How Big Is the Long-Term Debt Problem?
By James Kwak
Articles about the deficits and the national debt generally talk about unsustainable long-term deficits that will drive the national debt up to a level where scary things happen. Sensible commentators usually acknowledge that our current deficits are a sideshow and the real problems happen in the 2020s and 2030s due to modestly increasing Social Security outlays and rapidly increasing health care spending. I admit that this has generally been my line as well; for example, in a previous post I said that the ten-year deficit problem is entirely a product of extending the Bush tax cuts, but that even if we let them expire things will get worse over the next two decades.
But looking at the numbers, it’s not clear that the long-term picture is really that bad. Here I’ll lay out the numbers, and then, as they say on Fox News, you can decide. The summary is the chart above; the details are below.
The most common source for long-term national debt scare stories is the CBO’s 2011 Long-Term Budget Outlook (LTBO), published in June. That report has two scenarios. In the extended-baseline scenario, which is pretty close to current law, the national debt reaches 76% of GDP in 2021 and 84% in 2035 (Table 1-2, p. 8)—bad, but not that bad. In the alternative fiscal scenario, which is intended to be more realistic (and which everyone cites for dramatic effect), the debt reaches 101% in 2021 and 187% in 2035—the latter being a level that virtually everyone would consider too high.
But things have changed since June, notably because of the Budget Control Act signed in August. In the CBO’s Budget and Economic Outlook: An Update (BEO), published in August, the 2021 national debt in the baseline scenario (which the CBO uses as the first part of its long-term extended-baseline scenario) is down from 76% to 61% of GDP (Table 1-2, pp. 4–5). We can also tell what the alternative fiscal scenario’s updated 2021 projection would be by reapplying the adjustments included by the CBO in that scenario (extension of all expiring tax cuts, Medicare doc fix, and Iraq/Afghanistan drawdown); in that case, the projected 2021 debt has fallen from 101% of GDP to 80%.* (More on this later.)
So far this is is child’s play, since I’m just adding together numbers from different CBO tables. But we’re really interested in 2035 for now. The questions I want to answer are: first, what would the CBO’s new extended baseline say for 2035? and second, what is a realistic 2035 forecast?
2035 Extended-Baseline Scenario
The task here is to apply the same method used in the LTBO extended-baseline scenario to the new numbers. On the revenue side, revenues grow slowly because of real bracket creep (tax brackets are indexed for inflation, but over time average wages rise faster than inflation) and demographic trends. In the LTBO, the growth rate of revenues as a percentage of GDP from 2021 to 2035 is 0.78 percent per year (calculated from the supplemental data), so I use that in all of my projections. (That’s 0.78 percent per year, not 0.78 percentage points per year.) Under that assumption, revenues grow from 20.9% of GDP in 2021 to 23.3% in 2035.
On the spending side, the LTBO breaks spending into four categories: Social Security; Medicare, Medicaid, CHIP, and exchange subsidies (most but not all health care entitlements); other noninterest spending; and net interest. There’s no reason to change the June baseline estimates for the first two categories (6.1% and 9.2% for 2035, respectively). For other noninterest spending, the LTBO approach is to assume that this category remains constant as a share of GDP except for a few adjustments that make it fall slightly between 2021 and 2035, from 8.3% to 7.8%. But because of the Budget Control Act, this category has already fallen to 7.7% by 2021, so in my spreadsheet I just leave it at 7.7% through 2035.** Given the 2021 debt, the 2021 budget, a real interest rate of 2.7% (LTBO, p. 17), a real growth rate of 2.1% (LTBO supplemental data), and an inflation rate of 2.5% (LTBO supplemental data), you can project out future budgets. Net interest spending in 2035 turns out to be 3.0% of GDP. Total spending is 26.0% of GDP, the deficit is 2.7%, and the national debt is only 59% of GDP—smaller than it is today (and down from the 84% predicted in June).
2035 Realistic Scenario
Now, everyone knows that the CBO baseline is not realistic, so it’s time for the adjustments. First we apply the uncontroversial ones: the AMT patch, the Medicare doc fix, and the more rapid drawdown of troops from Iraq and Afghanistan. As I’ve pointed out before, these basically cancel out over ten years, and the 2021 national debt stands at 60% of GDP (down from 61% in the baseline). The CBO’s alternative fiscal scenario also assumes that discretionary spending grows with GDP before 2021, rather than with inflation as in the baseline scenario. But the Budget Control Act capped discretionary spending growth below the rate of inflation. So if we assume that Congress follows the law it set for itself, then discretionary spending remains low, even in the alternative scenario.
Going out to 2035, revenues are starting at a lower base (20.3% of GDP). Applying the same growth rate as in the LTBO, they grow to 22.7%.
For spending, first I use the same Social Security projection as in the extended baseline. (The LTBO uses the same numbers for both scenarios.) For health care spending, I use the numbers from the LTBO’s alternative fiscal scenario—the more pessimistic ones. Other noninterest spending is where things get interesting. In 2021, this category is down to 7.1% of GDP because of the Budget Control Act (which capped discretionary spending growth below the inflation rate) and the Iraq/Afghanistan drawdown. Again, I assume this remains constant as a share of GDP (not falling slightly, as in the LTBO).
With these assumptions, you get 2035 spending of 26.9% of GDP, a deficit of 4.2%, and a national debt at 69% and growing slowly. That’s with business as usual for Social Security, Medicare, Medicaid, and the health insurance exchanges. You could make a strong case that this is a problem, especially if you worry (like I do) about things like major financial crises, and the national debt should be lower than 69% of GDP. But it’s not obviously, unequivocally a national emergency. And the idea that we have to gut entitlement programs doesn’t make sense.
(There are two other reasons this is a conservative (high) estimate. First, I use the CBO’s assumption that the Joint Select Committee reduces deficits by a total of $1.2 trillion between now and 2021. But I assume that they do so in a way that does not affect deficits in any year after 2021, which is highly unrealistic. Second, since this realistic scenario has lower debt levels than in the CBO extended baseline, economic growth should be slightly higher on the CBO’s theory of the world. (See LTBO, chapter 2.))
Now, there is a big problem in this world: it’s that the rest of the government has been crippled by the Budget Control Act. By 2021, discretionary spending will be down to 5.5 percent of GDP—the lowest level since 1931. I’m not saying this is a happy outcome. I’m just saying it’s an outcome where the national debt is probably not our biggest problem—it’s the fact that we’ve eliminated almost everything except Social Security and health care programs.
2035 Realistic Scenario with Tax Cuts
Of course, everything gets worse if you assume that the Bush-Obama income and estate tax cuts are made permanent, as the CBO does in its alternative fiscal scenario. In that case, the 2021 national debt is up to 76% of GDP.*** 2021 revenues are only 18.4% of GDP; growing at the same rate as above, they only make it to 20.5% by 2035. Spending is the same as above, except that since deficits are so much bigger every year between now and then, total spending (including interest) is 29.3% of GDP in 2035. The deficit is 8.7% of GDP and the national debt is 117% and growing fast. I think most people would agree that that is a big problem.
2035 Alternative Fiscal Scenario
We can update the CBO’s alternative fiscal scenario, too, as I showed two figures above. It’s just the previous scenario (realistic with tax cuts), plus extending all other expiring tax cuts (which I think is a dubious assumption for reasons I won’t go into here), plus an assumption that tax revenues remain at their 2021 share of GDP indefinitely. (That is, Congress periodically cuts tax rates to compensate for a slowly expanding real tax base.) I think this assumption is gratuitously pessimistic (from a budgetary standpoint). Basically, it is an example of assuming your conclusion: if you want to conclude that we’re going to have a huge deficit problem, there’s no better trick than assuming that Congress constantly makes sure that tax revenues are too low.
In that scenario, the national debt in 2035 is 142% of GDP and climbing rapidly.
So the bottom line is: If we extend the Bush tax cuts, we have very big deficit problems over the next ten years and the next twenty-five years. If we let them expire, there is no ten-year problem. That’s the same as in my earlier post, and I don’t think that’s controversial to anyone who understands the numbers.
What’s more controversial is my claim that if we let the tax cuts expire, there is a twenty-five year problem, but it’s not a huge one. Many other people argue that even if we let the tax cuts expire, we still have to cut Social Security and Medicare. On my reading, the problem is a national debt at 69% of GDP and growing steadily. If we have another financial crisis, or we start losing our status as the reserve currency, that could be a serious problem. My opinion is we should do something about it. But it’s not necessarily the end of the world.
I’ve already pointed out a few places where my estimates are conservative, so I should also point out a couple of ways I could be underestimating future deficits (other than the obvious one that the tax cuts will probably be extended). Probably the biggest is that I import the CBO assumption that “other noninterest spending” will be a constant share of GDP after 2021—after it has already been slashed by the Budget Control Act. There is probably a fair chance that Congress will find a way to repeal those caps even before 2021.
Another, smaller issue is that I don’t have an allowance for “other means of financing” to increase deficits after 2021. In theory OMF shouldn’t systematically increase or decrease the debt (except a little bit, since financing accounts need to grow a little as the economy grows), but maybe there’s a practical reason why it usually ends up adding to deficits. If so, can someone let me know?
Of course, the interest rate and growth rate assumptions (from the CBO) could be wrong, but they both look conservative to me, especially a 2.1% real growth rate.
I happen to think we do have a long-term deficit problem. I just think it may be somewhat smaller than the conventional inside-the-Beltway wisdom would have you believe.
For those interested, I’ve attached my spreadsheet with all of the scenarios discussed above. It includes the CBO spreadsheets that I used as my starting point.
* The BEO provides a 2021 debt estimate of 82 percent of GDP under certain alternative policies (pp. 2–3), but it uses different adjustments than the LTBO. The BEO does not include the Iraq/Afghanistan drawdown, which is included in the LTBO, and it extends fewer tax cuts than the LTBO.
** This is a conservative assumption; if I mimicked the LTBO exactly, it would fall from 7.7% to 7.4% or lower.
*** This is basically the 82% figure on page 3 of the BEO, less the deficit savings from the Iraq/Afghanistan drawdown.