Efforts by Republican lawmakers to repeal the Affordable Care Act in recent months have repeatedly collapsed after the Congressional Budget Office (CBO) scored the various bills using a closely guarded analytic process done entirely behind closed doors.
As Republicans released proposed legislation, the CBO announced its views about its effect on the federal budget and how it would affect the number of people without health insurance but didn’t provide a clear account of how those numbers were reached.
The lack of transparency about the basis for the CBO’s findings makes it difficult for either experts or the broader public to evaluate their credibility or predict how alternatives to the legislation might be scored. As a result, the scores have been surprisingly bleak, forecasting far more people without health insurance than many experts expected, and opened the non-partisan agency to charges of having a liberal bias.
The models used by the CBO to produce these scores are confidential and unavailable for public scrutiny. Often they exist only on the hard drive of a single staffer or a small group of staffers with responsibility for certain types of legislation. The identities of these custodians are not publicly known, although very often lobbyists in Washington, D.C. know which staffers have custody of the models that cover areas relevant to the lobbyists’ clients.
Disclosure of the models would be a fairly simple task, merely requiring the CBO to post the spreadsheets it uses to make its estimates. So far, however, the CBO has resisted calls for greater transparency and public accountability.
The resistance may be undermining the CBO’s reputation as a non-partisan score-keeper of legislation. White House Budget Director Mick Mulvaney (pictured) recently said that the time for the CBO has “come and gone” and suggested changes to make the agency less influential in the legislative process.
Depending on which type of legislation is being considered, the CBO’s level of transparency varies. Its annual Long-Term Budget Outlook, for example, includes a somewhat detailed explanation of the basis for its forecasts, although even their the agency hides the precise formulas it uses. The healthcare bills are on the opposite end of the spectrum, cloaked in secrecy. For example, CBO’s most recent health care report, produced on July 19 with the staff of the Joint Committee on Taxation, estimated that the latest house bill would increase the number of uninsured by 32 million in 2026, double insurance premiums for individuals who purchase insurance, and decrease budget deficits by $472 billion over the next ten years. But it offered only vague and often unhelpful explanations for how it arrived at these figures.
The CBO says it uses a “microsimulation model” to estimate how rates of coverage and sources of insurance would change under proposed legislation. This was first developed in 2005, and underwent major expansions in 2008, 2009 and 2010, according to a presentation about the model by Jessica Banthin, the CBO’s Deputy Assistant Director for Health, Retirement and Long-Term Analysis. Banthin has worked at the CBO since 2011 and has held her current position since 2013.
The model incorporates data from surveys of individuals, households and companies in order to simulate how they would respond to various legal changes, according to the presentation. These include nationally representative samples of individual health care spending and large samples of business establishments. The model also relies on other CBO models, including the agency’s forecasts about the growth of the economy, the employment situation, interest rates and the federal budget.
One area that has come under intense criticism is the CBO’s secret method for forecasting the number of uninsured. Health policy analyst and journalist Avik Roy has complained that the CBO appears to “ascribe near-magical powers to the individual mandate.” This causes the agency to vastly over-estimate the increases in the number of uninsured in any bill that removes the mandate, according to Roy. But the CBO refuses to disclose precisely what portion of the uninsured are those that would voluntarily drop insurance due to the lack of a mandate.
One result has been public confusion over the impact of the various Republican bills. A recent ABC news story, for example, carried the headline “CBO estimates 22 million would lose insurance in a decade under the GOP health care bill.” That is untrue. Instead the CBO estimated that the number of uninsured would be 22 million higher under the bill than under current law–while refusing to disclose how many people would lose insurance coverage and how many people would choose to go without coverage.
Recently, however, those numbers have come to light. Roy says he obtained from a congressional staffer a CBO estimate of the impact of repealing the individual mandate. According to Roy, the CBO projects that repealing the mandate alone would lead to 16 million fewer insured by 2026, three-fourths of the total increase.
Roy argues that number is far too high, citing a 2016 article in the New England Journal of Medicine that found that the mandate did not have a big affect on overall coverage rate.
The larger issue, however, is not whether Roy or the CBO is right about the mandate. It is that the CBO’s secrecy about its calculations have sown confusion, misinformation and stymied public debate on health policy. Because the CBO has not disclosed this number or how it arrived at it, it is impossible for the public to judge whether the estimate is credible.
When the CBO scored proposals in 2015 to repeal Obamacare, it concluded that it would increase federal budget deficits by $137 billion over the following ten years. It said that if it excluded “macroeconomic feedback”–primarily, the extra economic growth generated by from eliminating Obamacare regulations and taxes–the deficit would increase by $353 billion over the period. According to the CBO, the economy would grow by an extra 0.7 percent during the period thanks largely to an increase in the labor supply, which the CBO thinks would rise by between 0.8 and 0.9 percent when Obamacare’s disincentives to work went away.
The CBO also said that some of the economic growth generated by the larger labor supply would be offset by the larger deficits, which the CBO believes reduce private investment and therefore overall economic output. This has its own feedback loop: higher deficits slow down economic growth, which means deficits grow even more, producing even slower growth.
That’s not an outlandish view. Plenty of economists believe that government deficits “crowd out” private investment. But it is a view that has come under criticism in recent years and does not seem well supported by history. Yet once again, the lack of transparency makes it difficult to see how much of a role “crowding out” plays in the CBOs analysis. If the CBO were to release its model, outside analysts would be able to change assumptions like those about crowding out and produce alternative scenarios for economic growth.
In an era of big data and open-sourcing, the CBO’s veil of secrecy seems outdated and unnecessary. While revealing the guts of its analysis might open the agency up to some criticism about particular assumptions and calculations, the far more important consequence would be a more robust and transparent analysis of legislative proposals.
The CBO was created in 1974 in hopes of making the federal budget process less opaque and less subject to partisanship. Forty-three years later, it is now one of the sources of opacity and partisan bickering about the budgets and legislation.