Illustration by Rob Dobi
This article appears in the April 2023 issue of The American Prospect magazine. Subscribe here.
In the final year of my doctorate program, I was fortunate enough to win a fellowship that would bring me to Washington for a year to work as a legislative staffer in Congress. Fellowship in hand, I began interviewing with potential House and Senate offices. Several offices required me to complete a skills test as part of the interview process. One of those tests stood out.
The prompt was straightforward but rather challenging: Design a policy to increase net retirement savings for families in the United States without spending a single net federal government dollar. I got to work designing a proposal to allow individuals to divert their tax refunds into a tax-preferred retirement vehicle like an IRA (more or less coming up with what would soon become the Obama administration’s now-defunct “myRA” proposal). This was hardly a revolutionary or even ideal policy proposal—after all, the best way to increase retirement savings in the United States would be to actually increase incomes, so people have some money left over at the end of the month to save. But it fulfilled the parameters of the assignment and I got the job.
Better models won’t solve the underlying problem: that Congress has needlessly abdicated too much power to the scorekeepers.
At the time, I thought the test was more like a wonky IQ test, a brain twister designed to help the congressional office determine whether I was smart enough to tackle the kinds of tough questions I would be expected to face when I arrived on the Hill. Now, after six years writing legislation as a senior Hill aide in the House and Senate, and a decade working in economic policy in Washington, I understand the assignment differently. The test was quite literally an assessment of whether I was capable of designing policy that could clear the primary gatekeepers in Congress: the scorekeepers at the Congressional Budget Office (CBO) and the Joint Committee on Taxation (JCT).
The exercise got at the fundamental element of legislative policymaking that is not well understood by those outside of Washington. The merits of a given policy are rarely evaluated based on whether the policy would help the most people, or even the people who need it most. Utilitarian or even normative questions about whether a policy is “right” or “good” are subjugated to narrow budgetary questions of how much a policy costs. This is doubly true when legislating via the arcane budget reconciliation process, on which Congress has relied almost exclusively to achieve major policy accomplishments since 2010, in the absence of a filibuster-proof majority for either party.
Legislative staffers must write policy with the scorekeepers in mind, at every step of the way. It’s convoluted, suboptimal, and frustrating—a bit like teaching kindergarten “to the test” instead of instilling a love of learning. Successful staffers study the scorekeepers and know how to game the rules, never letting the perfect be the enemy of the CBO-approved. This is how you end up with policies that sunset after only a short period of time, like the recent expiration of the one-year expansion of the Child Tax Credit and the impending 2025 expiration of the 2017 Trump Tax Cuts and Jobs Act. It’s how you get policies that don’t kick in until years after a bill is signed into law, like the 2021 proposal to offer Medicare dental that wouldn’t have gone into effect until 2028. And it’s how you wind up with perennial budget gimmicks, like shifting around the date for which Pension Benefit Guaranty Corporation premiums are due to move money from one budget year to another.
One of the most brilliant and effective legislative staffers I came across during my time on the Hill, a tax expert, took it a step further, making it a priority to establish a strong personal connection with the scorekeepers at the Joint Committee on Taxation. By effectively bringing JCT along for every step of the policymaking process, they were able to eliminate any scoring surprises, and even influence scoring choices that would have a sizable impact along the way. This was legislation by charm offensive, not something the Constitution’s architects anticipated.
AS IS PROBABLY BECOMING CLEAR, the problem is bigger than the scorekeepers, whose biases and limitations have been discussed in this issue at length. It’s the way scorekeeping seeps into every nook and cranny of the policymaking process, shaping not just what makes it into law but also what gets dreamed up in the first place. It’s a bit like what the sociologist Donald MacKenzie described in his book An Engine, Not a Camera, when he explained that financial models that were meant to evaluate financial markets (a camera) had instead started to shape the markets themselves (an engine).
Senators, congresspeople, and their staffs are supposed to be the engines of legislative policymaking, responding to and accountable to the preferences of their constituents. CBO is supposed to be the camera, taking a look at these policies and reflecting back their impacts on the federal budget and the broader economy. Over the last few decades, the roles have reversed, with members of Congress simply writing policy that reflects back what modelers and scorekeepers dictate. It’s the CBO’s world, and Congress is just living in it.
The impacts of this reversal are pernicious. CBO is often weaponized during internal disagreements over policy. Committee chairs who want to close ranks before a markup will proclaim that they aren’t accepting amendments from their fellow committee members that “score” (i.e., cost money). Legislative staffers who don’t want to engage with a policy on the merits will tell their colleagues their boss would love to co-sponsor their legislation, but first wants to see the CBO score. This is a particularly annoying one since it’s very difficult to get the CBO to score legislative proposals from rank-and-file members unless they are poised to become law.
In these examples, the mere threat of a bad CBO score is shaping (or rather eliminating) policy that never even makes it to CBO’s desk. And there are outside modelers who bring their estimates to Washington, also with the power to strangle policy before it even gets off the ground. As this issue has indicated, those outside modelers often use the same assumptions that invariably frustrate the enactment of any policy that intervenes in markets or increases public investment.
One of the low points of my legislative career was being asked to nickel-and-dime a group of health care advocates over a universal long-term care provision (a policy I personally supported) that would cost hundreds of billions of dollars. The proposal was to be attached to Medicare for All legislation that already clocked in at tens of trillions of dollars, rendering the entire exercise a bit silly. It’s no wonder Congress hasn’t made any progress on the looming long-term care crisis, despite a rapidly aging population.
It’s not only what gets “scored” that undermines effective policymaking, it’s also what doesn’t. As Nick Hanauer and others explain in this issue, congressional scorekeepers don’t assess impacts outside the ten-year budget window, like the economic benefits that accrue to society as the investments we make in early-childhood education mature when toddlers become workers.
Sometimes CBO doesn’t bother to estimate a set of relevant economic impacts at all. As Phil Rocco points out in his piece, CBO has basically taken a pass on estimating the economic impacts of climate change mitigation because “many of the linkages between climate change and the federal budget require further assessment.” In other words, Congress is kneecapped when trying to solve the climate crisis, because while the CBO will happily tally up the costs of reducing carbon emissions, they’ve determined that it’s simply too hard to estimate the economic benefits. As Lee Harris explains in this issue, this has led a group of private-sector climate risk modelers, armed with dodgy estimates that they allege can map the threat of extreme weather events with fine-grained precision, to enter the fray and sell the illusion of certitude to businesses, cities, and even parts of the federal government.
And it’s not just the budgetary impact of models that dictates the terms of policy debates in Washington. It’s also the way the scorekeepers model (or don’t model) the broader macroeconomic impacts of a proposal. Jobs numbers can sink a proposal, like when CBO determined that a $15 minimum wage would kill 1.4 million jobs, despite a plethora of economic evidence to the contrary. GDP estimates can as well. Good luck trying to move legislation after the CBO or the Penn Wharton Budget Model issues a dictum saying it will “shrink” the economy.
So if this issue hasn’t convinced you that our current system of legislative modeling and scorekeeping is broken, I’m not sure what will. But the really tough question we must address is how to reform this system, so Congress can do the important work their constituents sent them to Washington to do—solving the climate crisis; providing affordable child care, housing, higher education, and health care; and making the overdue investments in our families and communities that will help us to build a stable, healthy, and equitable economy in this country.
FIXING OUR MODELING SYSTEM WILL REQUIRE two sets of reforms: developing better models, and rightsizing the role of models in our policymaking process. We can and should work on both along parallel tracks. The good news is that we are starting to see early signs of a modeling renaissance. The National Academy of Sciences has begun work to advance macroeconomic modeling of climate risk, and the White House Council of Economic Advisers has made this a priority as well. These could serve as a counterweight to the kind of private-sector climate models being offered by First Street and others.
Several think tanks are in the beginning stages of developing alternative budget and macroeconomic models. And American University just launched a new Institute for Macroeconomic and Policy Analysis (IMPA) to develop next-generation models that will better account for the modern features of markets, including financialization and market power. (Full disclosure: My Groundwork Collaborative colleague Rakeen Mabud is on the board of this effort.) Modelers in the United States can also look to their European colleagues, who have begun undertaking similar efforts.
We should welcome these efforts, and federal grantmaking institutions as well as philanthropic foundations should support them. CBO and those who share its assumptions can’t be the only game in town. New models can help CBO make needed adjustments to its model, so that it is more accurate, transparent, and better reflects the latest empirical economic research. CBO should also adopt a more transparent posture, opening up the guts of its models and their underlying assumptions for other researchers to replicate, critique, and improve.
But better models won’t solve the underlying problem: that Congress has needlessly abdicated too much power to the scorekeepers. Voters elect representatives who they think will best represent their interests. They are effectively choosing a proxy, and they expect these individuals to use their judgment, not the CBO’s, when making decisions about which policies to advance. Elected leaders can and should exercise that judgment when faced with CBO estimates that are highly uncertain or don’t paint the full picture.
Sometimes the right thing to do (e.g., saving the planet) is obvious, and when this is the case our elected leaders can and should say so. And while economic estimates and budget scores may be helpful in comparing the relative benefits of multiple competing proposals for investing in our people, the fact that the United States should invest in our future should be self-evident. When common sense, or a gut check, or even a smell test will suffice, Congress should not overcomplicate it.
Models provide bits and pieces of useful information about the future impacts of a policy. Economic data is helpful, but it’s not the only form of data that matters. Most congressional offices tally constituent calls in favor of or against legislation. That’s a pretty useful piece of data too. Congress also receives reams of expert testimony in committee hearings. When ten out of ten early-childhood educators endorse a child care policy, Congress can listen to that and even favor that evidence over a boiled-down budget score and some educated guesses at the long-run economic impact.
Rightsizing the role of scorekeepers doesn’t require dismissing them altogether. Nor does it have to reflect a deeper anti-intellectual sentiment. In fact, placing economic estimates in the proper context, and using them judiciously and in conjunction with other relevant information, is precisely how modeling is intended to be used. It is the modern Congress that has bestowed upon the scorekeepers outsize power, not divine ordinance or even the Founding Fathers. Leaders in Congress have the unilateral ability to end their self-imposed tyranny from the scorekeepers. It’s past time they did so.
At the end of my time in Congress, I had authored a handful of bills and amendments that made it to the Oval Office for signature, spending a grand total of zero net federal dollars across them all. Although I think these proposals improved policy in a number of arenas—from labor policy to financial regulation to immigration—putting money behind them would have exponentially increased their impact. Until we adopt a better set of modeling tools, and rightsize the role of models and modelers as gatekeepers in our policymaking process, we’ll continue to under-deliver for Americans, achieving consistently suboptimal policy results. Or, as my late father used to say, we’ll continue to “get what we pay for.”
