LOS ANGELES — AS Republicans take control of Congress this month, at the top of their to-do list is changing how the government measures the impact of tax cuts on federal revenue: namely, to switch from so-called static scoring to “dynamic” scoring. While seemingly arcane, the change could have significant, negative consequences for enacting sustainable, long-term fiscal policies.
Whenever new tax legislation is proposed, the nonpartisanCongressional Budget Office “scores” it, to estimate whether the bill would raise more or less revenue than existing law would.
In preparing estimates, scorekeepers try to predict how people will respond to a new tax law. For example, if Congress contemplates raising the excise tax on cigarettes, scorekeepers consider existing trends in cigarette consumption, the likelihood that the higher taxes will induce some smokers to quit, and the prospect that higher prices will increase incentives for cigarette smuggling. There are no truly “static” revenue estimates.
These conventional estimates do not, however, include any indirect feedback effects that tax law changes might have on overall national income. In other words, they do not incorporate macroeconomic behavioral changes.
Dynamic scoring does. Proponents point out, correctly, that if a tax proposal is large enough, then those sorts of feedback effects can aim the entire economy on a slightly different path.
Such proponents argue that conventional projections are skewed against tax cuts, because they do not consider that cutting taxes could lead to higher economic output, which would make up at least some of the lost revenues. They maintain that dynamic scoring will, therefore, be both more neutral and more accurate than current methodologies.
But the reality is more complex. In order to look at the effects across the entire economy, dynamic modeling relies on many simplifying assumptions, like how well people can predict the future or how much they care about their children’s future consumption versus their own.
Economists disagree on the answers, and different models’ predicted feedback effects vary wildly, depending on the values selected for those uncertain assumptions. The resulting estimates are likely to incorporate greater uncertainty about the magnitude of any revenue-estimating errors and greater exposure to the risk of a political thumb on the scale.
Consider the nonpartisan scorekeepers’ estimates of the consequences of a tax-reform bill proposed last year by Representative Dave Camp, Republican of Michigan. Using different models and plausible inputs, the scorekeepers estimated that, under the bill, total gross domestic product might rise between 0.1 percent and 1.6 percent over the next decade — a 16-fold spread in projected outcomes. Which result should be the basis of congressional scorekeeping?
But the bigger problems lie deeper. Federal deficits are on an unsustainable path (as it happens, because of undertaxation, not excessive spending). Simply cutting taxes against the headwind of structural deficits leads to lower growth, as government borrowing soaks up an ever-increasing share of savings.
The most optimistic dynamic models get around this by assuming that the world today is in fiscal equilibrium, where the deficit does not grow continuously as a percentage of gross domestic product. But that’s not true. If you add the reality of spiraling deficits into those models, they don’t work.
To make these models work, scorekeepers must arbitrarily assume either that we tax more and spend less today than is really the case — which is what they did for the Camp bill — or assume that a tax cut today will be followed by a spending cut or tax increase tomorrow. Economists describe such a move as “making counterfactual assumptions”; the rest of us call it “making stuff up.”
In practice, these models are political statements. They show the biggest economic effects by assuming that tax cuts are financed by unspecified future spending cuts. The smaller size of government, not the tax cuts by themselves, largely drives the models’ results.
Further, the models are not a step toward more neutral revenue estimates, because they assume that, while individuals make productive investments, government does not. In reality, government spending contributes significantly to economic output. Truly dynamic modeling would weigh the forgone economic returns of government investments against the economic gains from lower taxes.
The Republicans’ interest in dynamic scoring is not the result of a million-economist march on Washington; it comes from political factions convinced that tax cuts are the panacea for all economic ills. They will use dynamic scoring to justify a tax cut that, under conventional scorekeeping, loses revenue.
When revenues do in fact decline and deficits rise, those same proponents will push for steep cuts in government insurance or investment programs, because they will claim that the models demand it. That is what lies inside the Trojan horse of dynamic scoring.