Congressional Budget Office consistently wrong on health-care estimates
To understand the budget impact of pending legislation, Congress routinely submits proposals to its accounting arm for analysis.
In the case of the health-care bills, the Congressional Budget Office (CBO) came back with a $1.042 trillion price tag over 10 years for the House bill and a $848 billion cost for the original Senate bill.
Lawmakers often cite CBO figures as holy writ and use them in arguments supporting or opposing proposed measures. But can the CBO estimate costs of complex programs down to the last billion dollars? Do CBO numbers present an accurate picture to legislators and to the American people?
There are a number of problems associated with CBO’s estimates. Some have to do with the games Congress itself plays with numbers. In the case of highly complex programs like health care, a myriad of variables can throw estimates off. In fact, the government’s track record for estimating health-care program costs is poor.
Congress usually asks CBO to judge a bill’s budget impact over 10 years. In the case of health-care reform, Congress is seeking to fudge the budget impact of the program by starting the program in 2013. But the CBO’s 10-year estimates start with 2010, so that they actually include costs for seven program years.
In addition, the CBO estimates show costs rising dramatically over time. “The estimate on the Senate bill could give the impression the program will cost $85 billion a year,” said Marron.
But the CBO estimate shows costs of more than $200 billion at year 10 and rising. In other words, costs for the original Senate bill could run well over $2 trillion during the second decade of the program.
CBO estimates often use sophisticated economic models that try to predict how everyone, from consumers and workers to employers and governments, would respond to a policy change. That’s a lot of moving parts.
For example, the pending health-care bills would loosen eligibility requirements for Medicaid, potentially adding millions of people to that program’s rolls. In order to figure out the increase in Medicaid costs, “you would need to model future economic trends and income distributions in order to estimate how many people would qualify,” explained Marron. “You would also have to model how people decide how to enroll in Medicaid and the behavior of health-care providers and their willingness to provide services. You would also need to estimate average health expenditures, how they are rising over time and how they vary across different population groups.”
There is evidence that governments chronically underestimate health-care program costs. A report released over the summer by Sen. Sam Brownback (R-Kansas) argues that “health care appears to be an area with great room for overly optimistic assumptions regarding” changes in the behavior of patients and providers, the impact of technology, future health cost inflation and the likely success of cost controls.
The report cites numerous instances of faulty health-care estimates from Medicare and Medicaid, the State Children’s Health Insurance Program (SCHIP), the Massachusetts universal-coverage plan and the United Kingdom’s National Health Service.
For example, the House Ways and Means Committee estimated that the original Medicare hospital insurance program would cost $9 billion annually by 1990. Actual spending that year was $67 billion.
The same committee predicted in 1967 that the total Medicare program would cost $12 billion in 1990. Actual spending was $110 billion.
In the case of Medicaid DSH — a program that reimburses states for payments to hospitals that treat Medicaid and uninsured patients — CBO estimated in 1987 that payments would amount to less than $1 billion in 1992. The actual cost that year was $17 billion. The flaw in this case, according to the Brownback report, was that “lawmakers had failed to detect loopholes in the legislation that enabled states to draw significantly more money from the federal treasury than they would otherwise have been entitled to claim.”
To be fair, CBO health-care estimates have also proved to be quite useful. In July 1989, CBO doubled its cost estimate for a Medicare catastrophic coverage benefit to take effect in 1990, citing new data showing it had significantly underestimated cost growth. Congress killed the program before it could take effect.