may prove useful to other states
considering implementing such programs. In particular, states may want to
consider the trade-offs that come with implementing a Partnership program.
First, a Partnership program's potential impact on Medicaid expenditures
should be considered. Based on our scenario comparison and survey data, we
anticipate that Partnership programs in California, Connecticut, Indiana,
and New York are unlikely to result in savings for their state Medicaid
programs and could result in increased Medicaid expenditures. This is
largely due to the modifications of state Medicaid eligibility
requirements states have to make in order to offer asset protection to
Partnership policyholders and survey data showing that the majority of
Partnership policyholders would have purchased traditional long-term care
insurance had the Partnership program not existed. However, given the
amount of long-term care insurance benefits and income and asset levels of
current Partnership policyholders, we also anticipate that relatively few
policyholders will access Medicaid in the four states with Partnership
programs. Therefore, the impact of Partnership programs on state Medicaid
programs will likely be small.

66It is possible that Partnership policyholders with higher incomes could
meet Medicaid income thresholds because the four states with Partnership
programs allow individuals to deduct medical expenses from their income
when determining Medicaid eligibility. However, the individuals would
still need to contribute their income toward the cost of care. Therefore,
this limits Medicaid's liability for individuals with higher incomes.

While Partnership programs are not likely to reduce states' Medicaid
expenditures, the programs do offer some benefits to some consumers. The
asset protection feature, which states require Partnership policies to
offer at no additional premium cost, can benefit policyholders who exhaust
their Partnership benefits and who access Medicaid. Even if individuals do
not end up using their Partnership insurance or Medicaid, the availability
of asset protection may provide peace of mind for those who fear the risk
of having to spend their assets on their long-term care. However, states
that implement Partnership programs should recognize that, because of
their cost, Partnership policies generally do not benefit all consumers.
The cost of annual premiums for long-term care insurance may not be
affordable to individuals with moderate incomes, and as a result long-term
care insurance policyholders, including Partnership policyholders, tend to
be wealthier than those without such insurance.

Agency and State Comments and Our Evaluation

We received written comments on a draft of this report from HHS (see
appendix III) and from the four states with Partnership programs,
California, Connecticut, Indiana, and New York (see appendix IV).

HHS commented that the results of our study should not be considered
conclusive because the results do not adequately account for the effect of
estate planning efforts such as asset transfers. Specifically, HHS was
concerned that the simplified scenarios were flawed in that they did not
account for individuals who engage in estate planning activities prior to
expending all of their own funds on long-term care costs. HHS further
noted that the data sources used in our report would not likely yield
accurate data on asset transfers and criticized the report for not
incorporating a review of the literature on this issue and reporting on
analyses of the experience of the four states with Partnership programs.
The four states with Partnership programs disagreed with our conclusion
that the Partnership programs are unlikely to result in Medicaid savings
and, like HHS, commented that our scenarios did not adequately account for
the impact of asset transfers. California, Connecticut, and New York
raised concerns about our methodology for estimating the financial impact
of the Partnership program on Medicaid. California and Connecticut noted
that we had excluded two Partnership policyholder survey questions from
our analysis that they consider in their own analysis of the Partnership
program.

We maintain that the evidence suggests that the Partnership program is
unlikely to result in savings for Medicaid, despite limited data and
program experience. We agree with HHS and the four states with Partnership
programs that Medicaid savings could result from those individuals who
would have transferred assets in the absence of the Partnership program.
However, our scenarios suggest that the savings associated with asset
transfers are likely to offset the potential costs associated with
policyholders who would have purchased traditional long-term care
insurance in the absence of the Partnership programs. Further, the
assumptions used by California, Connecticut, and Indiana to predict
savings could underestimate the percentage of Partnership policyholders
that represent a cost to Medicaid and overestimate the percentage that
represent savings to Medicaid. We did not provide an overview of the
literature on asset transfers in our draft report because, as we noted in
our March 2007 report, the evidence on the extent to which individuals
transfer assets to become financially eligible for Medicaid coverage for
long-term care is generally limited and often based on anecdote.^67 We did
not comment on states' analyses of their experience with the Partnership
programs because, according to our analysis, their methodology overstates
potential savings and understates potential costs.

  Impact of Asset Transfers on Medicaid

In appendix II of our draft report we acknowledged that some savings could
result for Medicaid if, in the absence of a Partnership program, an
individual would have self-financed his or her long-term care and
transferred assets. We also acknowledged how a Partnership program can
result in Medicaid savings if, in the absence of the Partnership program,
an individual would have purchased a traditional long-term care insurance
policy and transferred assets that were at least equal to the value of the
traditional long-term care insurance policy. However, our analysis
suggests that these savings would be limited to those individuals who,
prior to requiring long-term care, would have transferred assets to become
eligible for Medicaid in the absence of the Partnership program. Further,
the larger percentage of policyholders who represent a potential cost to
Medicaid are likely to offset savings attributable to asset transfers.

67See [43]GAO-07-280 .

While the literature on the extent of asset transfers is generally limited
and anecdotal, in March 2007, we published a report that included an
analysis of asset transfers by nursing home residents using HRS data. We
complemented that analysis by examining a sample of Medicaid applications
in three states to identify the extent of asset transfer activity.^68 Both
of these analyses suggested that about 10 to 12 percent of individuals
transferred assets before applying for Medicaid, and the median amount
transferred based on analysis of the HRS data and state Medicaid
applications was $1,239 and $15,152, respectively.^69 The relatively low
incidence of asset transfers and the small amounts transferred relative to
the costs associated with long-term care suggest that the impact of asset
transfers on Medicaid may be limited. While the results of this study are
not specific to Partnership policyholders, we found no other credible
evidence suggesting that Partnership policyholders would transfer
sufficient assets to offset the costs to Medicaid associated with the
large number of individuals who would have purchased traditional long-term
care insurance in the absence of the Partnership program. Also, although
the overall impact of DRA on Medicaid eligibility is uncertain, DRA
reduces the opportunity for people to transfer assets in order to become
Medicaid eligible by increasing the period Medicaid programs can
"look-back" at an individual's assets to 5 years. In response to HHS'
comments about asset transfers, we have amended our draft report to make
the discussion of asset transfers more prominent in the body of our report
and to include reference to our March 2007 study.

  Methodology for Assessing Medicaid Savings

California, Connecticut, and New York raised concerns about our
methodology for estimating the financial impact of the Partnership program
on Medicaid. California and Connecticut noted that we had excluded two
Partnership policyholder survey questions from our analysis that they
consider in their own analysis of the Partnership program. These questions
asked Partnership policyholders whether they would have transferred assets
to become eligible for Medicaid in the absence of the program and whether
the Partnership program influenced their decision to buy long-term care
insurance.^70

68See [44]GAO-07-280 .

69The HRS analysis was based on transfers during the 4 years prior to
nursing home entry by elderly nursing home residents who were
Medicaid-covered. The analysis of a sample of Medicaid applications in
three states was based on transfers during the 3-year look-back period by
approved Medicaid applicants.

We maintain that our methodology is sound and that the methodology
California, Connecticut, and Indiana use underestimates the potential for
Medicaid costs and overestimates the potential for Medicaid savings. We
relied on a question that asked Partnership policyholders whether they
would have purchased traditional long-term care insurance in the absence
of the Partnership program.^71 We disagree with California, Connecticut,
and Indiana regarding the appropriateness of including additional survey
information because of concerns about ambiguous wording and these states'
assumption that policyholders' responses can be used to predict the
likelihood of future asset transfers. We did not present the states'
analyses for evaluating Medicaid spending in our draft report because we
believe the states' analyses overstate potential savings and understate
potential costs.

In our analysis, we estimated that about 80 percent of policyholders would
have purchased traditional long-term care insurance in the absence of the
program, and we estimated that these individuals generally represented a
potential cost to Medicaid. Our 80 percent estimate was based on analysis
of the survey question about how Partnership policyholders would have
financed their long-term care in the absence of the Partnership program.
The methodology that California, Connecticut, and Indiana use to estimate
potential costs is based on a policyholders' response to the following
criteria, obtained from three survey questions:

           1. The policyholder would have purchased traditional insurance in
           the absence of the Partnership program;
           2. the Partnership program had no influence on the policyholders'
           decision to purchase insurance; and
           3. the policyholder would not have transferred assets in the
           absence of the Partnership program.

70One survey question asks: "Did the [partnership program] influence your
decision to purchase long-term care insurance? (yes or no)." The other
asks: "Why did you decide to purchase long-term care insurance?" Eight
response options are provided, one being "As an alternative to
transferring assets to qualify for Medicaid."

71The California, Connecticut, and Indiana surveys asked: "Would you have
purchased long-term care insurance in the absence of the Partnership?"
(yes or no).

By adding the two additional criteria to determine whether an individual
represents a potential cost to Medicaid, the states' estimate of the
percentage of policyholders who fell into this category was more
restrictive than ours. We have several concerns with the wording of the
survey questions used to define the additional two criteria. In addition,
according to our analysis, the criteria that define costs are not
correctly specified because there are some circumstances when the second
criterion would represent a cost to Medicaid whether or not the
Partnership program had an influence on the policyholder's decision to
purchase insurance.

The wording of the survey question about whether the Partnership program
influenced the policyholder's decision to purchase insurance was not
specific with regard to how the decision was influenced. In particular,
the Partnership programs' influence could have been to influence the
policyholder to purchase a different benefit package, to change the timing
of the policyholder's purchase, or to change the policyholder's decision
to purchase at all. Given the ambiguity of the question, it is not clear
how a response should be interpreted. Moreover, how this question is
interpreted could influence the outcome of an analysis of the likely
impact of the Partnership program on Medicaid spending. Our analysis
suggests that even if the Partnership program influenced policyholders to
purchase enhanced benefits, the Partnership program still represented a
potential cost, just a smaller cost. Adding this criterion incorrectly
narrows the number of policyholders who represent potential costs to
Medicaid.

We also disagree with the states' assumption that policyholders' responses
to the asset transfer question can be used to approximate the extent to
which individuals would or would not transfer all of their assets in the
future and--in the absence of the program--become eligible for Medicaid.
The respondents were asked about events that are unlikely to occur for 15
to 20 years, and to speculate on what their actions would be in the future
if there was no Partnership program. California, Connecticut, and New York
reported that about 25 percent of respondents said they would have
transferred assets to become eligible for Medicaid. All of these
individuals are excluded from the pool of policyholders who represent a
potential cost to the program in the state cost estimates. The assumption
that all of these individuals would have transferred all of their assets
is inconsistent with our March 2007 report regarding the incidence of
asset transfers and amount of assets transferred for the purposes of
becoming eligible for Medicaid. We agree that some individuals would have
transferred assets in the absence of the program, but do not agree that
this question provides an adequate measure of the extent to which it
occurs. Therefore, we believe using the responses to this question may
overstate the extent to which respondents would actually transfer all of
their assets.

We have similar concerns with the methodology California, Connecticut, and
Indiana used to estimate savings, because it is based on the same three
questions, two of which we view as inadequate. We assumed in our scenarios
that individuals who would have self-funded their long-term care without
insurance were likely to be budget neutral, but acknowledged there were
several circumstances that would cause these individuals to become a
potential source of savings. However, we did not attempt to quantify the
percentage who would become a source of savings because of data
limitations and because the savings were likely to be outweighed by the
larger percentage of policyholders who likely represented a cost to the
program. In contrast, California, Connecticut, and Indiana consider a
policyholder to represent savings if:

           1. the policyholder would have purchased a Partnership policy as
           an alternative to transferring assets; and
           2. (a) the Partnership program influenced their decision to
           purchase insurance, or (b) the policyholder would not have
           purchased long-term care insurance in the absence of the
           Partnership program.

As we noted in the discussion above regarding the states' methodology for
estimating Medicaid costs, we disagree with the reliance on the asset
transfer question as a measure of the extent to which individuals would
have transferred assets. We also believe it was incorrect to predict
Medicaid savings for those respondents who said the Partnership program
influenced their decision to purchase a policy. Even if the Partnership
program influenced the policyholder's decision to purchase enhanced
benefits, our analysis suggests the Partnership program would not result
in savings but would rather result in a reduction of costs to Medicaid.

New York commented that our analysis was not applicable to their state.
They cited preliminary results of a 2006 survey that estimated the number
of recent Partnership policyholders who would have financed their care
with a traditional policy in the absence of the Partnership program. Their
estimates were considerably lower than the 80 percent we estimated based
on our results from California, Connecticut, and Indiana. New York used
different questions in their survey than California, Connecticut, and
Indiana.^72 As such, their results were not comparable to those of the
other states. We believe New York's question was less direct for the
purposes of our analysis than the question used by California,
Connecticut, and Indiana in their survey of Partnership policyholders. New
York's question asked policyholders--using a multiple choice format--how
they would pay for long-term care in the future, if they had not purchased
a Partnership policy. One of the possible responses was that they would
purchase traditional insurance. This required policyholders to speculate
about future behavior, and to respond to a more complex question and
answer format. California, Connecticut, and Indiana asked directly about
decisions made in the past--whether the Partnership policyholder would
have purchased long-term care insurance in the absence of the Partnership
program, with a simple yes or no response.

  Impact on Medicaid Savings of Purchasing Different Levels of Benefits by
  Partnership and Traditional Policyholders

California, Connecticut, and New York also commented that our finding that
Partnership policyholders tended to have more extensive benefits than
traditional policyholders was inconsistent with our scenarios that assumed
that the policyholder would have purchased comparable benefits in the
absence of the Partnership program. Assuming comparable benefits in our
scenarios allowed us to assess the impact of the Partnership program on
Medicaid savings in a simpler framework. As we explain in appendix II,
some Partnership policyholders may have more coverage than if they had
purchased a traditional policy. We show that if the value of the insurance
policy is less than the amount of assets owned by the policyholder, the
person will still take longer to become Medicaid eligible with a
traditional long-term care insurance policy with a lesser value than with
a Partnership policy. However, the amount of additional time it would take
for the individual with a traditional policy to become eligible for
Medicaid would be less than if the two policies had the same amount of
benefits.

HHS, the Indiana Partnership program, and the New York Department of
Insurance provided us with technical comments and clarifications, which we
incorporated as appropriate.

72Whereas California, Connecticut, and Indiana surveys asked: "Would you
have purchased long-term care insurance in the absence of the
Partnership?" (yes or no), New York's survey asked: "Had you not purchased
Partnership insurance, what would be your plan to pay for LTC you may
need?" Seven response options were provided, one being "I would purchase
non-Partnership long-term care insurance."

As agreed with your offices, unless you publicly announce its contents
earlier, we plan no further distribution of this report until 30 days
after its issue date. At that time, we will send copies of this report to
the Secretary of Health and Human Services, congressional committees, and
other interested parties. We will also 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 questions about this report, please contact me
at (202) 512-7119 or [email protected]. Contact points for our Office of
Congressional Relations and Public Affairs may be found on the last page
of this report. GAO staff who made key contributions to this report are
listed in Appendix IV.

John E. Dicken
Director, Health Care

Appendix I: Data and Methods for Analysis of Long-Term Care Insurance
Benefits and Demographics

In this appendix we describe the data and methods that we used to examine
the benefits of Partnership and traditional long-term care insurance
policyholders. We also describe the data and methods we used to assess
income, assets, age, gender, and the marital status of Partnership program
long-term care insurance policyholders, traditional long-term care
insurance policyholders, and people without long-term care insurance.

Examining Long-Term Care Insurance Benefits Purchased by Partnership
Policyholders and Traditional Long-Term Care Insurance Policyholders

We examined the benefits purchased by Partnership long-term care
policyholders and the benefits purchased by traditional long-term care
policyholders, using 2002 through 2005 data from two sources. Our data
source for the benefits purchased by Partnership policyholders was the
Uniform Data Set (UDS)--a data set supplied to us by each of the four
states with Partnership programs that contained information on all
Partnership policyholders who had purchased long-term care Partnership
policies. The UDS was developed collaboratively among the four states with
Partnership programs, insurers, the National Program Office at the Center
on Aging, University of Maryland, and the Program Evaluator, Laguna
Research Associates. The UDS contains information submitted by insurers
with Partnership policyholders and summarized by each of the states on a
quarterly basis. Insurers are required to submit data to the state
Partnership program on: (1) newly insured people,^1 (2) people who dropped
their policies, (3) applicants for insurance who were assessed for
long-term care insurance eligibility, and (4) the amount of payments for
services and utilization. We used the data set for newly insured people to
analyze the benefits purchased by Partnership policyholders. These data
contain information on daily benefit amounts, the length of the benefit
period, the length of the elimination period, and the type of coverage,
including whether the coverage is comprehensive coverage or for facilities
only. To obtain data about the benefit characteristics of insurance
policies purchased by traditional long-term care policyholders, we
surveyed five large insurance companies selling long-term care insurance.
We selected these five insurance companies on the basis of the total
number of policies and amount of annualized premiums in effect in the
individual market, as of December 31, 2004. The five insurance companies
were AEGON USA,^2 Bankers Life and Casualty Company, Genworth Financial,
John Hancock Life Insurance Company, and Metropolitan Life Insurance
Company. All five insurance companies sold policies in the individual
market, and two of the five carriers--John Hancock Life Insurance Company
and Metropolitan Life Insurance Company--were also among the five largest
carriers that sold products in the group market. We requested data on the
number of enrollees in the individual market who chose selected benefit
options for new long-term care insurance policies sold from July 1, 2002,
to March 31, 2005. We collected data on coverage types, daily benefit
amounts, elimination periods, benefit periods, inflation protection
options, and optional benefits offered.

1This part of UDS data contains information on each person who applied for
a Partnership policy and who passed the underwriting process.

2AEGON USA left the long-term care insurance market on March 31, 2005.

Examining Income and Asset Distributions Among Partnership Policyholders and
Comparison Populations in Two Partnership States and Nationally

We used three data sources to examine the income and assets of Partnership
policyholders, traditional long-term care insurance policyholders, and
people without long-term care insurance: Partnership program surveys of
Partnership policyholders; the 2004 American Community Survey (ACS); and
the 2004 Health and Retirement Study (HRS).

To examine the household income and household assets of Partnership
policyholders, we used data from Partnership program surveys of a sample
of Partnership policyholders at the time they first purchased insurance
coverage. We restricted our analysis of the income and assets of
Partnership policyholders to surveys conducted by the California and
Connecticut Partnership programs because the Indiana Partnership program's
data were not sufficiently detailed to include in our analysis, and the
New York Partnership program was not able to provide us with data from
recent years. In addition, because the surveys were of a sample of
Partnership policyholders--40 percent of Partnership policyholders in
California and 50 percent of Partnership policyholders in Connecticut--we
increased the number of observations by analyzing more than 1 year of
data. We included data from 2003 and 2004 for California, and data from
2002 through 2005 for Connecticut.

To approximate the household income of individuals without long-term care
insurance in California and Connecticut, we used the 2004 ACS. Household
asset information was not available in these states. The ACS is conducted
by the Census Bureau, as a part of the Decennial Census Program, and
provides information about the characteristics of local communities. The
ACS publishes social, housing, and economic characteristics for
demographic groups, including household income and assets, covering a
broad spectrum of geographic areas in the United States and Puerto Rico.
It is the largest household survey in the United States, with an annual
sample size of about 3 million. In order to make appropriate comparisons
between the income data from the California and Connecticut Partnership
program surveys and the ACS, we restricted our calculations in the income
analysis to respondents who were aged 55 and over, when we calculated our
household income ranges.^3 In our analysis, we used the ACS state
population data as a proxy for people without any long-term care
insurance. Approximately 12 percent of people over age 55 have long-term
care insurance, so our measure is likely to contain approximately 88
percent of people without long-term care insurance.

To examine national-level data on household income and assets of
individuals with and without long-term care insurance, we used information
from the 2004 HRS, the most recent year available for that survey data
set. The HRS is a longitudinal national panel survey of individuals over
age 50, and is sponsored by the National Institute on Aging and conducted
by the University of Michigan. The HRS includes individuals who were not
institutionalized at the time of the initial interview and tracks these
individuals over time, regardless of whether they enter an institution.
Researchers conducted the initial interviews in 1992 in respondents' homes
and conducted follow-up interviews over the telephone every second year
thereafter. HRS questions pertain to physical and mental health status,
insurance coverage, financial status (including household income and
assets), family support systems, employment status, and retirement
planning. We used data from the HRS to calculate the household income
distribution nationally for people with long-term care insurance and for
people without long-term care insurance. To make our income analysis of
HRS data consistent with the income analysis of Partnership policyholders
and individuals without long-term care insurance in California and
Connecticut, we restricted the HRS income analysis to individuals age 55
and over. The HRS data for people with insurance do not differentiate
between Partnership and traditional insurance policyholders and
approximately 2 percent of people with long-term care insurance nationwide
have Partnership policies. Therefore, although the HRS data may contain a
small number of Partnership policyholders, about 98 percent of all
long-term care policyholders are likely to have traditional long-term care
insurance.

3We restricted our analysis of income to people age 55 and older because
long-term care policies tend to be purchased by people in their late 50s
or early 60s, and people in this age group may have a different level of
income compared to the average for the population as a whole.

Examining Demographic Characteristics--Age, Gender, and Marital Status--of
Partnership Policyholders and Other Populations in Partnership States and
Nationally

To compare the age, gender, and marital status of Partnership
policyholders and other populations in Partnership states and nationally,
we used data from the UDS from 2002 through 2005 and the HRS data from
2004. The UDS data contain information on Partnership policyholders, while
the HRS was used to calculate estimates for traditional policyholders and
for those people without long-term care insurance.

Data Reliability

We took several measures to ensure the reliability of the data used in
this report. For the UDS and Partnership policyholder surveys conducted by
the states with Partnership programs, we interviewed the officials at the
state offices familiar with these data in order to establish whether the
data were reliable and suitable for the purposes of our report. For the
GAO survey of traditional long-term care insurance carriers, we
interviewed each of the carriers about their data to ensure the accuracy
and reliability of the data provided. For the ACS and HRS data sets, we
collected and examined the data documentation and sought information from
the providers of the data. In addition, we took steps to ensure that the
data were valid and within reasonable ranges. To do this, where
appropriate, we examined the distribution of our key variables,
calculating estimates of central tendency, ranges, and frequencies,
missing values, and sample size. We determined that these data sets were
sufficiently reliable for the purposes of this report.

We performed our work from September 2005 through May 2007 in accordance
with generally accepted government auditing standards.

Appendix II: Explanation of the Simplifying Assumptions Used in the
Illustrative Scenarios

To analyze the impact of the Partnership programs on Medicaid, we used
scenarios that are illustrative of the options individuals have to finance
their care. This appendix provides additional information on the
construction of the three scenarios and how adjusting the simplifying
assumptions affects the length of time it takes for the individual to
become eligible for Medicaid in the scenarios.

Illustrative Scenarios for Time Taken to Become Eligible for Medicaid

We developed three illustrative scenarios based on the three basic options
that are available to an individual for financing his or her long-term
care: (1) self-financing without any long-term care insurance, (2)
financing using traditional long-term care insurance, and (3) financing
using Partnership long-term care insurance. For illustrative purposes, our
scenarios are based on an individual who owns assets worth $300,000. If
the individual has a long-term care insurance policy, this policy covers 3
years of nursing facility care at a cost of $70,000 per year: that is, the
policy covers $210,000 of nursing facility care costs. We then used the
scenarios to compare the time it would take for the individual to become
eligible for Medicaid using each of the three financing options.^1

           o Self-financing Without Any Long-term Care Insurance (Scenario
           C): The calculations underlying the self-financing scenario are
           the simplest. If the individual self-finances in the absence of
           the Partnership program, the individual pays for his or her own
           care, essentially spending his or her assets down to Medicaid
           eligibility thresholds before becoming eligible for Medicaid. In
           this case, the number of years it takes for the individual to
           become eligible for Medicaid equals the total assets^2 divided by
           the cost of a year of nursing facility care. In our example, this
           is $300,000 in assets divided by $70,000 in nursing facility costs
           per year, or about 4.3 years until the individual is eligible for
           Medicaid. The equation for this calculation can be expressed as

           Equation 1: Self-financing time to Medicaid

           = Assets / cost per year
		   
1We used constant dollars in our scenarios. This means that the purchasing
power of dollars in our scenarios is constant over time. We also hold the
individual's assets at a fixed dollar amount over time.

2In our illustrative scenarios, we assume the individual spends his or her
assets to zero. In other words, we disregard the effect of Medicaid
allowing beneficiaries to retain some assets. We address the effect of
adjusting this assumption in this appendix.

           o Traditional Long-term Care Insurance (Scenario B): Next we
           calculated the time it would take to become Medicaid eligible if
           the same individual has a traditional long-term care insurance
           policy. In this scenario, the insurance policy pays for care up to
           the limits of the policy. After the insurance policy is exhausted,
           the individual spends his or her own assets to pay for long-term
           care. Once the assets are exhausted, the individual is eligible
           for Medicaid. In our example, the individual has a traditional
           long-term care insurance policy worth 3 years of care in a nursing
           facility or $210,000. The time to Medicaid in this example is
           $210,000 in insurance coverage plus $300,000 in assets, all
           divided by $70,000 in nursing facility costs per year, or about
           7.3 years. The equation for this calculation can be expressed as

           Equation 2: Traditional insurance time to Medicaid

           = (Insurance policy value + assets) / cost per year

           o Partnership Long-term Care Insurance (Scenario A): Finally, we
           calculated the time it would take to become Medicaid eligible if
           the same individual has a Partnership policy. In this scenario,
           the insurance pays for care up to the limits of the policy, and
           then the individual has to self-finance using unprotected assets.
           Once those assets are exhausted, the individual is eligible for
           Medicaid because protected assets do not have to be spent on care.
           In a dollar-for-dollar model, the protected assets are equivalent
           to the value of the insurance policy. The time to become eligible
           for Medicaid in this example is $210,000 in insurance coverage
           plus $90,000 in unprotected assets, all divided by $70,000 in
           nursing facility costs per year or 4.3 years. The equation for
           this calculation can be expressed as

           Equation 3: Partnership insurance time to Medicaid

           = (Insurance policy value + unprotected assets) / cost per year

           Under the assumptions of our illustrative scenarios, with a
           dollar-for-dollar policy, the sum of the insurance policy and the
           unprotected assets is equal to total assets: the same as for the
           self-financing scenario. Therefore, the time to Medicaid is the
           same for the individual in both the Partnership and self-financing
           scenarios, and it is greater if the individual purchases a
           traditional policy than if he or she purchases a Partnership
           policy. These relationships are shown graphically in the report in
           figure 1.
		   
		   Evaluating the Effects of Adjusting the Assumptions Underlying
		   the Illustrative Scenarios

           Underlying our illustrative scenarios were several simplifying
           assumptions. When we adjusted these simplifying assumptions, we
           found that some resulted in no change, some resulted in
           accelerated Medicaid eligibility, and some resulted in delayed
           Medicaid eligibility. Overall, we believe that the survey data
           showing that 80 percent of Partnership policyholders would have
           purchased a traditional long-term care insurance policy in the
           absence of the Partnership program represent compelling evidence
           that, as currently structured, the Partnership programs are
           unlikely to result in Medicaid savings.

           While some of the 20 percent of Partnership policyholders who
           would have self-financed their care and become eligible for
           Medicaid may represent a source of savings, others may represent a
           source of increased spending and still others will result in
           neither savings nor spending. We believe that in the aggregate the
           savings potential from the Partnership programs of these 20
           percent of individuals is outweighed by the 80 percent of
           individuals who will likely result in increased Medicaid spending.
           Specifically, we made the following simplifying assumptions in our
           scenarios and discuss the effect on our results of adjusting these
           assumptions:

           o The individual depicted in the scenarios has assets and benefits
           that are typical of many individuals with long-term care
           insurance. The individual depicted in the scenarios has $300,000
           in assets and 3 years of long-term care insurance--assets and
           benefits that are typical of many individuals with long-term care
           insurance. The individual also receives long-term care in a
           nursing facility with costs for a year of care of $70,000 that are
           roughly equivalent to average nursing facility costs nationwide in
           2004. In our example, the individual has assets of $300,000 and,
           in two of our scenarios, a long-term care policy worth $210,000.
           The cost of a year of nursing facility care is $70,000. As long as
           the individual has assets that are greater than the value of the
           insurance policy, we can insert any numbers into equations (1),
           (2), and (3), and the individual becomes eligible for Medicaid at
           the same time as with a Partnership policy and if he or she
           self-finances, but it takes longer if the individual has
           traditional insurance.

           o The individual spends eligible assets to zero as a condition for
           Medicaid eligibility. While states allow individuals to keep a
           small amount of assets, these assets are in addition to anything
           that needs to be spent to become eligible for Medicaid. Including
           these assets has little impact on the scenarios since the
           individual can keep the same assets in all three scenarios, and
           these assets are outside of any spend-down requirement. Using our
           examples above, we decrease the assets in equations 1 and 2, and
           decrease the unprotected assets in equation 3 by the amount of
           assets the individual is allowed to keep. We find that having a
           traditional policy will still result in more years to Medicaid
           eligibility than having a Partnership policy or self-financing,
           and the Partnership and self-financing scenarios still result in
           the same time to Medicaid eligibility.

           o The individual purchases the same amount of insurance benefits
           under the Partnership and traditional long-term care insurance
           scenarios. While this is our simplifying assumption, we recognize
           that the Partnership policyholder might have more coverage than if
           he or she had purchased a traditional policy because of the extra
           benefit requirements of Partnership policies, such as inflation
           protection, that are not required of traditional policies.
           Provided the individual has assets that are no less than the value
           of the insurance policies, the Partnership policyholder will still
           not take as long to reach Medicaid eligibility as he or she will
           with a traditional policy, although the difference is narrower
           than if the benefits are the same. For example, if we change the
           value of the benefits in the traditional policy to $150,000 (and
           keep the value of the Partnership policy at $210,000 and the value
           of assets at $300,000) the equation for the traditional policy
           becomes ($150,000 + $300,000)/$70,000, which is about 6.4 years
           and is still longer than the approximate 4.3 years with the
           Partnership policy but less than the approximate 7.3 years
           expected if the policies in the traditional and Partnership
           scenarios were the same.

           o The individual has income below Medicaid eligibility thresholds.
           We made this assumption because Medicaid income eligibility
           thresholds vary across states. However, increasing the
           individual's income up to Medicaid eligibility thresholds has no
           impact on the scenarios since the individual can keep the same
           income in all three scenarios. If the income exceeds Medicaid
           eligibility thresholds, the individual is ineligible for Medicaid
           in all three scenarios.

           o The individual's assets are greater than the value of the
           insurance policy. We assumed that most individuals would have
           assets that are worth more than the value of the insurance policy.
           Individuals have a disincentive to purchase long-term care
           insurance with a value exceeding their assets because it might
           increase their premium unnecessarily. While we do not have
           information about the amount of assets that Partnership
           policyholders have at the time they use their benefits, available
           evidence suggests that most individuals do not overinsure the
           value of their assets at the time of purchase, though their status
           could change over time. Survey data from California and
           Connecticut show that while 53 percent of Partnership
           policyholders have more than $350,000 worth of household assets at
           the time of purchase, only about 32 percent of these Partnership
           policyholders have more than 5 years of coverage equal to
           $350,000.^3

           However, it is possible that some policyholders will spend some or
           all of their assets by the time they require long-term care and
           will have more insurance than assets. If we modify our example
           above, and assume the individual's insurance policy has greater
           value than the assets in our scenarios, we see that with a
           Partnership policy, the individual will still become eligible for
           Medicaid sooner than with a traditional policy, but later than if
           he or she self-financed. Therefore, for the 20 percent of
           individuals who would have self-financed their care in the absence
           of a Partnership program, and who have more insurance than assets,
           the Partnership program results in savings to Medicaid.
           Specifically, if we assume the insurance policy is worth $210,000,
           and the individual has assets equal to $150,000, we obtain the
           following results from our scenarios:

           Self-finance: assets / cost per year = $150,000 / $70,000 = 2.1
           years

           Traditional insurance: (insurance + assets) / cost per year =
           ($210,000 + $150,000) / $70,000 = 5.1 years

           Partnership insurance: (insurance + unprotected assets) / cost per
           year =

           ($210,000 + 0) / $70,000 = 3 years

           o The individual is unmarried. In our illustrative scenarios, we
           assume the individual is unmarried--the most likely marital status
           of policyholders at the time nursing home care is required. On the
           other hand, if the individual is married, Medicaid allows spouses
           to keep a certain amount of jointly owned assets (i.e., half of
           the value of the assets up to a maximum amount that was
           approximately $100,000 in 2007). In general, the spousal exemption
           that is deducted from assets would be the same across all three
           scenarios and would not affect the basic relationships among the
           three scenarios unless the net assets after the spousal exemption
           are of less value than the insurance policy. If the amount of
           insurance exceeds the value of assets net of the spousal
           exemption, there is a potential for Medicaid savings for a
           Partnership policyholder who would have self-financed. If that
           individual would have purchased a traditional policy, Medicaid
           spending would increase. Our scenarios illustrate this point. If
           we assume the individual is married and the spouse has already
           taken the spousal exemption such that the individual's assets are
           $300,000, the results do not change and our original formulas
           remain intact. Alternatively, if we assume the individual's spouse
           is entitled to half of the household assets of $300,000, up to the
           maximum of $100,000, then our results do change and the
           policyholder becomes overinsured. In this instance, if the
           individual self-finances, the spousal exemption would be $100,000,
           leaving the individual with $200,000 in assets. If the individual
           has traditional insurance, the spouse is also entitled to
           $100,000. However, if the individual has a Partnership policy,
           $210,000 of the assets are protected, leaving $90,000 in
           unprotected assets. The spouse would be entitled to half of the
           $90,000, or $45,000. The formulas are presented below.
		   
3Some of these individuals may be married and the household assets may be
shared. In order for the assets to be fully protected for married
individuals, both individuals need to have a Partnership policy.
Partnership policies are not transferable, and if a surviving spouse of a
Partnership policyholder requires long-term care and does not have a
Partnership policy, the assets would not be protected.

           Self-finance: assets / cost per year = $200,000 / $70,000 = 2.9
           years

           Traditional insurance: (insurance + assets) / cost per year =
           ($210,000 + $200,000) / $70,000 = 5.9 years

           Partnership insurance: (insurance + unprotected assets) / cost per
           year =

           ($210,000 + $45,000) / $70,000 = 3.6 years

           o The individual uses the same long-term care services in all
           three scenarios. An individual who self-finances might have an
           incentive to use fewer or less expensive services than if he or
           she were insured by either a Partnership or traditional policy
           because the individual would be paying for services. If the
           individual uses fewer services when self-financing, the assets
           last longer, enabling the individual to pay for care longer and
           postponing Medicaid eligibility. In this situation, the cost per
           year of self-financing would be smaller than if he or she had
           either Partnership or traditional insurance. This would result in
           an increase in the time it takes to become Medicaid eligible for a
           person who self-finances relative to what it would have taken if
           he or she had purchased either a Partnership or traditional
           insurance policy.

           o The individual does not save premiums paid if he or she would
           have self-financed their care such that assets are equal in all
           three scenarios. We made this assumption to make our scenarios
           easier to understand. Premium payments may be
           substantial--potentially as much as $3,000 per year--so it is
           possible that if the individual would have saved their premium
           payments by instead self-financing his or her long-term care, the
           individual would have more assets than either Partnership or
           traditional policyholders would when they begin to use their
           benefits. If this is the case, by self-financing, the individual
           would have more assets to pay for long-term care before becoming
           eligible for Medicaid, which would delay the time to Medicaid.
           Therefore, individuals who purchase Partnership policies would
           have saved their premium dollars and not purchased long-term care
           insurance represent a potential cost to the Medicaid program.
           Using our examples above and assuming 15 years of payments saved
           at $3,000 per year, by self-financing, the individual would save
           $45,000 in additional assets that would otherwise have been spent
           on Partnership premiums. Therefore, the self-financing individual
           has assets of $345,000. Using our equation we see that the time to
           Medicaid is delayed ($345,000 / $70,000 = 4.9 years) if the
           individual self-finances, while relative to this option a
           Partnership policy would accelerate the individual's time to
           Medicaid by 0.6 years. In this case, the Partnership and
           traditional policy scenarios would not change because premiums are
           required to be identical for the Partnership and traditional
           polices.

           o The individual is equally likely to transfer assets in all three
           scenarios. An individual who self-finances or uses traditional
           insurance might be more likely to transfer assets to a spouse or
           other family members than he or she would with a Partnership
           policy, because assets are protected under Partnership policies.
           If the individual self-finances and transfers assets, he or she
           would likely take less time to become eligible for Medicaid than
           with a Partnership policy (and assuming no transfers with the
           Partnership policy), resulting in Medicaid savings. If the
           individual would have purchased traditional insurance, the amount
           of assets transferred would have to be equal to at least the value
           of the insurance policy purchased in order for the Partnership
           program to result in Medicaid savings. If the amount of asset
           transfer is less than the value of the insurance policy, the
           increase in Medicaid spending attributable to the Partnership
           program would be less than without the asset transfer, but would
           still be an increase.
		   
		   Appendix III: Comments from the Department of Health & Human
		   Services
		   
		   Appendix IV: Comments from the four states with Partnership programs,
           California, Connecticut, Indiana, and New York 

           Appendix V: GAO Contact and Staff Acknowledgments

           GAO Contact

           John E. Dicken (202) 512-7119 or [email protected]
		   
		   Acknowledgments

           Christine Brudevold, Assistant Director; Krister Friday; Michael
           Kendix; Julian Klazkin; Elijah Wood; and Suzanne Worth made key
           contributions to this report.
		   
		   GAO's Mission

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

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

           The first copy of each printed report is free. Additional copies
           are $2 each. A check or money order should be made out to the
           Superintendent of Documents. GAO also accepts VISA and Mastercard.
           Orders for 100 or more copies mailed to a single address are
           discounted 25 percent. Orders should be sent to:
		   
		   U.S. Government Accountability Office 441 G Street NW, Room LM
           Washington, D.C. 20548

           To order by Phone: Voice: (202) 512-6000 TDD: (202) 512-2537 Fax:
           (202) 512-6061
		   
		   To Report Fraud, Waste, and Abuse in Federal Programs

           Contact:

           Web site: www.gao.gov/fraudnet/fraudnet.htm E-mail:
           [email protected] Automated answering system: (800) 424-5454 or
           (202) 512-7470
		   
		   Congressional Relations

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

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


(290497)

www.gao.gov/cgi-bin/getrpt?GAO-07-231 .

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

For more information, contact John E. Dicken at (202) 512-7119 or
[email protected].

Highlights of [52]GAO-07-231 , a report to congressional requesters

May 2007

LONG-TERM CARE INSURANCE

Partnership Programs Include Benefits That Protect Policyholders and Are
Unlikely to Result in Medicaid Savings

Partnership programs allow individuals who purchase Partnership long-term
care insurance policies to exempt at least some of their personal assets
from Medicaid eligibility requirements. In response to a congressional
request, GAO examined (1) the benefits and premium requirements of
Partnership policies as compared with those of traditional long-term care
insurance policies; (2) the demographics of Partnership policyholders,
traditional long-term care insurance policyholders, and people without
long-term care insurance; and (3) whether the Partnership programs are
likely to result in savings for Medicaid.

To examine benefits, premiums, and demographics, GAO used 2002 through
2005 data from the four states with Partnership programs--California,
Connecticut, Indiana, and New York--and other data sources. To assess the
likely impact on Medicaid savings, GAO (1) used data from surveys of
Partnership policyholders to estimate how they would have financed their
long-term care without the Partnership program, (2) constructed three
scenarios illustrative of the options for financing long-term care to
compare how long it would take for an individual to spend his or her
assets on long-term care and become eligible for Medicaid, and (3)
estimated the likelihood that Partnership policyholders would become
eligible for Medicaid based on their wealth and insurance benefits.

California, Connecticut, Indiana, and New York require Partnership
programs to include certain benefits, such as inflation protection and
minimum daily benefit amounts. Traditional long-term care insurance
policies are generally not required to include these benefits. From 2002
through 2005, Partnership policyholders purchased policies with more
extensive coverage than traditional policyholders. According to state
officials, insurance companies must charge traditional and Partnership
policyholders the same premiums for comparable benefits, and they are not
permitted to charge policyholders higher premiums for asset protection.

Partnership and traditional long-term care insurance policyholders tend to
have higher incomes and more assets at the time they purchase their
insurance, compared with those without insurance. In two of the four
states, more than half of Partnership policyholders over 55 have a monthly
income of at least $5,000 and more than half of all households have assets
of at least $350,000 at the time they purchase a Partnership policy.

Available survey data and illustrative financing scenarios suggest that
the Partnership programs are unlikely to result in savings for Medicaid,
and may increase spending. The impact, however, is likely to be small.
About 80 percent of surveyed Partnership policyholders would have
purchased traditional long-term care insurance policies if Partnership
policies were not available, representing a potential cost to Medicaid.
About 20 percent of surveyed Partnership policyholders indicate they would
have self-financed their care in the absence of the Partnership program,
and data are not yet available to directly measure when or if those
individuals will access Medicaid had they not purchased a Partnership
policy. However, illustrative financing scenarios suggest that an
individual could self-finance care--delaying Medicaid eligibility--for
about the same amount of time as he or she would have using a Partnership
policy, although GAO identified some circumstances that could delay or
accelerate Medicaid eligibility. While the majority of policyholders have
the potential to increase spending, the impact on Medicaid is likely to be
small because few policyholders are likely to exhaust their benefits and
become eligible for Medicaid due to their wealth and having policies that
will cover most of their long-term care needs.

Information from the four states may prove useful to other states
considering Partnership programs. States may want to consider the benefits
to policyholders, the likely impact on Medicaid expenditures, and the
income and assets of those likely to afford long-term care insurance.

HHS commented on a draft of the report that our study results should not
be considered conclusive because they do not adequately account for the
effect of estate planning efforts such as asset transfers. While some
Medicaid savings could result from people who purchase Partnership
policies instead of transferring assets, they are unlikely to offset the
costs associated with those who would have otherwise purchased traditional
policies.

References

Visible links
  37. http://www.gao.gov/cgi-bin/getrpt?GAO-05-1021R
  38. http://www.gao.gov/cgi-bin/getrpt?GAO-07-280
  39. http://www.gao.gov/cgi-bin/getrpt?GAO-06-401
  40. http://www.gao.gov/cgi-bin/getrpt?GAO-06-401
  41. http://www.gao.gov/cgi-bin/getrpt?GAO-07-280
  42. http://www.gao.gov/cgi-bin/getrpt?GAO-07-280
  43. http://www.gao.gov/cgi-bin/getrpt?GAO-07-280
  44. http://www.gao.gov/cgi-bin/getrpt?GAO-07-280
  52. http://www.gao.gov/cgi-bin/getrpt?GAO-07-231
*** End of document. ***