A Review of CBO’s Estimate of Spending From the Department of Defense’s Medicare-Eligible Retiree Health Care Fund

At a Glance

The Medicare-Eligible Retiree Health Care Fund (MERHCF) was created by the Floyd D. Spence National Defense Authorization Act for Fiscal Year 2001 (Public Law 106-398). That fund pays for health care programs that supplement Medicare for senior beneficiaries of the Military Health System.

In a cost estimate prepared for the Congress in October 2000, the Congressional Budget Office estimated that spending from the MERHCF would total about $57 billion over the 2003–2010 period; actual costs over that period were 3 percent less—about $55 billion. This report provides details about why that estimate differed from actual expenditures and discusses how CBO will use that information to assess and improve the accuracy of future cost estimates and reports.

CBO concluded the following:

  • Although the $2 billion difference between the cost estimate and actual expenditures from the MERHCF was relatively small, there were significant differences between the various components of the estimate and actual spending. Specifically, costs for the TRICARE for Life Medicare wraparound benefit and reimbursements to military treatment facilities were lower than CBO estimated. Those differences were largely offset by the costs of the Senior Pharmacy benefit, which were higher than CBO estimated.
  • The differences between the agency’s estimates and actual spending were attributable to a combination of factors: Specifically, behavioral responses among the affected beneficiaries were different from what CBO anticipated, administrative decisions by the Department of Defense and subsequent legislation affected the amounts paid by the MERHCF for the Senior Pharmacy benefit, and various other factors resulted in outcomes that differed from CBO’s estimates.

Analysis of the cost estimate has generated valuable insights that CBO is using to improve its baseline projections of spending for the MERHCF and also to improve cost estimates for other proposals. For instance, on the basis of its experience preparing the MERHCF cost estimate, and experience derived from projecting the costs of other federal programs, the agency has taken a more cautious approach when estimating the rate at which people begin using new government benefits. CBO has also gained insight into how to better interpret and use data and to highlight the presence of uncertainty when preparing cost estimates.

Notes

Notes

Unless otherwise indicated, all years referred to in this report are federal fiscal years, which run from October 1 to September 30 and are designated by the calendar year in which they end.

Numbers in the text and tables may not add up to totals because of rounding.

Summary

To improve its estimating methods and provide transparency about its estimates, the Congressional Budget Office reviews previous budget projections and, when feasible, its cost estimates for legislation. This report, prepared as part of that process, describes the methods the agency used in October 2000 to estimate spending from the Department of Defense’s (DoD’s) Medicare-Eligible Retiree Health Care Fund (MERHCF).1 The report also provides details about why those estimates differed from actual expenditures and discusses how CBO will use that information to enhance the accuracy of future cost estimates and reports.

The MERHCF was created by the Floyd D. Spence National Defense Authorization Act for Fiscal Year 2001 (2001 NDAA, Public Law 106-398). That fund pays for the TRICARE for Life (TFL) health benefit, a supplemental benefit for military retirees and their dependents that covers the portion of health care expenses not covered by Medicare. It also pays for the Senior Pharmacy benefit and any costs incurred by military treatment facilities (MTFs) for care provided to Medicare-eligible beneficiaries of the Military Health System (MHS).

CBO originally estimated that spending from the MERHCF would total about $57 billion over the 2003–2010 period; actual costs over that period were 3 percent less—about $55 billion (see Table 1). (CBO’s cost estimate covered the years from 2001 through 2010. However, the 2001 NDAA specified that the MERHCF was not to begin making payments until 2003.) The costs for TFL and reimbursements to MTFs were lower than CBO’s estimates by $1.5 billion and $5.8 billion, respectively. However, those differences were largely offset by the costs of the Senior Pharmacy benefit, which were $5.4 billion higher than CBO’s estimate.

Table 1.

Summary of Differences Between CBO’s Estimate of Spending From the MERHCF and Actual Spending Over the 2003–2010 Period

Billions of Dollars

Sources: Congressional Budget Office; Department of Defense.

MERHCF = Department of Defense Medicare-Eligible Retiree Health Care Fund; MTFs = military treatment facilities.

a. CBO’s cost estimate for the MERHCF was prepared as part of Congressional deliberations on H.R. 4205, which led to the enactment of H.R. 5408, the Floyd D. Spence National Defense Authorization Act for Fiscal Year 2001. That act subsequently became Public Law 106-398. The cost estimate covered the period from 2001 to 2010, but P.L. 106-398 specified that the MERHCF was not to begin making payments until 2003.

b. For purposes of this analysis, amounts for MTFs include costs for the Uniformed Services Family Health Plan.

In CBO’s assessment, several factors explain many of the differences between the agency’s estimates and actual spending:

  • Unanticipated Behavioral Responses. Certain behavioral responses were different from those CBO anticipated. For instance, CBO expected that Medicare-eligible beneficiaries of the MHS (referred to in this report as MHS seniors) would use military treatment facilities at higher rates than they had before the creation of TFL, but the evidence indicates there was a shift away from care and pharmaceuticals received at MTFs toward use of TFL and the Senior Pharmacy benefit.
  • Administrative Decisions and Subsequent Legislation. Administrative decisions by DoD and subsequent legislation related to copayments increased the cost of prescription drugs relative to the amounts in CBO’s cost estimate. Neither CBO’s baseline projections nor its cost estimates incorporate assumptions about future legislative changes or their effect on spending.2 When CBO’s cost estimate was prepared, beneficiaries’ cost sharing for prescription drugs purchased through the TRICARE retail pharmacy network and mail order averaged about 25 percent. CBO projected that DoD would increase copayments over time to maintain that same level of cost sharing, as the department had legal authority to do. However, through a combination of administrative decisions and legislation, pharmacy copayments remained unchanged throughout the 2003–2010 period. As the price and use of prescription drugs increased over time, a larger portion of the cost was paid from the MERHCF.
  • Other Factors. Various other factors resulted in outcomes that differed from CBO’s estimates. For instance, the agency’s estimates of per capita expenditures for TFL were higher than the amounts that were actually spent. However, that overestimate was largely offset by an underestimate of the population of eligible users. In the case of the TRICARE Senior Pharmacy benefit, the number of prescriptions per user appears to have increased at a faster rate than CBO projected, which contributed to the agency’s underestimate of spending for that benefit.

Background and CBO’s October 2000 Cost Estimate

In October 2000, CBO published an estimate of the costs of the new health benefits for MHS seniors and spending from the MERHCF. The estimates were based on historical spending by DoD for MHS seniors at military treatment facilities and for special programs designed to provide health benefits to MHS seniors affected by closures of military installations. CBO also examined health care spending for seniors in the economy at large.

The MHS encompasses all health care programs and benefits for active-duty military personnel, military retirees, and their eligible dependents and survivors. It includes government-owned MTFs, pharmacy benefits, and dental programs, as well as programs and activities designed to maintain the medical readiness of military forces. A large part of the MHS consists of TRICARE, which is a collection of health plans that cover uniformed service members, retirees, and their dependents and survivors in the United States and abroad. The plans include managed care options, as well as fee-for-service options with preferred provider networks.

Retirees and their dependents and survivors generally are eligible for the same benefits as active-duty households, although they usually have higher out-of-pocket costs and can access MTFs only on a space-available basis. Also, before enactment of the 2001 NDAA, retirees and their dependents were not eligible for the TRICARE health benefit when they became eligible for Medicare, which for most beneficiaries occurs at age 65. When the 2001 NDAA was enacted, the number of Medicare-eligible TRICARE beneficiaries totaled about 1.5 million.3

The 2001 NDAA established TRICARE for Life, a supplemental (sometimes called wraparound) benefit that covers the portion of health care expenses not covered by Medicare. It also created the TRICARE Senior Pharmacy benefit, which allows MHS seniors to fill prescriptions for medication through the TRICARE National Mail Order Pharmacy and the TRICARE civilian retail pharmacy network. Participants’ out-of-pocket costs for those benefits are usually less than costs incurred under other health plans available to seniors.

The act also established a new Treasury account called the Department of Defense Medicare-Eligible Retiree Health Care Fund. The MERHCF pays for all costs related to health benefits for MHS seniors, including TRICARE for Life, the TRICARE Senior Pharmacy benefit, and any expenses for services and drugs provided at MTFs.4

Amounts in the MERHCF are derived from payments to the fund by the uniformed services and the Treasury. Those payments, which are classified as discretionary spending, are intragovernmental transactions and have no net effect on federal spending because the costs to the uniformed services and the Treasury are offset one for one by the receipt of those payments into the fund. (For additional information about the accounting methods used for those contributions, see Box 1.) Amounts in the MERHCF are then used to pay for health benefits for MHS seniors, which is when a federal outlay is recorded. Those payments from the MERHCF are classified in the budget as mandatory spending because they can be made without further appropriations.5

Box 1.

Intragovernmental Payments to the MERHCF

Amounts in the Department of Defense’s (DoD’s) Medicare-Eligible Retiree Health Care Fund (MERHCF) are derived from two sources: annual accrual payments made by the uniformed services and unfunded liability contributions from the Treasury. Annual accrual payments equal the estimated present-value cost of future benefits from the MERHCF for those who are currently in the service.1 An independent board of actuaries approves the methods and assumptions used by DoD to determine how much each of the services should contribute. That way, after considering many factors—including, but not limited to, the growth of health care spending, life expectancy, and the probability that service members will earn a military retirement—DoD can ensure there will be enough funds in the MERHCF when those members become eligible for Medicare (usually at age 65). Those accrual payments are recorded in the budget as discretionary spending from appropriations for the uniformed services.

In addition to accrual payments from the uniformed services, the MERHCF receives payments from the Treasury. Those payments are made to cover the costs of beneficiaries who had already established eligibility for benefits from the MERHCF when that fund was created or who had already served for periods that contributed to future eligibility and for which no accrual payments were ever made—often referred to as the unfunded liability.2 Under the actuarial board’s oversight, DoD updates its estimate of that amount each year. The remaining unfunded liability will be amortized gradually over a set number of years. The last payment is currently scheduled to occur in 2039.

The accrual payments and unfunded liability payments have no net effect on federal spending. What matters for calculations of federal spending are the payments from the MERHCF for health care. Both the accrual payments and unfunded liability payments are intragovernmental transactions—outlays from one budget account and offsetting receipts (which are negative outlays) in another budget account. Amounts in the MERHCF are then used to pay for health care benefits for Medicare-eligible beneficiaries of the Military Health System, which results in federal outlays. Those payments from the MERHCF are classified in the budget as mandatory spending because they can be made without further appropriations.

All amounts in the MERHCF are held in the form of nonmarketable Treasury instruments and credited with interest until they are needed to pay benefits. As of September 30, 2018—the date used for the most recent published valuation data—DoD estimated that the accrued liability of the MERHCF was $436 billion. On that date, the balance credited to the MERHCF totaled $266 billion, leaving an unfunded liability of $170 billion.3

Because the accrual payments represent a cost to the uniformed services, CBO included those payments in the cost estimate it provided to the Congress in October 2000.4 CBO projected that the accrual payments would total $29 billion over the 2003–2010 period. The actual accrual payments to the MERHCF were about $83 billion over that period, almost triple CBO’s estimate (see the table). The differences between CBO’s estimate and the actual payments arise from differences between CBO’s projections and the estimating assumptions set by the actuarial board. The long-term growth rate of health care spending was a significant factor in those differences. CBO estimated that, in the long run, health care spending would grow by about 4.5 percent per person per year and that the fund would be credited with interest at a rate of 6.5 percent annually. The board, once established, set both the long-run growth assumption for per-person health care spending and the amount of interest credited to the fund at 6.25 percent per year.

If the projected interest rate is significantly greater than the anticipated growth in per capita spending (as it was in CBO’s projections), a smaller amount needs to be set aside in the fund to pay future benefits. A substantial amount of time elapses between when accrual payments are made for current members of the uniformed services and when benefits are paid out of the MERHCF for those same members. Given that fact, the difference between the 4.5 percent growth rate for per capita spending and the 6.5 percent interest rate was significant. About half of all military personnel are younger than 25, and they will not begin receiving benefits from the MERHCF for 40 or more years. With at least 40 years of compounding, if the interest rate was projected to be 2 percent higher than the per capita growth rate, the estimated accrual payment would be less than half what it would be if the per capita growth rate and interest rate were projected to be equal.5


1. A present value is a single number that expresses a flow of future income or payments in terms of an equivalent lump sum received or paid at a specific time. Accrual estimates accelerate the recognition of long-term costs and display the expected net cost of new federal commitments when they are incurred and thus are most controllable by policymakers. For more information on accrual accounting, see Congressional Budget Office, Accounting for Federal Retirement and Veterans’ Benefits: Cash and Accrual Measures (September 2019), www.cbo.gov/publication/55499.

2. The unfunded liability can also change on the basis of new technical assumptions or changes to the program. For instance, if legislation was enacted that increased benefits paid from the MERHCF for existing retirees, it would cause an increase in the unfunded liability.

3. See Department of Defense, Office of the Actuary, Valuation of the Medicare-Eligible Retiree Health Care Fund, September 30, 2018 (February 2020), p. 3, https://go.usa.gov/xwBJN (PDF, 863 KB).

4. See Congressional Budget Office, A Comparison of the Budgetary Effects for Two Proposals to Extend Health Care Benefits to Retirees of the Uniformed Services Who Are Age 65 and Older (October 11, 2000), www.cbo.gov/publication/12788.

5. In its most recent valuation report, the actuarial board assumed that both the growth rate for long-run per capita spending and the rate of interest earned by the MERHCF would be 5 percent. See Department of Defense, Office of the Actuary, Valuation of the Medicare-Eligible Retiree Health Care Fund, September 30, 2018 (February 2020), p. D-3, https://go.usa.gov/xwBJN (PDF, 863 KB).

TRICARE for Life

In CBO’s cost estimate, projected costs for TFL totaled $19.5 billion over the 2003–2010 period (see Table 2). To estimate the cost of TFL, CBO first estimated an average cost per beneficiary and then multiplied that by its projection of eligible beneficiaries. CBO’s estimate of the average cost per beneficiary was based on the cost of premiums for medigap insurance, which provides benefits similar to those available through TFL.6 The medigap information was adjusted to take into account CBO’s estimate that MHS seniors would increase their use of the MTFs and that about 10 percent of the eligible population would continue to use other health insurance options. After those adjustments, CBO estimated that the average cost of TFL in 2003 would be $1,176 per eligible beneficiary. That amount was projected to increase by about 6.5 percent each year as a result of price inflation and an increase in the use of health care services.

Table 2.

Comparison of CBO’s Estimate of Spending From the MERHCF With Actual Spending Over the 2003–2010 Period

Billions of Dollars

Sources: Congressional Budget Office; Department of Defense.

MERHCF = Department of Defense Medicare-Eligible Retiree Health Care Fund; MTFs = military treatment facilities; * = between -$50 million and $50 million.

a. CBO’s cost estimate for the MERHCF was prepared as part of Congressional deliberations on H.R. 4205, which led to the enactment of H.R. 5408, the Floyd D. Spence National Defense Authorization Act for Fiscal Year 2001. That act subsequently became Public Law 106-398. The cost estimate covered the period from 2001 to 2010, but P.L. 106-398 specified that the MERHCF was not to begin making payments until 2003.

b. In CBO’s cost estimate, the Medicare wraparound component (TRICARE for Life) and the costs related to retail and mail-order pharmacies were combined into a line labled “Tricare Costs.” For this display, CBO broke out the components to match cost data received from the Department of Defense. The cost estimate also included costs related to an increase in Medicare usage. Although an increase in Medicare usage represents an increase in costs to the government, those costs are not included in the figures above because they are not paid from the MERHCF and there is no way to seperately track the effect of the TRICARE for Life benefit on overall Medicare spending.

c. For purposes of this analysis, amounts for MTFs include costs for the Uniformed Services Family Health Plan.

On the basis of DoD’s projections, CBO estimated that about 1.5 million beneficiaries would be eligible for TFL in 2003 and that this population would grow by about 1 percent each year, as the military retiree population became older and therefore eligible for the new benefits. Besides the requirement to enroll in Medicare Part B (Medical Insurance), beneficiaries who use TFL pay no premiums and usually have no out-of-pocket costs. Therefore, CBO estimated that almost all eligible beneficiaries would begin using the new benefit as soon as it became available.

TRICARE Senior Pharmacy

CBO’s estimate for the Tricare Senior Pharmacy benefit—$15.1 billion over the 2003–2010 period—was based on observed prescription drug use among MHS seniors affected by the closure of numerous military installations during the 1990s. Before the Senior Pharmacy benefit became available, certain MHS seniors who had lost access to MTF pharmacies because of those closures had access to subsidized prescription drugs through a mail-order program or select retail drugstores. About 10 percent of the MHS population actively used those benefits, and CBO used data on those beneficiaries to estimate the number and cost of prescriptions for the Senior Pharmacy benefit.

After adjusting for those who would not use the new benefit, including those CBO expected would continue to obtain prescription drugs at MTFs, CBO estimated that the total cost of the Senior Pharmacy benefit in 2003 would be about $950 for each of the 1.5 million eligible beneficiaries. Furthermore, CBO estimated that the cost per beneficiary would increase by about 6 percent per year; that estimate accounted for drug price inflation, changes in the mix of prescriptions, and increases in the number of prescriptions filled per person.

Military Treatment Facilities

In CBO’s 2000 analysis, payments to MTFs from 2003 through 2010 were projected to total $22.6 billion. CBO’s estimate of costs for care at military facilities consisted of two parts. First, the agency allowed for the fact that care provided to MHS seniors at MTFs would be reimbursed by the MERHCF. (Before enactment of the 2001 NDAA, that care was funded by discretionary appropriations to the Defense Health Program and military personnel accounts.) On the basis of data from DoD, CBO estimated in 2000 that military treatment facilities were providing about $1.5 billion worth of medical care and pharmaceuticals to about 0.3 million MHS seniors. After including the effects of price inflation, population growth, and an increase in the amount of services used per person, CBO estimated that those costs would increase to about $2.0 billion in 2003.

Second, CBO estimated that DoD would increase the amount of services provided to beneficiaries through military treatment facilities. MHS seniors receive free services and pharmaceuticals from MTFs on a space-available basis and as budgetary resources allow. Before enactment of the 2001 NDAA, at least some demand was not being met because of limited funding, in CBO’s assessment. With mandatory funding for those costs, MTFs could provide goods and services to MHS seniors without the constraints of annual appropriations. CBO estimated that the amount of care received by MHS seniors at MTFs would increase by about $0.2 billion per year, or by about 10 percent, so that total spending for beneficiaries receiving care at MTFs would be $2.2 billion in 2003.

Why Did CBO’s Estimates Differ From Actual Spending Between 2003 and 2010?

Judging the eventual accuracy of cost estimates for legislation that are prepared before the legislation is enacted can be difficult for several reasons. Enacted legislation may differ from an earlier version that CBO analyzed before it was amended by the Congress. In addition, subsequent legislation can cause the costs to differ from amounts originally estimated in ways that are difficult to separate. Finally, the actual costs or savings resulting from enacting legislation are often a small part of a large budget account or revenue stream and cannot be clearly identified.

Expenditures from the MERHCF, however, are clearly identified by DoD and the Treasury; so it is possible to evaluate CBO’s estimate for the provision in the 2001 NDAA that established the fund. However, it is difficult to identify and quantify the exact source of the differences because DoD provides cost data at an aggregate level only. CBO’s estimates were based on factors such as the eligible population, usage rates, prescriptions filled per capita, medical cost inflation, and where retirees and their dependents received care. Although the department provides some statistics on the population and the number of prescriptions filled, the level of detail is not sufficient to directly compare those data with the factors that CBO considered in constructing its cost estimate. CBO concludes that many of the differences between the agency’s estimates and actual spending can be explained by several factors (see Table 3). Those include unanticipated behavioral responses, administrative decisions and subsequent legislation, and other factors.

Table 3.

Sources of Differences Between CBO’s Estimate of Spending From the MERHCF and Actual Spending

Source: Congressional Budget Office.

MERHCF = Department of Defense (DoD) Medicare-Eligible Retiree Health Care Fund; MTFs = military treatment facilities; TFL = TRICARE for Life; n.a. = not applicable.

Unanticipated Behavioral Responses

Certain behavioral responses differed from those CBO anticipated:

  • Whereas CBO expected that MHS seniors would increase their use of military treatment facilities, the evidence indicates there was a shift away from care and pharmaceuticals received at MTFs toward use of TFL and the Senior Pharmacy benefit.
  • CBO also expected that beneficiaries would begin to use the new TFL and Senior Pharmacy benefits more quickly than they actually did, as evidenced by CBO’s overestimate of spending from the MERHCF over the 2003–2004 period.

Changes in Usage of MTFs. Figures published by DoD show that overall inpatient and outpatient services received by MHS seniors at military treatment facilities decreased after the creation of the MERHCF. Between 2002 and 2004, the outpatient and inpatient workload related to MHS seniors at military hospitals and clinics declined by 31 percent and 13 percent, respectively. In contrast, the use of TFL increased over that same period.7

The number of prescriptions filled at military pharmacies also differed from CBO’s estimate. CBO expected that those who filled their prescriptions at MTFs before the creation of the Senior Pharmacy benefit would continue to do so. DoD estimated that in 2001, the year before the start of the Senior Pharmacy benefit, each senior TRICARE beneficiary filled about 19.5 prescriptions at MTFs on average.8 By 2010, DoD reported, the number had declined to about 16.6.9 Although per capita prescription drug use in the United States increased by about 40 percent over that period, the number of prescriptions filled by senior beneficiaries at MTFs fell by 15 percent, which suggests that a significant number of prescriptions that would have been filled at those facilities were instead obtained through the retail network or mail order.

Despite that evidence, one cannot say conclusively that the creation of TFL alone caused the reduction in workload at MTFs. A decline in the use of MTFs and an increase in the use of nonmilitary care occurred among all retiree beneficiary groups over the 2002–2004 period, including those who were too young to qualify for Medicare and TFL.10 The closing of MTFs and a subsequent reduction in the number of beneficiaries living near military bases may have contributed to the reduction.11 Increased security after 9/11 also made it more difficult to access medical facilities located on military bases.

The reduction in the use of MTFs also has implications for Medicare that are not fully reflected in the comparisons between CBO’s cost estimate and actual expenditures from the MERHCF. When an MHS senior receives care at an MTF, the MERHCF reimburses the military facility for the full cost of that care. However, if that same beneficiary uses TFL, the MERHCF pays only that portion of the bill not covered by Medicare. Those added Medicare costs are not included in the comparisons because it is not possible to separate the costs for TFL beneficiaries from the hundreds of billions spent annually from the Medicare trust funds. If determining those additional costs was possible, it could further clarify the differences between CBO’s cost estimate and the actual expenditures because more of the total cost to the government would be included.12

Utilization Rates of New Benefits. CBO also overestimated the rate at which beneficiaries would begin to use the new TFL and pharmacy benefits. Because the TFL and Senior Pharmacy benefits require beneficiaries to pay minimal or no out-of-pocket costs, CBO expected that eligible beneficiaries would maximize their use of the new benefits almost immediately. However, CBO overestimated the cost of the new benefits over the 2003–2004 period, suggesting that adoption of the new benefits was slower than the agency expected (see Table 2). It is not clear why that was the case. By law, TRICARE pays last after almost all other health insurance.13 It is possible that many beneficiaries were slow to understand the value of the TFL and Senior Pharmacy benefits and were reluctant to cancel their private Medicare supplemental insurance plans.

Administrative Decisions and Subsequent Legislation

Administrative decisions by DoD and subsequent legislation had significant effects on amounts paid from the MERHCF over the 2003–2010 period. On net, those policies increased the cost of prescription drugs relative to the amounts in CBO’s cost estimate. They also decreased spending on TFL by changing the amounts paid for skilled nursing facilities and home health care.

Copayments for Prescriptions. The 2001 NDAA authorized DoD to require senior beneficiaries to make the same copayments as other TRICARE beneficiaries who filled prescriptions through the retail network or mail-order program. CBO estimated that, on average, those copayments equaled about 25 percent of the overall cost of the drugs and that DoD would increase those copayments as prescription costs increased to maintain that same 25 percent contribution, as was allowed by the law at that time. However, TRICARE pharmacy copayments did not change over the 2003–2010 period, and because prescription drug costs were increasing, the portion of the costs paid by beneficiaries was less than 10 percent, on average, over the period. Thus, the MERHCF ended up paying a larger share of the cost of each prescription than CBO estimated.

For several years, DoD chose not to increase the copayments. When DoD considered that option, the Congress temporarily prohibited the department from doing so. In particular, the National Defense Authorization Acts for Fiscal Years 2007, 2008, and 2009 (Public Laws 109-364, 110-181, and 110-417, respectively) each froze pharmacy copayments for one year.14 In total, CBO estimated, those acts increased outlays from the MERHCF by about $600 million over the 2003–2010 period, relative to the baseline in effect when those estimates were completed (see Table 4).15 In total, CBO estimates, the combination of both the administrative and legislative actions related to copayments increased the cost of the TRICARE Senior Pharmacy benefit by about $2 billion over the 2003–2010 period relative to CBO’s original estimate.

Table 4.

CBO’s Estimates of the Effects of Subsequent Legislation on Spending From the MERHCF

Millions of Dollars

Source: Congressional Budget Office.

Changes affected mandatory spending.

MERHCF = Department of Defense Medicare-Eligible Retiree Health Care Fund.

a. See Congressional Budget Office, cost estimate for S. 1438, the National Defense Authorization Act for Fiscal Year 2002 (December 31, 2001), www.cbo.gov/publication/13477.

b. See Congressional Budget Office, cost estimate for H.R. 5122, the John Warner National Defense Authorization Act for Fiscal Year 2007 (October 11, 2006), www.cbo.gov/publication/18229; cost estimate for H.R. 4986, the National Defense Authorization Act for Fiscal Year 2008 (January 25, 2008), www.cbo.gov/publication/19449; and cost estimate for S. 3001, the Duncan Hunter National Defense Authorization Act for Fiscal Year 2009 (October 7, 2008), www.cbo.gov/publication/20362.

c. See Congressional Budget Office, cost estimate for H.R. 4986, the National Defense Authorization Act for Fiscal Year 2008 (January 25, 2008), www.cbo.gov/publication/19449.

Prescription Drug Prices. In addition to freezing pharmacy copayments for an additional year, P.L. 110-181 required drug manufacturers to discount the cost of drugs dispensed through the TRICARE retail pharmacy network. The mandated discounts are calculated so that the price of the drug does not exceed the prices set under the Federal Ceiling Price Program.16 CBO estimated that the discounts would reduce spending from the MERHCF by about $1 billion over the 2008–2010 period.17

Payments for Skilled Nursing and Home Health Care. Other legislation also reduced spending from the MERHCF and explains some of the differences between CBO’s cost estimate and actual expenditures. The National Defense Authorization Act for Fiscal Year 2002 (P.L. 107-107) authorized DoD to set maximum allowable charges for skilled nursing facilities and home health care up to two years earlier than it could have before the law was enacted. CBO estimated that P.L. 107-107 would reduce spending for TRICARE for Life by almost $300 million in 2003.18

Other Factors

Some other factors, such as costs per person and the number of beneficiaries, differed from those underlying CBO’s cost estimate.

Average Cost per Person of TFL. On the basis of available information, it appears that CBO overestimated the average cost per person of TFL by about 25 percent over the 2003–2010 period. A possible reason for that overestimate is related to the use of information on medigap insurance premiums to develop the cost estimate. Medigap premiums include insurers’ administrative costs and profit. But because TFL undertakes very little marketing and does not require staff to set and analyze premium rates, the administrative costs are relatively low, so those premiums may not have been a good proxy. Surveys from the 2003–2010 period show that administrative costs and profits made up about 20 percent of medigap insurance premiums.19

Use of TFL. Even though CBO overestimated the per capita cost of Tricare for Life, the agency’s total estimate for TFL was only 8 percent higher than actual expenditures over the 2003–2010 period, mainly because CBO underestimated the number of people who would eventually use the program. As discussed earlier, there is evidence that, for some of their care, beneficiaries shifted from MTFs to TFL at rates greater than CBO expected. Also, evidence suggests that the population of TFL-eligible beneficiaries under the age of 65 was larger than CBO expected.20

Number of Prescriptions per Beneficiary. In addition to the factors discussed above, another reason the cost of the Senior Pharmacy benefit was higher than CBO estimated may be related to the agency’s projection of the number of prescriptions filled per user. CBO’s original projection was based on information from 1999, the most recent year for which complete data were available. CBO used that information to estimate the number of prescriptions the average beneficiary would use each year and then estimated that the figure would grow by about 4.1 percent per year. For the economy as a whole, estimates put the actual per-person growth in the number of prescriptions filled over that period at about 4.8 percent annually.21 When that difference in the growth rate is compounded over the period of CBO’s estimate, it leads to an 8 percent difference in the number of prescriptions filled per beneficiary by 2010.

Considerations for Future Analysis

Shortly after the MERHCF began making payments in 2003, CBO compared the actual payments with its estimates. That process allowed the agency to gain valuable insights that it used to improve its projections of MERHCF spending in the baseline and also to improve cost estimates of other proposals.

For example, it became apparent early on that the number of people who chose to use the new TFL and Senior Pharmacy benefits in the first years of the program was smaller than CBO estimated, although use of those benefits eventually exceeded CBO’s expectations. On the basis of that experience, as well as experience with other federal programs, CBO has taken a more cautious approach when estimating the rate at which people begin using new government benefits. For various reasons, even if the new benefits are generous, there tends to be at least some delay before the program reaches its maximum participation level.

Examination of actual spending can also reveal where adjustments to important estimating inputs may be necessary. In the case of CBO’s estimates for TFL, the use of information on premiums for Medicare supplements, which included administrative costs and profits, was one of the main reasons CBO’s estimates exceeded actual costs. Although it is not possible to identify all adjustments needed to correctly apply data and information to a cost estimate, examination after the fact can offer insights into how to use data in future estimates.

CBO’s projections of spending from the MERHCF highlight the effects of uncertainty in estimates. In almost all cost estimates, the assumptions underlying those estimates, including agencies’ decisions about how to administer programs and individuals’ decisions about how to use government programs and benefits, can contribute to uncertainty. In the case of the MERHCF, DoD’s decision not to increase copayments was one of the reasons CBO underestimated spending for pharmaceuticals. Similarly, although CBO expected that providing mandatory appropriations through the MERHCF would allow MTFs to increase their spending and meet the unmet demand for goods and services provided to MHS seniors, it appears instead that other factors caused an overall decrease in the use of MTFs by those beneficiaries.

In recent years, CBO has begun to include more discussion of uncertainty in its cost estimates and reports. In any cost estimates in which uncertainty is significant, CBO includes a discussion of the topic and highlights the inputs that could cause significant variation in the final costs. When possible, the agency also provides in its reports a range of possible outcomes to better inform policymakers. For instance, in a 2017 report on the Military Health System, CBO acknowledged that it was unclear how certain policies would affect spending by MTFs.22 Those discussions of uncertainty benefit from the agency’s analysis of programs such as TFL and the Senior Pharmacy benefit.


1. CBO’s cost estimate for the MERHCF was prepared as part of Congressional deliberations on H.R. 4205, which led to enactment of H.R. 5408, the Floyd D. Spence National Defense Authorization Act for Fiscal Year 2001. That act subsequently became Public Law 106-398. See Congressional Budget Office, cost estimate for H.R. 4205, an act to authorize appropriations for fiscal year 2001 for military activities of the Department of Defense, for military construction, and for defense activities of the Department of Energy, to prescribe personnel strengths for such fiscal year for the Armed Forces, and for other purposes (October 26, 2000), www.cbo.gov/publication/12873.

Also see Congressional Budget Office, A Comparison of the Budgetary Effects for Two Proposals to Extend Health Care Benefits to Retirees of the Uniformed Services Who Are Age 65 and Older (October 11, 2000), www.cbo.gov/publication/12788.

For details about how to map the figures in CBO’s cost estimate to the figures in this report, see the appendix.

2. See Congressional Budget Office, How CBO Prepares Baseline Budget Projections (February 2018), www.cbo.gov/publication/53532.

3. Department of Defense, Evaluation of the TRICARE Program: FY 2002 Report to Congress (2002), Chapter 4, p. 4-3.

4. In this report, all references to and costs for the MTFs include the Uniformed Services Family Health Plan (USFHP), which is an association of six health maintenance organizations available only at selected locations. See the appendix for additional details.

5. Mandatory spending can generally occur as needed, with no further Congressional action. (In contrast, benefits funded with discretionary appropriations are often constrained by the amount and timing of the appropriation.) However, legislative changes to mandatory spending are often subject to House and Senate budget enforcement rules and also to the Statutory Pay-As-You-Go Act of 2010 (title I of P.L. 111-139).

6. Medigap is private insurance that is purchased to help pay for costs not covered by Medicare, including deductibles and coinsurance.

7. See Department of Defense, Evaluation of the TRICARE Program: FY 2005 Report to Congress (March 1, 2005), pp. 59–64, https://go.usa.gov/xGpdy.

8. See Department of Defense, Evaluation of the TRICARE Program: FY 2004 Report to Congress (March 1, 2004), p. 61, https://go.usa.gov/xGpvk.

9. See Department of Defense, Evaluation of the TRICARE Program: Fiscal Year 2013 Report to Congress (February 28, 2013), p. 79, https://go.usa.gov/xGpvU.

10. See Department of Defense, Evaluation of the TRICARE Program: FY 2005 Report to Congress (March 1, 2005), pp. 57–65, https://go.usa.gov/xGpdy.

11. Ibid., p. 3.

12. In 2000, CBO estimated that those additional costs to Medicare would total about $2.7 billion through 2010. See Congressional Budget Office, A Comparison of the Budgetary Effects for Two Proposals to Extend Health Care Benefits to Retirees of the Uniformed Services Who Are Age 65 and Older (October 11, 2000), p. 4, www.cbo.gov/publication/12788.

13. One notable exception is Medicaid. When a beneficiary is eligible for Medicaid, all other benefits, including those from TRICARE, must be used first.

14. Those laws did not affect DoD’s ability to change the pharmacy copayments after the specified period, but the department chose not to increase those copayments.

15. See Congressional Budget Office, cost estimate for H.R. 5122, the John Warner National Defense Authorization Act for Fiscal Year 2007 (October 11, 2006), www.cbo.gov/publication/18229. Also see Congressional Budget Office, cost estimate for H.R. 4986, the National Defense Authorization Act for Fiscal Year 2008 (January 25, 2008), www.cbo.gov/publication/19449, and cost estimate for S. 3001, the Duncan Hunter National Defense Authorization Act for Fiscal Year 2009 (October 7, 2008), www.cbo.gov/publication/20362.

16. The Federal Ceiling Price Program, which was established by the Veterans Health Care Act of 1992 (38 U.S.C. §8126), set a cap on the prices that manufacturers can charge the Department of Veterans Affairs, DoD, the Public Health Service, and the Coast Guard. The caps on those prices are based on a measure of average manufacturer prices and inflation.

17. On the basis of data from DoD, CBO estimates that the discounts, which take the form of rebates, reduced actual outlays from the MERHCF by $700 million over the 2008–2010 period, or about $300 million less than CBO’s original estimate. That difference can be attributed to legal challenges that delayed the collection of some rebates until 2013. See Congressional Budget Office, cost estimate for H.R. 4986, the National Defense Authorization Act for Fiscal Year 2008 (January 25, 2008), www.cbo.gov/publication/19449.

18. See Congressional Budget Office, cost estimate for S. 1438, the National Defense Authorization Act for Fiscal Year 2002 (December 31, 2001), www.cbo.gov/publication/13477.

19. For instance, see National Association of Insurance Commissioners, “Medicare Supplement Loss Ratios in 2005” (2006), p. 5, https://tinyurl.com/y3jppzmx (PDF, 4.62 MB).

20. Although most people become eligible for Medicare upon turning age 65, at least 100,000 TFL users, or between 5 percent and 10 percent of the population, are under 65. They become eligible for Medicare, and hence TFL, after qualifying for benefits from the Social Security Disability Insurance program or upon a determination that they have end-stage renal disease.

21. All amounts in this report that reference growth in prescription drug use for the economy as a whole are estimated using information found in Rabah Kamal, Cynthia Cox, and Daniel McDermott, “What Are the Recent and Forecasted Trends in Prescription Drug Spending?” (Peterson-KFF Health System Tracker infographic, posted February 20, 2019), https://tinyurl.com/y22zrsdv. This infographic includes a comparison of drug spending before and after the introduction of the Medicare Part D benefit in 2006. It is not clear what, if any, effect the introduction of Part D had on the utilization and costs of the TRICARE Senior Pharmacy benefit.

22. See Congressional Budget Office, Approaches to Changing Military Health Care (October 2017), www.cbo.gov/publication/53137.

Appendix: CBO’s Cost Estimate and Actual Spending From the MERHCF—A Crosswalk

In October 2000, the Congressional Budget Office transmitted an estimate of the budgetary effects of the Floyd D. Spence National Defense Authorization Act for Fiscal Year 2001 (2001 NDAA) to the Congress. CBO’s cost estimate included the projected effects on mandatory (direct) spending and spending subject to appropriation of the benefits provided to Medicare-eligible beneficiaries of the Military Health System (MHS seniors).1 This appendix provides a crosswalk between CBO’s original estimate of spending for MHS seniors and the categories of expenditures used by the Department of Defense (DoD).

CBO’s October 2000 Cost Estimate

For the most part, the estimates of mandatory spending in CBO’s October 2000 cost estimate represented the amounts that would be paid from a new health care trust fund called the Medicare-Eligible Retiree Health Care Fund (MERHCF). (That fund pays for the TRICARE for Life, or TFL, health benefit, the Senior Pharmacy benefit, and costs incurred by military treatment facilities, or MTFs, for care provided to MHS seniors.) The MERHCF was required to start paying benefits at the start of fiscal year 2003, which is why most of the mandatory spending was projected to begin in that year (see Table A-1). Although the MERHCF did not begin making payments until 2003, the 2001 NDAA required that the new Senior Pharmacy benefit begin providing benefits on April 1, 2001, and that TFL begin paying claims at the start of fiscal year 2002. Until the MERHCF was established, those benefits were paid from discretionary funds, which is why CBO estimated additional discretionary spending in 2001 and 2002.

Table A-1.

CBO’s Estimates of Spending for Health Benefits for Medicare-Eligible Beneficiaries of the Military Health System

Billions of Dollars

Source: Congressional Budget Office.

CBO’s cost estimate for the MERHCF was prepared as part of Congressional deliberations on H.R. 4205, which led to the enactment of H.R. 5408, the Floyd D. Spence National Defense Authorization Act for Fiscal Year 2001. That act subsequently became Public Law 106-398.

Medicare-eligible beneficiaries of the Military Health System are retirees of the armed forces and their dependents and survivors.

MERHCF = Department of Defense (DoD) Medicare-Eligible Retiree Health Care Fund; MTF = military treatment facility; * = between zero and $50 million.

a. For purposes of this analysis, amounts for MTFs include costs for the Uniformed Services Family Health Plan.

b. TRICARE health benefits for retirees of the other uniformed services (the Coast Guard, National Oceanic and Atmospheric Administration, and Public Health Service) are paid from mandatory appropriations. In CBO’s original cost estimate, it was assumed that the new health benefits would be paid from those existing mandatory appropriations in 2001 and 2002 and that the new benefits would then be paid from the MERHCF beginning in 2003.

c. Public Law 106-398 included a one-year extension of a demonstration program that allowed Medicare to reimburse military treatment facilities for the cost of operating a health maintenance organization (HMO) for Medicare-eligible beneficiaries of the Military Health System. CBO estimated that the cost of operating that HMO would be higher than the cost of traditional Medicare.

d. Annual accrual payments equal the estimated present-value cost of future benefits from the MERHCF for those who are currently in service. Those accrual payments are recorded in the budget as discretionary spending from appropriations for the uniformed services.

Because MHS seniors would be receiving care with minimal or no out-of-pocket costs, CBO estimated that there would be an increase in overall health care usage by that population, which would increase costs to Medicare. The 2001 NDAA also included a one-year extension of a Medicare subvention demonstration.2 Those costs are paid from the Medicare trust funds, not from the MERHCF. Excluding that estimated $2.7 billion in spending from the Medicare trust funds, CBO projected that spending from the MERHCF would total $57.1 billion over the 2003–2010 period.

Rearranging CBO’s Cost Estimate

In this appendix, CBO’s cost estimate for the 2001 NDAA has been rearranged to more closely match the classification of expenditures provided by DoD (see Table A-2). The estimates of mandatory spending for TRICARE are broken out to reflect the separate costs for TFL and the Senior Pharmacy benefit.

Table A-2.

Mapping CBO’s Estimates of Mandatory Spending Into MERHCF Spending Categories

Billions of Dollars

Source: Congressional Budget Office.

CBO’s original cost estimate covered the period from 2001 to 2010, but Public Law 106-398 specified that the MERHCF was not to begin making payments until 2003. Costs for 2001 and 2002, which were mostly paid from discretionary accounts and are more difficult to trace, are not included in this analysis.

MERHCF = Department of Defense Medicare-Eligible Retiree Health Care Fund; MTFs = military treatment facilities; n.a. = not applicable; * = between zero and $50 million.

a. For purposes of this analysis, amounts for MTFs include costs for the Uniformed Services Family Health Plan.

b. The costs associated with increased Medicare use and Medicare subvention are not paid from the MERHCF, and they cannot be easily broken out from other spending by the Medicare trust funds. Therefore, this report does not include a comparison of those estimates with actual expenditures.

c. Public Law 106-398 included a one-year extension of a demonstration program that allowed Medicare to reimburse military treatment facilities for the cost of operating a health maintenance organization (HMO) for Medicare-eligible beneficiaries of the Military Health System. CBO estimated that the cost of operating that HMO would be higher than the cost of traditional Medicare.

Comparing CBO’s estimate with actual payments by the MERHCF to military treatment facilities is more complicated. CBO’s estimate of payments to MTFs incorporated the costs of certain demonstration programs that existed before the enactment of the 2001 NDAA; those costs, including payments to private network and mail-order pharmacies, were paid for by DoD. To accurately compare the expenditures reported by DoD, the costs related to non-MTF pharmacies have been separated so that all that remains in the line labeled “MTF Costs” are CBO’s estimates of payments to the MTFs, including CBO’s estimate of the increase in MTF costs.3

Costs related to the retirees of other uniformed services (the Coast Guard, National Oceanic and Atmospheric Administration, and Public Health Service) can also be divided into costs for TFL and the Senior Pharmacy benefit. Before creation of the MERHCF, health benefits for those beneficiaries were paid from mandatory appropriations to their respective retirement accounts.

Actual outcomes cannot be traced for some parts of CBO’s cost estimate, so a comparison with actual expenditures is not possible. The mandatory costs associated with increased Medicare use and Medicare subvention are paid from the Medicare trust funds and cannot be separately identified from the hundreds of billions of dollars paid annually from those funds.

Similarly, before the MERHCF started making benefit payments in 2003, almost all new TFL and pharmacy benefits were paid from the discretionary account for the Defense Health Program. In 2002, that program spent about $18 billion, and budget documents from that period do not detail how much of that amount was spent on MHS seniors. Using information from annual TRICARE reports, CBO estimates that DoD spent between $1.7 billion and $2.0 billion on TFL and the Senior Pharmacy benefit in 2002.4 That is more than the $1.7 billion in discretionary spending that CBO estimated.


1. See Congressional Budget Office, cost estimate for H.R. 4205, an act to authorize appropriations for fiscal year 2001 for military activities of the Department of Defense, for military construction, and for defense activities of the Department of Energy, to prescribe personnel strengths for such fiscal year for the Armed Forces, and for other purposes (October 26, 2000), www.cbo.gov/publication/12873.

The estimate used for H.R. 4205 is also discussed in Congressional Budget Office, A Comparison of the Budgetary Effects for Two Proposals to Extend Health Care Benefits to Retirees of the Uniformed Services Who Are Age 65 and Older (October 11, 2000), www.cbo.gov/publication/12788. That publication provides additional details about the estimate and is used as the starting point for this crosswalk.

2. The 2001 NDAA included a one-year extension of a demonstration program that allowed Medicare to reimburse military treatment facilities for the cost of operating a health maintenance organization (HMO) for Medicare-eligible beneficiaries of the Military Health System. CBO estimated that the cost of operating that HMO would be higher than the cost of traditional Medicare.

3. In this report, all references to and costs for the MTFs include the Uniformed Services Family Health Plan (USFHP), which is an association of six HMOs available only at selected locations. In CBO’s analysis, the costs of USFHP—which is administered by private entities—are included in the costs for MTFs in the interest of conciseness and also because that is where USFHP was included in the October 2000 cost estimate. In that cost estimate, the MTF estimate represented all benefits available to MHS seniors before the creation of TFL and the Senior Pharmacy benefit, including those provided by USFHP. In 2003, when the MERHCF first began making payments, about 30,000 Medicare-eligible TRICARE beneficiaries were enrolled in USFHP, which cost the MERHCF about $350 million.

4. See Department of Defense, Evaluation of the TRICARE Program: FY 2005 Report to Congress (March 1, 2005), p. 72, https://go.usa.gov/xGpdy. The estimate of actual spending for 2002 was compiled using data from page 27 and pages 59 through 68. Those same data also report information for 2003 and 2004 that differ from the expenditures from the MERHCF reported by the Treasury, so it is not clear how accurate those figures are.

About This Document

This report evaluates a Congressional Budget Office cost estimate. Such evaluations help guide CBO’s efforts to improve the quality of its projections and, as background information, are intended to assist Members of Congress in their use of the agency’s estimates. In keeping with CBO’s mandate to provide objective, impartial analysis, the report makes no recommendations.

Matt Schmit wrote the report, with guidance from Kathleen FitzGerald, Theresa Gullo, David Newman, and Sam Papenfuss. Comments were provided by Elizabeth Bass, Ann E. Futrell, Heidi Golding, and F. Matthew Woodward.

Joel Sitrin of the Department of Defense provided helpful comments. (The assistance of an external reviewer implies no responsibility for the final product, which rests solely with CBO.)

Jeffrey Kling and Robert Sunshine reviewed the report. Loretta Lettner was the editor, and Jorge Salazar was the graphics editor. The report is available on CBO’s website (www.cbo.gov/publication/56653).

CBO continually seeks feedback to make its work as useful as possible. Please send any comments to communications@cbo.gov.

Phillip L. Swagel

Director

October 2020