I testified today before the House Financial Services Committee at a hearing the Republicans called “The perils of ignoring the national debt.” As I tweeted earlier today, the hearing was delayed a hour because the R’s went off to vote on a tax cut adding another $100 billion to the 10-year debt. That bill is unlikely to get very far in the Senate, but it does raise a somewhat existential question about the urgency of all this hand-wringing about the debt.
Here’s my testimony; see the intro for a summary and bullet points. There are two points I’ll highlight here. First, see the discussion of “revealed preferences.” I hear a lot of chin music from all sides about how much they really want to cut spending, yet they rarely do so, and, to the contrary, are often quick to raise spending on their preferences, many of which, to be clear, I share.
But I was taught to pay a lot more attention to what people do, not what they say, which economists label “revealed preferences.” Thus, I conclude in my testimony that:
Spending and “revealed preferences:” My point in these NDD and defense discussions is that Congress’s “revealed preferences” suggest our deficits are born of an unwillingness to raise the revenue we need to meet the spending we believe is warranted. To label that a “spending problem” is fundamentally inaccurate.
While hearings like this show we can have substantive disagreements on the extent of those needs and spending levels, spending deals don’t sign themselves. This reality, in tandem with the existence of large, persistent deficits and debt, reveals that year upon year, Congressional majorities accept existing spending levels. However, the current Congressional majority has not only been unwilling to raise the revenues necessary to pay for its revealed preferences, it has significantly cut its revenue inflows.
The essence of our fiscal problem is thus neither a revenue problem nor a spending problem. Instead, it is this: Congress has long been unwilling to raise the resources necessary to pay for the institution’s revealed preferences. Given that framing of the problem, policy makers must either reduce Americans’ expectations about the role of government in our economy and their lives or, over the long-term, raise the revenues necessary to meet those expectations.
Next, a quantitative analysis of budget forecasts suggests that the logic of a framing our fiscal imbalance as a spending problem, full stop, is indefensible. I paste that section in below and urge you to take a good look at the Figure 4 (the third figure below). It’s dense, I grant you, but I think it’s persuasive.
Washington, we have a revenue problem
The above observations show the “spending-problem-only” framework to be clearly misguided. In fact, one of my biggest concerns about the impacts of the Republican tax plan is the extent to which it has broken a critical linkage in public finance: that between a full-capacity economy and lower deficits and debt through higher revenue flows. (Because it so heavily favors wealthy households, the other major concern is the extent to which the tax law exacerbates after-tax income inequality.)
Figure 2 below shows that because more economic activity—lower unemployment in the figure—has historically spun off more tax revenue, as the economy has closed in on full capacity, the budget deficit has gotten smaller, and vice versa (deficits are shown as positive shares of GDP).  Outside of major wars, and before the 2017 tax law, when economic growth led to more employment and diminished labor market slack, deficits typically came down. In fiscal year 2000, for example, the unemployment rate was 4 percent and the budget was in surplus.
In fact, using data back to the mid-1940s, the average deficit as a share of GDP over every year that the unemployment rate was lower than or equal to 4.5 percent comes to -0.4 percent. If I take last year’s and this year’s deficits (-3.5 and -3.8 percent) out of that average, the result is a small surplus (0.1 percent).
The end of the figure shows just how different our fiscal stance is today and, based on CBO projections, in the future. The unemployment rate is well below 4 percent, but the deficit, also about 4 percent, is far above its average at low unemployment. In fact, a simple regression of the deficit-to-GDP ratio against unemployment predicts a deficit of about 1 percent in FY18, almost 3 percentage points below its actual value.
Is this divergence driven by a negative shock to revenues or a positive shock to spending? CBO data reveal the answer to be a negative revenue shock. The figure below shows that in the summer of 2017, before the tax cuts and spending deal, the budget office predicted that we’d spend 20.5 percent of GDP in 2018, which, as the actual spending bar shows is almost exactly the right number (20.3 percent). However, CBO also thought — remember, this is pre-tax-cut — that we’d collect 17.7 percent of GDP in revenues when the actual share was, as shown, just 16.4 percent. This diminished revenue figure is the key difference between what CBO expected then and what occurred. In fact, the spending share—20.3 percent of GDP—is precisely equal to the 50-year average.
To be clear, the point of this comparison is not to argue that our current spending of about 20 percent of GDP is optimal (though it is “average” in historical terms). Instead, it shows that the jump in the 2018 deficit was a function of the tax cuts leading to diminished revenue collection, an outcome that is especially disappointing given the near-full-capacity economy.
A comparison of CBO long-term projections further underscores the revenue shortfall point. Figure 4 shows CBO’s forecasts for primary outlays (outlays other than interest payments) and revenues from two different vintages of their long-term budget outlook, one from 2010 and the most recent, from 2018. The logic of the “spending problem” case implies projected outlays should be accounting for a higher share of GDP in the 2018 projection, with the projected revenue share either higher or similar to that of the earlier forecast. That is, the “spending problem” scenario should show fiscal gaps being driven by more spending, not less revenues.
In fact, the opposite is the case. Not only are primary outlays lower in the 2018 than the 2010 forecast—by 2 percentage points of GDP, on average, over the forecast period—but revenues in the 2018 budget outlook are much lower than in 2010’s outlook—by 4.5 points, on average. And they would be even lower were Congress to extend the Trump tax cuts. In other words, current law in 2010 (not all of which was followed, to be clear; i.e., the George W. Bush tax cuts did not fully sunset) called for both higher spending and higher revenues than today’s current law. And given that the revenue decline between these two forecasts has been more than twice that, on average, as the primary spending decline, the “spending problem” framework is not defensible.
These are forecasts, but a longer-term analysis of the actual path of revenues and outlays by Robert Kogan of the Senate Budget Committee staff makes a similar point. Had we kept the Clinton-era tax code in place—meaning no George W. Bush or Trump tax cuts—but let all the spending that occurred since then proceed apace, including the military actions, the Affordable Care Act, and so on, the result would be a debt-to-GDP ratio that is 27 percentage points, or about a third, below where it stands today, about 51 percent instead of 78 percent.
 Paul Van de Water, “2017 Tax Law Heightens Need for More Revenues,” Center on Budget and Policy Priorities, November 15, 2018, https://www.cbpp.org/research/federal-tax/2017-tax-law-heightens-need-for-more-revenues#_ftnref16.
 The regression runs from 1949 to 2017 with the unemployment rate at time t and with one lag, along with a lag of the dependent variable. All coefficients other than the constant are significant at the p<0.01 level; R-sq=0.81 and DW=1.77.
 Primary outlays are an appropriate choice here because the “spending problem” refers to programmatic spending. Consider a revenue increase with unchanged program spending. Thanks to the higher revenues, deficits and debt service would fall, even with no spending cuts. However, for completeness, I show the same figure with total outlays in an appendix.