Illustration by Rob Dobi
This article appears in the April 2023 issue of The American Prospect magazine. Subscribe here.
In 2019, as a part of her presidential bid, Sen. Elizabeth Warren (D-MA) released a bold policy proposal: a wealth tax on the richest Americans. The 2 percent annual tax on total assets above $50 million, rising to 6 percent above $1 billion, would have hit plutocrats hard, from Jeff Bezos to Warren Buffett. According to the Warren campaign, the wealth tax was estimated to raise $3.75 trillion over ten years, and perhaps even more critically, decrease the political power of wealthy individuals.
With similar proposals on the table from Sen. Bernie Sanders (I-VT), it seemed like there was a real acknowledgment of the way that tax policy could be leveraged to reshape the power dynamics of the U.S. economy. But with one report, the momentum behind these proposals ground to a halt.
An institution called the Penn Wharton Budget Model (PWBM) conducted an analysis of Sen. Warren’s proposed tax and found that it would raise $1 trillion less than what her campaign had argued. More importantly, according to the model used by Penn Wharton, wealthy individuals would reduce their investment to avoid taxes, which would depress long-run economic growth and even lower wages across the economy.
More from Rakeen Mabud | David Dayen
The PWBM failed to factor in many aspects of Sen. Warren’s proposal, including a heightened tax enforcement regime and especially the economic benefits of programs that she proposed funding with the tax revenue (universal child care, increased education funding, and student loan forgiveness), which would expand the economy on their own. But the widespread news coverage of the findings, blaring headlines linking the policy to recession, ultimately provided important ammunition for conservatives to push back against the proposed tax and paint it as bad policy. In the fight over Build Back Better, while a minimum tax on corporations did make it into law, wealth taxes were largely sidelined.
Because of the high profile of Warren’s plan and the presidential race, we know about this particular policy, and the pushback that ultimately doomed it. But there are dozens of other ideas that are more quietly shoved aside, or not even contemplated, thanks to a series of obscure gatekeepers who have come to dominate the way we deal with America’s most pressing challenges. They use the language of math and the presumption of certainty to dismiss innovative solutions before they can build a coalition of popular support. They claim to be neutral arbiters reflecting rigid realities about how the world works. But their methods are uncertain, their biases barely concealed, and their ideological goals apparent—if you know where to shine the light.
The key mechanism used to assert this dominance is the macroeconomic policy model. It’s important to be very precise here. Models can be constructive tools that process the best available evidence and incorporate, to the extent possible, the desires of policymakers and elected representatives. The Environmental Protection Agency’s air quality model that formed the basis for how to fight acid rain was a sterling example of the best approximation of the real-world effects of policy driving to a positive result.
Though economists may not want you to believe it, the assumptions they build into their models are actually embedding a perspective about the economy.
As an economist and a journalist who rely on data and evidence, we are in no way reflexively skeptical of the use of numbers. We would not condemn models per se, any more than we would condemn a calculator or a slide rule. But we also know the old adage that predictions are hard, especially about the future. The bigger the system being modeled, the harder it is for those predictions to represent the truth. That’s particularly true when you’re talking about enormous sectors like health care or the environment, or in the case of macro models, the entire national economy over a span of years or even decades.
We believe there should be more humility about the results of macro policy models when they are used in policy debates, especially given how they frequently serve as political weapons. While they produce potentially useful metrics for assessing policy, models are only as good as the assumptions built into them and the information that gets inputted.
The specific modelers most listened to in Washington today are a particular set of economists, think tanks, and government agencies. Budgetary scorekeepers at the Congressional Budget Office and the Federal Reserve, and outside practitioners like the Penn Wharton Budget Model, the State Tax Analysis Modeling Program (STAMP), REMI, and the Tax Foundation’s Taxes and Growth Model, hold immense power over what gets considered. They help ingrain narratives into the public consciousness, dictating how policymakers, journalists, and other players understand what is good or bad for the economy. This creates invisible shackles on policymakers, guardrails that cannot be touched, problems that cannot be solved, no matter how ingenious the prescription.
The modelers and the models they use are rarely scrutinized deeply, even by those most closely attuned to the world of policymaking. Yet for all the claims of neutrality and rigor, these modelers are neither dispassionate, comprehensive, or even accurate much of the time.
In this special issue, we shine a bright light on these gatekeepers to try to understand what the models they use are, where they get their power, and how they acquired so much influence. We will dig into the history of macro models, and try to understand why fixating on the macroeconomy can often lead us down the wrong path. We will better understand the ways that macro models constrain our scope of imagination about what is possible in the policy landscape, and what needs to change in order to make these tools useful.
Molding the Models
As Elizabeth Popp Berman detailed in her book Thinking Like an Economist, over the past several decades an economic style of thinking has taken hold over policy debates in Washington. Prior to this era, major legislation about societal problems could target power dynamics, or uphold universal rights. But starting in the 1960s, this gave way to an approach that sees policy questions through the lens of market dynamics. The best policy, under this framework, is that which finds the most efficient path to solving the problem.
This is how we got housing vouchers over public housing, cap-and-trade over mandated reductions in pollution, “bending the cost curve” over Medicare for All. The economic style tended to downplay regulation in favor of behavioral nudges. It tended to prefer market forces over public forces. “Economists, and the economic style, are not the primary reason that Democratic policy positions moved away from the high liberalism associated with the Kennedy-Johnson era and the Great Society,” Berman writes. “But … economists, and the economic style, were the channel through which this change took place in the Democratic Party.”
Modeling became the further channel for economists to adopt this policy approach, as the primary way to quantify gains and weigh them against costs. And structures rose to account for this.
JACQUELYN MARTIN/AP PHOTO
Models hold immense power over what bills get considered.
The Congressional Budget Office was established in 1974, with the goal of wrestling expertise out of the executive branch and back toward Congress. In the hands of its first director, Alice Rivlin, CBO became a go-to forecaster, beyond just presenting the budgetary outlook for particular bills. Its long-term estimates of deficits, interest rates, demographic shifts, expected economic and employment growth, and other economic indicators became one way that the public heard about the state of the nation.
CBO has a point of view, described in the Prospect in 2020 as short-run Keynesian and long-run classical. As Nick Hanauer explains in this issue, CBO deliberately minimizes the long-term economic benefits of public investments, building into the model the assumption that over time, public spending crowds out private spending. And the entire project of CBO, to assess policy solely in terms of its budgetary cost rather than its total effects on national income and quality of life, stresses the politics of deficits and debt over equality and prosperity. That perspective is broadly in line with other practitioners of macro models, both inside and outside of government.
These models reinforce, justify, and calcify a particular theory of change, backed by the same players that have been trying to embed a conservative, neoliberal ideology in Washington politics for decades. And it’s clear why this has been so successful: In the hands of a politician, an estimate that makes your policy look good or an opponent look bad can be extremely powerful. Even if the modeler makes caveats about ranges of estimates, having one number to use as a cudgel can be seductive. And that number can launch a thousand headlines, without the context or uncertainty that underlies it.
Every time a politician or media figure ties budgetary numbers to a particular policy, it reinforces the macro model’s importance. But when that evidence is rooted in assumptions and ideologies that ultimately damn progressive ideas to the dumpster, these models are only useful for those who agree with them.
Excavating the Assumptions
The economy, of course, is not a simple set of equations that can be neatly resolved with some elegant number-crunching. The economy is a complicated, messy system, featuring hundreds of millions of individual actors pushing and pulling in different directions. Boiling progressive policy proposals down to one number belies the complexity of systemic, long-standing crises, and gives policymakers a pass from actually grappling with that complexity and the inevitable trade-offs. Many have tried to map the economy scientifically with precision; many have failed, as you will see in this issue.
Though economists may not want you to believe it, the assumptions they build into their models to make the math work are actually embedding a perspective about the economy. Some of these assumptions can belie common sense. There’s the idea that investments in children or the climate that pay off over the long term are too costly in the short term to justify; or that distributional outcomes by race, gender, wealth, and other salient characteristics are irrelevant to policy considerations. Other assumptions fail to incorporate the last several decades of economic research, such as the idea that public investment and private investment are substitutes rather than complements, or the idea that concentration in goods production or labor markets doesn’t matter for long-run economic outcomes.
That these assumptions are divorced from reality is not just a technical matter. The lack of grounding in reality also means that these models are wrong. The Trump tax cuts were rolled out with the imprimatur from both CBO and Penn Wharton that they would expand the economy over the next ten years—CBO’s model predicted an increase in GDP of 0.7 percent, while Penn Wharton provided a range from 0.6 to 1.1 percent. Notwithstanding the point that the models couldn’t have foreseen the pandemic and the emergency economic measures taken to counteract it, if you control for that it remains clear that this did not bear out. Despite the macro religion, the tax cuts were simply not expansionary.
The fact that these assumptions are hidden under tangles of math has enormous implications. The lack of transparency means that massive decisions that affect the lives of millions are being justified using a set of equations that only a few people fully understand and control. And yet, when we look under the hood, it becomes clear that the same money and power that governs influence all over Washington has its hand in these technical gatekeepers as well. Ideologies should be challenged in the marketplace of ideas, not buried under a mountain of numbers.
Those who publicize the work of the macro modelers, whether in politics or the press, are only interested in the results, and have no interest in the way the sausage was made. They trust the figures as the product of impartial observers who simply apply the math. They don’t question the methods, or the biases. That’s what gives the models their power; it’s the way they are used, or abused, by the modern political system that gives them outsized sway over the policies that govern our lives.
Where Do We Go From Here?
In this issue, we dig inside two of the biggest modelers, the CBO and the Penn Wharton Budget Model, exposing their blind spots and in some cases their hidden ideologies. We detail the questionable assumptions built into macro models, and the ways in which economists have come to rely on a process that excuses them from looking at the real world. We give space to progressive economists and policymakers who have had to work under the world that the macro modelers built, constrained by the hidden handcuffs on their ambitions and the illogic that faulty models predict. And we try to sketch out another world, mindful of the boldness of the past but also the necessary perspective of the future, that can give economic analysis its proper place in the policymaking process: as an aid to problem solving, not a roadblock to overcome.
Leon Keyserling, President Harry Truman’s chair of the Council of Economic Advisers who was a major figure in constructing the Fair Deal, once wrote to his immediate predecessor, Edwin Nourse, complaining about the flaccid state of the economic profession. “While we economists have long talked in the refined atmosphere of theoretical underpinnings,” Keyserling said, “we live in a world where prices and wages and profits are being made.” That real world is where policy must emanate from, not through retreating to perfect dioramas of a model economy.
Our economic challenges continue to grow ever larger in scale and scope, in ways that crude algorithms of dubious quality cannot reach. Combating existential threats like climate change and racial injustice will require analyses that account for the benefits that come from shifting our economy fundamentally over the long run, even if they come with significant up-front costs. The models of the future will have to recognize that not everything can be boiled down into a simple cost-benefit analysis.
In other words, if tackling the challenges of the future requires us to rethink the trickle-down, neoliberal approach to policymaking, the models must follow.
