Page images
PDF
EPUB

Appendix E

Assumptions Used to Estimate Costs for Long-Term Care Recommendations

Estimates of the costs of the long-term care program proposed by the Pepper Commission were made using the Brookings/ICF Long-Term Care Financing Model.1 This appendix discusses the methodology used in estimating the costs of the proposed program. The first section presents an overview of the model structure, the databases used in the model, and the model's underlying logic. The key relationships and behavioral parameters of the model are then discussed. The final section explains the implementation of the Commission's proposal within the model's framework.

OVERVIEW OF THE MODEL

The Brookings/ICF Long-Term Care Financing Model simulates the utilization and financing of both institutional and noninstitutional long-term care services for elderly families over the period 1986 to 2020. Institutional services include nursing home care provided by skilled nursing facilities (SNFs) and intermediate care facilities (ICFs). Noninstitutional services include home health, homemaker, personal care, and meal preparation services. The model simulates the number of individuals receiving these services and the costs of these services as financed by various public and private sources.

The model was jointly developed by The Brookings Institution and Lewin/ICF in 1986. Significant funding for the components of the model was received from the U.S. Department of Labor, the President's Commission on Pension Policy, the American Council of Life Insurance, the Employee Benefits Research Institute, the Congressional Budget Office, and the assistant secretary for planning and evaluation of the U.S. Department of Health and Human Services. Lewin/ICF and The Brookings Institution staff made major revisions to the model in 1989 with the support and guidance of the assistant secretary for planning and evaluation, U.S. Department of Health and Human Services, and the Robert Wood Johnson Foundation.

The objective of the model is to simulate the effects of various financing and organizational reform options.

on future public and private expenditures for nursing home and home care. The model can be used to simulate long-term care financing assuming changes in private methods (such as increased purchase of private long-term care insurance) or new public financing programs. These simulations are greatly affected by the choice of assumptions about the economy (such as the rate of growth of the overall economy and nursing home prices) and individual behavior (such as rates of nursing home utilization and insurance purchase). Because of the uncertainty surrounding some of these events, the model is designed to facilitate use of alternative assumptions to show how sensitive the results are to the assumptions chosen.

The key advantage of this model over actuarial models is its ability to simulate the effects of various policy options on the distribution of benefits for dif ferent economic and demographic groups. For example, the model can be used to generate estimates of the effects of proposals on demographic groups, such as low-income unmarried women age 85 and over, who have few assets.

The model begins with a nationally representative sample of the adult population with a record for each person's age, sex, income, and other characteristics assembled from the sources described below. The model simulates life events for each individual in the sample population from 1986 to 2020, using a Monte Carlo simulation methodology.2 The model first simulates future demographic characteristics, labor force participation, and income and assets of the elderly. It then simulates disability, admission to and use of nursing home and home care, and methods of financing long-term care services. The model uses national data and does not take into account regional, state, or local variations.

The current version of the model is a major revision of the model that was developed jointly by Lewin/ICF and The Brookings Institution in 1986. The model was revised in 1988-89 to incorporate data from a number of newly available data sources, including the 1982-84 National Long-Term Care Survey (NLTCS), the 1985 National Nursing Home

Survey (NNHS), the 1984 Survey of Income and Program Participation (Wave 4), and Medicaid and Medicare program data provided by the Health Care Financing Administration (HCFA).

The six major components of the model are described below. A flowchart of these components is shown in Figure E-1.

Population Database

The model uses demographic and income information from the May 1979 Current Population Survey (CPS) for a nationally representative sample of 28,000 adult individuals of all ages in 1979. This 1979 database was chosen because social security earnings histories for each individual in the sample were matched to the CPS data. This is the last time that the CPS data were matched to social security earnings histories.

Income Simulator

The model then simulates demographic events such as mortality, marital status, labor force activity, income, and assets for each individual. The probabilities of the likelihood of marriage, work, and so forth for different demographic groups were estimated from matched CPS files and other sources. The aggregate probabilities of mortality are taken from Social Security Administration estimates. The macroeconomic assumptions underlying the simulations are generally those used in Alternative II-B of the 1989 Social Security Trustee's Report. These assumptions about future economic growth, inflation, unemployment, and wages greatly influence the model's results. These economic assumptions were selected in part because they are consistent with the demographic probabilities used in the model. Based on these assumptions, the model estimates retirement income from private sector defined benefit pension plans, public pension plans, social security, private sector defined contribution plans, Individual Retirement Accounts (IRAs), and Keoghs. The model also simulates the assets of elderly individuals, including the value of home equity, from the 1984 Panel of the Survey of Income and Program Participation (SIPP).

Disability of the Elderly

Using probabilities estimated from the 1982-84 NLTCS and the 1985 NNHS, this part of the model simulates the level of disability for persons age 65 and over. The model simulates the onset of disability, the level of disability, changes in disability, and recovery from disability.

Figure E-1 Brookings/ICF Long-Term Care Financing Model

Representative Population Data Base

o Income, assets data

o Family structure

o Earnings histories

Simulate Income. Labor Force Activity. Family Structure. Assets in Each Year in Future

o Uses modified version of PRISM

o Income from each major source of income (including assets)

o Assets, by type of asset (housing and financial)

Simulate Disability of the Elderly

o Entry into disability

o Level of disability

o Recovery from disability

Simulate Utilization of LTC Services

o Institutional services

o NonInstitutional services

Simulate Sources and Levels of Payment

o Eligibility for public programs

o Program rules

o Eligibility for private programs

Analysis Tabulations

o Public, private expenditures for LTC o Nursing home, home care utilization

Utilization of Long-Term Care Services

This part of the model uses probabilities estimated from the 1982-84 NLTCS and the 1985 NNHS to simulate admission to and length of stay in a nursing home. For noninstitutionalized persons, the model also simulates the use and length of receipt of paid home care services using probabilities derived primarily from the 1982-84 NLTCS and Medicare program data. Many policy changes can affect the utilization of services. Changes in the utilization of services resulting from changes in delivery or financing of longterm care are implemented in this module.

Sources and Levels of Payment

The fifth component of the model simulates the sources of payment and the level of expenditures for each individual receiving institutional or noninstitutional long-term care services. The model incorporates current Medicare eligibility and coverage provisions (post-Medicare Catastrophic Coverage Act provisions) and a set of uniform assumptions about the Medicaid program, including spousal impoverishment protection. State Medicaid program variations are not modeled. Altering the characteristics of existing programs or adding new programs to this component of the model permits analysis of the impact of policy options on the financing of care and the economic impact of policy changes on various demographic groups.

Aggregate Expenditures and Utilization

The sixth part of the model accumulates Medicare, Medicaid, private expenditures, and utilization for each simulated individual for each year. The final output file from the model provides detailed information for individuals age 65 and older, for each year from 1986 to 2020, on individuals' age, sex, marital status, disability, sources and amounts of income, assets, and use of and payment sources for nursing home and home care services. This file is tabulated to show aggregate long-term care expenditures for various demographic groups and sources of financing.

SPECIFICATION OF KEY MODEL RELATIONSHIPS

Under current policy, four key factors and relationships have the largest influence on the level and distribution of long-term care expenditures over the 1986 to 2020 period. These four factors are:

• Income and assets of the elderly, which influence whether individuals are able to pay privately for long-term care or must rely on Medicaid. In general, if income and assets increase faster than prices, the ability of the elderly to pay privately will increase.

• The level and severity of disability, which determines the size of the disabled population and the average number of years of disability prior to death. Whether disability rates will decrease as mortality rates decline will have a large impact on future long-term care expenditures.

• The likelihood of nursing home entry, which influences long-term care expenditures. If individ

uals use nursing homes less in the future, it may result in declines in aggregate long-term care expenditures.

• Nursing home and home care prices, which affect not only expenditures but also utilization. Whether prices will continue to increase faster than the consumer price index (CPI) as they have in recent years will have a large effect on the ability of individuals to pay privately for long

term care.

This section discusses how these key factors and relationships were modeled and how these analyses have been incorporated into the model.

Future Retirement Income and Assets

Approximately one-half of all long-term care expenditures for the elderly are accounted for by out-ofpocket payments made by the elderly. Because of the heavy reliance on out-of-pocket payments, understanding the level and distribution of the income and assets of the elderly is extremely important to understanding the ability of the elderly to pay for long-term care and the degree to which they must rely on public programs.

Retirement Income-In recent years the average income of the elderly has increased, and poverty rates for the elderly have declined. Much of these increases in income have stemmed from higher social security payments and the expanded role of employer pensions. In the future, the level of retirement income will depend on a number of key factors, including preretirement earnings histories of workers, the age of retirement, and the level of pension coverage.

The model simulates benefits from social security, private and public employee retirement plans, IRAs, earnings, asset income, and Supplemental Security Income (SSI) program benefits. Pension and social security benefits are simulated based on the simulated work history of an individual. Each time a worker is simulated to enter a pension-covered job, he or she is assigned to an actual pension plan sponsor selected from a representative sample of private and public retirement plan sponsors. If the individual becomes eligible for a pension benefit, the model then calculates the individual's benefit using the plan's actual benefit provisions.

The individual labor supply probabilities in the model are linked to aggregate estimates of employment levels and industry composition. The aggregate levels are based on Bureau of Labor Statistics fore

casts of labor force participation rates and the distribution of workers by industry of employment. 3. 4 Aggregate unemployment rates, inflation, and real wage growth, which also influence retirement income, are based on the Social Security Actuary's IIB assumptions from the 1989 Social Security Trustees' Report. These reflect long-run real wage growth of approximately 1.3 percent per year. Labor supply in the model is described in detail elsewhere.5

The model's estimate of labor supply reflects the assumption that, over time, a higher proportion of workers will be employed in the services industries and a lower proportion in the manufacturing industry. This assumption has important implications for both wages and pension coverage.

Pension coverage rates in the model are assumed to remain constant on an industry-specific basis in the future. The overall pension coverage rate drops, however, because of the increasing proportion of jobs in the service industry. This is consistent with recent data from the 1983 and 1988 Special Pension Supplements to the CPS.

Assets in Retirement-For most individuals, assets are an important factor in financing long-term care expenditures. Not only are assets used to pay for long-term care, but Medicaid coverage is also conditioned upon the level of assets. The model simulates, in five steps, the level of assets and the income from these assets for persons age 65 and over. First, in 1979, each family unit with a member age 65 and over is assigned a level of assets. This level of assets is based on a distribution of assets from an analysis of the 1984 SIPP Wave 4, which was conducted by the Bureau of the Census.

The model assigns family units the level of assets of similar families from the 1984 SIPP on the basis of age, marital status, pension-receipt status, and income status. Actual individual data records from the 1984 SIPP, adjusted for inflation, are assigned to individuals simulated in the model. This allows a distribution of assets, rather than just an average amount for different demographic subgroups. The model imputes the distribution asset levels for two types of assets: home equity and all other financed assets.

In a second step, the model uses a similar procedure to assign a level and distribution of assets to families who reach the age of 65 after 1979. These probabilities are based upon the distribution and level of assets of persons who were age 63 to 67 in 1984 in SIPP. Before 1984, assets are reduced by a factor equal to the actual rate of change in the CPI over the time

period. The level of assets from 1984 to the present is increased by the actual rate of change in the CPI, and in the future by the projected rate of change in the CPI assumed under the Alternative II-B assumptions. This means that, over time, the real level of assets at age 65 is assumed to remain a constant percentage of income, after controlling for marital status, income, and pension receipt status.

In a third step, the assets of elderly families are adjusted over time to reflect saving/dissaving patterns during retirement. Assumptions concerning changes in the real value of net financial assets (that is, nonhousing) are more difficult to specify because of the lack of good empirical data on dissavings in retirement. Longitudinal data from SIPP for 1984 and 1985 were used to study the saving/dissaving patterns of the elderly by selected demographic and economic characteristics. Based on these preliminary analyses, the model allows some elderly families to save, some to dissave, and others to maintain a constant real level of financial assets.

In a fourth step, the model calculates an assumed level of income from nonhousing assets for family units with a person age 65 and over. In the long run, the rate of return is assumed to be 2 percentage points above the CPI. This assumption is derived from the long-run average interest rate assumed in the Alternative II-B projections of the Social Security Administration. The long-run interest rate is 6 percent. Therefore, with an assumed 4 percent annual increase in prices, the real rate of return on financial assets in the model is 2 percent.

Finally, to reflect real increases in real estate values, the model assumes that the value of net housing assets increases 1.6 percentage points faster than the CPI. The rate of return on housing assets is expected to be lower than that on financial assets as a result of slowing growth in housing demand.

Disability-Modeling disability level and the characteristics of the disabled population is important because almost all users of long-term care are disabled. In the model, disability status for the elderly is simulated in several steps. In the first year that an individual turns 65, he or she may be assigned a disability level. At the start of each subsequent simulation year, the model simulates the disability status of every individual. Various events occur during the year that affect the number of disabled elderly persons in the population, such as:

• Some persons become disabled.

• Some disabled persons have increased disability.

[blocks in formation]

In the model simulations, disabled individuals age 65 and over are defined as those who are unable to conduct at least one instrumental activity of daily living (IADL) (that is, doing heavy work, doing light work, preparing meals, shopping for groceries or other personal items, getting around inside, walking outside, managing money, or using the telephone) or are unable to conduct at least any one of five activities of daily living (ADLs) (that is, eating, bathing, dressing, "toileting," or getting in and out of bed)." In the model, an individual may be assigned one of four disability levels: (1) a deficiency in one or more instrumental activities of daily living (IADL only), (2) a deficiency in one activity of daily living (one ADL), (3) a deficiency in two or more activities of daily living (two or more ADLs), or (4) no disability.

The model simulates intrayear changes in disability status for persons who are admitted to or discharged from a nursing home (based on data from the 1985 NNHS) or who start to use noninstitutional services (based on data from the 1982-84 NLTCS). For all other persons in the model, changes in disability status are simulated at the start of each simulation year using an annualized disability transition matrix by age and marital status based on data from the 1982-84 NLTCS. The model takes these changes in disability status into account (that is, adjustments are made for remissions from disability and death) and simulates disability for additional persons, if necessary, to match the disability prevalence assumptions.

The model assumes that disability rates based on age/sex/marital status remain constant over time. Because mortality rates are projected to decline in the future, individuals are simulated to spend more years in disability. This has a large effect on the model results.

Long-Term Care Utilization-Another important component of the model is the simulation of nursing home use, including determination of entry and length of stay, and formal home care use.

Nursing Home Utilization-During each year some individuals are simulated to enter a nursing home. In these cases, the model simulates the length of stay and whether the individual will be discharged alive or dead. The model also simulates the individual's disability level while in the nursing home and at discharge, if discharged alive.

[blocks in formation]

Data from the 1982-84 NLTCS were used to estimate regression equations for the probability of admission to a nursing home. Logistic models of the two-year probabilities of nursing home entry were separately estimated for disabled and nondisabled persons. The primary reason for separate models for the disabled and nondisabled was the limited data available in the NLTCS for the nondisabled. Consequently, the variables used in the model to estimate nursing home entry include:

• For the nondisabled-age, sex, marital status, and any prior nursing home stays,

• For the disabled-age, marital status, any prior nursing home stays, and disability level.

Having estimated the two-year probabilities of entry, these probabilities were then annualized and their predictive accuracy was compared against a synthetic annual admission cohort estimated from the 1985 NNHS Discharge File. The annualized probabilities were found to understate admissions for persons over age 85 compared with the 1985 NNHS. Therefore, the NLTCS admission probabilities were adjusted to reflect totals from the NNHS by estimating a regression equation by age group and prior nursing home stay, and using the coefficients as the adjustment factors.

The probabilities in the model explicitly assume that the rates of nursing home entry for disabled and nondisabled persons will remain constant over time on an age/sex/marital status basis. Constant rates imply, among other things, that the nursing home bed supply will increase to accommodate admissions from an increasingly large elderly population.

Nursing Home Length of Stay-The 1985 NNHS Discharge File was used to estimate the nursing home length of stay. The 1985 NNHS contains data on each nursing home stay for each individual in the survey. To obtain the length of stay for complete episodes, lengths of stay were aggregated for persons who were discharged from nursing homes and who reentered a

« PreviousContinue »