America’s long-run fiscal problem is spending growth, not taxes

America’s long-run fiscal problem is spending growth, not taxes

Yesterday I wrote about the history of tax increases since World War II, and about the battle over the total level of taxation. Now I want to turn to spending.

I am a low-tax guy. I have worked on tax issues for 12 of my 15 years in Washington, helping elected officials lower taxes and prevent tax increases. You can see a list of the taxes President Bush cut here. I would like to cut taxes far below where they are today, and I will continue to make the case that America is better off with a bigger private sector and a smaller government. America’s long run fiscal debate, however, is instead principally about whether we will allow future spending increases to force taxes to increase dramatically above where they are today.

I believe that America’s greatest economic policy challenge is the projected long run growth of spending on three programs: Social Security, Medicare, and Medicaid.

I would like to show you why I believe this, and how it relates to taxes.

Let’s begin by looking at federal taxes over time, as we did yesterday. You can see that they bounce around quite a bit. Some of that is from changes to tax law by Congress, and some because total revenues track economic growth fairly closely. What is remarkable about this graph is that the Federal government’s take from the U.S. economy has remained basically flat since the end of World War II.

federal taxes

All that fluctuation is distracting. I will remove it and leave just three trend lines for reference. By doing so I remove much noise and lose little information.

The blue and green dotted lines below represent averages for different periods: the green line is the 50-year average, at 18.1% of GDP. The blue line is the average of the last 30-years, at 18.4% of GDP. The post-WWII average is below the blue dotted line, at 17.9%.

The red dotted line is a trend line. It shows a 5-year moving average of the graph above. There is a very gradual upwards slope, growing about 0.025 percentage points per year (that’s two and one-half hundredths of a percent). Since the end of World War II, total federal taxes have therefore increased at a rate of about 1 percentage point of GDP every 40 years.

federal taxes plus trend

Now I will make a judgment call. I argue that the remarkable flatness of these lines is the result of a broad-based policy consensus, or at least a long-term balancing point of political forces, of how much in total taxes we as a nation are comfortable taking from those who earn it and giving to the federal government. Since World War II, between 17 cents and 19.1 cents of each dollar earned have gone to the federal government. The average is between 18.1 cents and 18.4 cents, depending on what time frame you choose. The number has been creeping up, by about one penny per dollar each 40 years. The highest that trend (the 5-year moving average) has ever been is just under 19.1 cents per dollar. My personal preference would be well below any of these numbers.

I will now take those trends and extend them for another 70+ years, to 2080. I am letting the red line grow at its historic long run trend of +1 percentage point every 40 years. I hate that trend.

federal taxes long-run trends

What is nice about this graph is that we can zoom out from discussions of specific tax policies, and instead just focus on the total level of taxation. When current law has allowed taxes to creep up above the low 18s, the political pendulum swings and Congress “cuts” taxes. In aggregate terms, they are not really cutting taxes, they are instead preventing taxes from increasing as a share of GDP (which we learned yesterday is still an increase in real dollars taken by the government). Since the end of World War II, our political system has dynamically adjusted to change current law as needed to keep taxes in the low 18s as a percent of GDP. Unfortunately, the same appears to be highly unlikely for future spending.

Let us turn to historic total federal spending, again with a 5-year moving average trend line.

federal spending plus trend

I am going to take the spending trend line and superimpose it on our earlier long-run tax graph. Yellow is spending, and all other colors are taxes:

taxes and short-run spending

You can see that we usually run budget deficits, since the yellow spending line is generally above the red tax line. If spending is greater than taxes, then the government is running a deficit and has to issue Treasury bonds to borrow funds from the private sector. The bigger the gap between the yellow spending line and the other tax lines, the bigger the budget deficit. Bigger deficits are bad for the economy. So are higher taxes.

Now all we need to do is add projected spending under current law for the long run. I will draw it in white.

taxes and spending long term trends

Whoa! Hold on. What happened? The scale of the vertical axis changed, and that white long-run spending line is ridiculous. Surely that can’t be correct?

It is correct. If I could have you remember only one thing about economic policy, it would be this graph. You can see that the current law spending trend is clearly unsustainable in the long run. Under current law, total federal spending will grow steadily, reaching 40% of GDP by 2080.

The exact slope of that white spending line depends on a lot of assumptions, and there are esoteric debates about those assumptions. Some analysts would have it reach about 36% by 2080, and others in the mid 40s. Those are huge differences, but everyone’s line ends up looking similar, and the basic conclusion remains unchanged – the long-term spending line slopes up dramatically. The tax lines remain basically flat, as they have since the end of World War II. The difference between these slopes creates an unsustainable borrowing trend that, if left unchecked, would eventually cause the U.S. economy to collapse.

The white spending line I show here was done for me by OMB staff in 2007. Since then some spending assumptions have changed a bit, and we also have a new President. The dip you see in the early years of the white line will be shallower than you see here, because near-term spending will be higher than we assumed in 2007. Neither of these changes modify the basic conclusion.

For now, I will ask you to trust me that the increase in the white spending line is driven by the growth of three entitlement spending programs: Social Security, Medicare and Medicaid. I will show this later.

I posted yesterday about the battle on total taxation. The two ends of that debate, House Republicans and House Democrats, are separated by 1.4% of GDP on how much federal taxes they would collect. The House Republican plan would collect just under 18% of GDP in taxes. The President’s budget would collect just under 19%, and the budget passed through the House by Democrats would collect just over 19%. In each case, however, that is just setting the level of the flat flat line. Nobody is talking about a permanent upward change in the slope of tax lines, so you’re arguing about whether you should be a smidge below that flat blue line, or up to 0.8 percentage points above the flat green line. Either way, they are all flat tax lines.

Even the long-term red tax trend line, which slopes slightly upward and which I hate, hits 19.4% in 2080. That is higher than taxes have ever been for a sustained period of time, and it is less than half of projected spending in that year.

Even the highest level of taxes imagined in the current short term debate does no good in addressing this long-term fiscal problem, because the current law spending line will just keep growing above that, forever. At some point you have to bring the spending and tax lines together, or at least to be roughly as close as they have been in recent history. While undesirable, we can sustain deficits in the 2-ish% of GDP range indefinitely without the economy collapsing. We cannot sustain deficits that grow and keep growing as a share of the economy.

To close that gap, we have to reduce the slope of the spending line. Raising taxes as a share of the economy is an upward shift in the flat tax lines. This can delay the day of reckoning, but it cannot solve the problem. And raising taxes causes economic damage of its own.

America’s long run fiscal problem is the projected future growth of spending under current law. If we do not prevent that white line from growing as you see above, our economy will eventually implode, no matter what we do with taxes.

18 responses

  1. Our long term obligations to Social Security, Medicare and Medicaid is the 8,000 lb. elephant in the room. It's a problem that any thinking policy analyst and politician knows exists, but don't seem to address in meaningful ways, partially due to our constant preoccupation with short term political cycles and crisis such as the war in Iraq and our current economic woes.

    I look forward to hearing your continued thinking on this subject Keith.

  2. Thank you for the very understandable explanation. Politicians have not usually shown a willingness to get out ahead of their constituents and ask for sacrifice today to avoid greater future pain. Is it possible that enough voters will recognize spending growth as an issue to convince the political class to take preemptive action (as is being done today with climate change – whether or not it is real)?

  3. I hope you left a copy of these graphs behind in Larry Summers' office.

    I only disagree with this:

    'Since the end of World War II, our political system has dynamically adjusted to change current law as needed to keep taxes in the low 18’s as a percent of GDP.'

    Aside from the Reagan years, I think the adjustment mostly came from the people adapting their behavior to the tax code. I remember being told, in the 1970s, by a mailman, that he didn't want to work overtime, because after taxes the money wasn't worth the extra effort.

    And, he was a Democrat.

  4. I think this is a great in depth explanation of what many have been saying for awhile, and you put it in exactly the right context. But I think that the debate has moved on from here a little bit.

    The natural inclination of a Fiscal Conservative, upon seeing these numbers, is to call for private accounts on Social Security, or other changes that effectively eliminate these entitlements as a major component of future spending obligations. But many of my more liberal friends point out that the more likely alternative is that we keep making smaller reforms to the existing entitlements. I would be interested, then, to see what would happen if some of these smaller (procedural rather than structural) reforms were better modeled out.

    For example, what would happen if the eligible age to collect Social Security and Medicare was pushed out periodically? If we added an additional year to the minimum age every decade, would the spending gap decrease substantially? If not, what thresholds of eligibility would be necessary to close the gap to a reasonable level?

    On the one hand, I can understand how difficult it is to model the future based on legislation that doesn't exist. But on the other hand, as I said, most people already acknowledge that these entitlements are unsustainable with their current structure. While you and I would call for substantial reform, Economic Liberals can shut down debate on by asserting that the spending gap "crisis" is easily averted by periodic fine-tuning of the programs. So I would like to see us move this debate one step ahead, and see if these counter-arguments would really work.

  5. Keith,

    First, thanks for creating this blog. Regardless of whether or not I agree with your positions, I commend the way you lay out the relevant concepts and facts and generally present a clear, step-by-step argument in terms non-experts can easily understand.

    Now on to your post, and I hope you don’t mind a long comment.

    Many among the general public interpret statements such as “America’s long run fiscal problem is spending growth, not taxes” as an implication that we should solve our fiscal imbalance problem solely by reducing projected spending. I don’t know whether or not you intended such an implication, but if so, that argument would be presenting a false dichotomy. The choice we face is not between solving the long-term fiscal imbalance problem solely by increasing projected revenues or solely by reducing projected spending, but rather a choice among those two options or a combination of the two. Even if “spending growth is the problem”, it does not necessarily follow that reducing spending growth alone is the best solution.

    You are indeed correct (and few, if any, informed people would disagree) that our long-term projected deficits, which are unsustainably large vs. revenues per current tax policy and even more vs. revenues at historical averages as a percent of GDP soand debt-to-GDP per current policy, are driven by spending growth, and that this spending growth is driven mainly by entitlements (Medicare, Social Security and Medicaid, in that order).*

    And I believe the vast majority of Americans, if reasonably well-informed, would prefer (and rationally so) some reduction in overall projected government spending over letting effective overall taxation rise to around 40% of GDP, even knowing that those reductions in projected spending mean real sacrifice for at least some segments, not just painless cost reductions like efficiencies from digital medical records, elimination of “waste, fraud, and abuse”, etc. Not only would most Americans object on libertarian grounds (whether or not they’ve ever heard the term “libertarian” if forced to hand over so much of their own income, but if a taxation-only (revenue-only) approach were tried based on extremely heavy taxation of “the rich” (broadly-defined), it would probably hurt GDP enough (and experience enough diminishing returns in terms of incremental revenue due to dynamic effects) that most Americans would prefer some reduction in spending over that much higher taxation of “the rich”.

    Given the above assumptions, most serious, reasonably well-informed people (experts and laypeople), with the exception of some hardcore hyperpartisans, accept that we need policy changes that will achieve some reduction in projected spending, and beyond just supposedly (and in some cases actually) fairly painless, win-win efficiency gains.

    But it does not follow that we should solve the fiscal imbalance problem entirely on the spending side, as opposed to via some combination of sacrifices on the spending and taxation sides. For one to prefer and advocate such an approach, he would have to consider and present what policies would have to be changed in what ways to what degree (e.g., raising the retirement age to X, means testing Social Security and/or Medicare to cut out Y% of seniors, etc.) to achieve what revised projected spending levels, and contend that making all of those changes, with all the adverse they would have on people, were preferable to even a bit more taxation than the level he advocates. I don’t see any such argument in your post (although perhaps you plan to present it in a subsequent chapter in this series of posts), and, based on my sense of the magnitudes of spending “cuts” that would be necessary, I don’t think most Americans would prefer such an approach, and I don’t consider it politically plausible.

    So, based on economics, the priorities and values of most Americans, and political feasibility, I believe a realistic approach must involve a combination of reductions in projected spending and tax increases to increase projected revenues. The discussion/debate should be over how much of each, and in what manner (how to raise projected revenues and how to reduce projected spending), would be ideal. And I look forward to your thoughts on these matters.

    I also have some technical (but significant) questions. In your chart and in your argument you present “projected spending under current law for the long run”, and I realize that, as you noted, spending assumptions for the short term have changed somewhat, and I realize that great precision is not necessary for the purposes of this discussion. But a couple of assumptions can affect the slope of that long-term projected spending line to more than a minor degree. First, does it include interest expense? Second, if “current policy” is the basis, what is assumed regarding the AMT and expiration/extension of the Bush tax cuts? I realize that my second question pertains to revenue rather than spending, but if interest expense is included in your spending projection, than it is partly a function of deficits, which are a function of revenues as well as non-interest spending. If we assume that AMT is not patched, that all the Bush tax cuts expire, or even that taxes are substantially increased beyond that point, and thus we substantially increase projected revenues and substantially reduce projected deficits, then we substantially reduce projected interest expense and, in turn, projected total spending.

    * As a note, our fiscal health (debt-to-GDP) would deteriorate even faster than official projections imply, because such projections — such as CBO’s — typically do NOT factor in the reality that as deficits and debt-to-GDP grow substantially, interest rates would be higher, ceteris paribus, thus increasing interest expense (and thus total spending) and reducing GDP and revenues, ceteris paribus, (and higher transfer payments) all of which means that all factors in debt-to-GDP worsen (numerator, because of lower revenues and higher spending, and denominator, because of lower GDP). And of course those dynamics feed back and snowball. Rather than factoring in these dynamics, CBO, in its Long-Term Budget Outlook simply assumes a single set of annual GDP growth rates for all scenarios of varying levels and types of taxation and various levels and types of spending, and these annual growth rates certainly do not seem to reflect the kind of economic implosion that would eventually occur if continue for several decades per current policy and start to experience the projected growth in deficits and debt-to-GDP.

  6. As a note re: my comment above, I do assume that it is possible for a substantial portion of incremental revenues resulting from tax increases to cause lower deficits, ceteris paribus, as opposed to simply spurring an equivalent amount of incremental spending (with little to no reduction in deficits, ceteris paribus). In other words, although higher revenues may “feed the beast” to some extent (cause some degree of incremental spending), the dynamic is not necessarily dollar-for-dollar, and if part of a deficit-reduction deal, would probably not be anywhere close to dollar-for-dollar.

  7. Pingback: Kicking Over My Traces: Keith Hennessey’s Federal Spending Skyrocket

  8. Pingback: Stop them before they spend again |

  9. Pingback: How to measure health care cost control |

  10. Pingback: Kentucky Club for Growth

  11. Pingback: Miscellaneous Ramblings » America’s long run fiscal problem is spending growth, not taxes by Keith Hennessey

  12. Pingback: A Chart That Should Keep Progressives Up at Night | Progressive Fix

  13. Pingback: Spending is the Problem | TightWind

  14. Brooks: Too many moving parts. This is why, to quote the old saying, ” In Washington, nothing is ever done about a problem until there’s a crisis.” In my opinion, people are losing faith with the political and intellectual elites who are supposed to foresee and forestall the crisis. Therefor, there will be an explosion; or maybe a collapse.

  15. Pingback: Correct! Sort Of… « noot

  16. Pingback: A Chart That Should Keep Progressives Up at Night


Get every new post delivered to your Inbox.

Join 6,544 other followers