The Bond Market Wins
The Bond Market Wins
This article originally appeared in the January 2012 edition of Econ Journal Watch.
Download the PDF of "The Bond Market Wins."
I used to think if there was reincarnation, I wanted to come back as the president or the pope or a .400 baseball hitter. But now I want to come back as the bond market. You can intimidate everybody.
—James Carville, The Wall Street Journal, February 25, 1993
The United States is on an unsustainable fiscal path. How will the federal government address its long-term fiscal problems? There are four possible tools: Higher revenues, lower spending, inflation, and default. The last two hurt bondholders, and are the least likely.
The most likely scenario is one of chronic fiscal contrition, like the 2011 debt limit negotiations, but reenacted every few years. American leaders—driven by rating agencies, examples from overseas, and perhaps a pro-austerity branch of government—huddle together and lop a semi-credible $2 trillion off of future deficits.
This is just one scenario, and others are possible. It’s too early to tell how the bondholders are going to stay whole, but their victory is extremely likely. The long-run fiscal crisis is likely to be managed through a combination of tax increases and spending cuts. Below I discuss the most likely cause of default were it to occur: A GOP-led defense of the taxpayer. But in any case, a tipping point, where a AA+ or AA rated U.S. government is unable to borrow one week, is quite unlikely.
In predicting that bondholders will stay whole, I am disagreeing with the view of Laurence Kotlikoff and Scott Burns, who argue in The Coming Generational Storm (2004) that the soft default of inflation will be part of the solution to our looming fiscal crisis. I take this position seriously, but come to a different set of conclusions.
Why No Default? Why No Inflation?
One piece of evidence that U.S. default and high inflation are unlikely comes from the bond market itself: Real and nominal interest rates on long-term government bonds are exceedingly low. Investors have not charged the federal government the rates they charge high-risk countries. Further, U.S. long-term real and nominal rates have fallen since the financial crisis, a crisis that hurt expected tax revenues and made it clear that the U.S. government’s ability to pay its debtors had deteriorated. If the bond market is deeply worried about the U.S. government’s ability to pay its long-run obligations, it is not using market prices to demonstrate its concern.
Another reason to doubt default or hyperinflation is the maturity structure of Treasury debt. The overwhelming majority of US debt is in Treasury notes, investments that last from one to ten years. In recent years, the average loan to the U.S. government has lasted approximately five years. This means that the U.S is rolling over one-fifth of its debt every year. This gives the market a chance to vote on the federal government’s future creditworthiness on a regular basis.
Let us consider an extreme case: If the U.S. planned to inflate or default its way out of the debt in 2020, and the market foresaw that in 2015, markets would instantly stop lending extra money to the U.S. for any period longer than five years, or only lend to it at exorbitant rates of return. Financial markets would bring the planned future default to the present. So planned future defaults, whether hard (partial or total repudiation) or soft (inflation), are almost impossible.
What about partially anticipated defaults? And what about default plans that the markets only begin to anticipate a year or two in advance? In those scenarios the benefits of defaulting are still much lower than an almost-impossible abrupt, surprise default. One reason the benefits of default are likely to be lower than in an abrupt-surprise scenario is that nations that default almost always only partially default. Partly because they want to return to credit markets as soon as possible (especially to borrow in the short-term market, which cities and states use to cover payroll before taxes are paid), nations are reluctant to completely destroy their reputations as borrowers.
Taken together, this means that the incentives to default or inflate are weaker than many imagine. And since market participants quite likely know this fact already, an overnight fiscal crisis driven by fears of inflation or outright default is less likely than many imagine.
Consider the following scenario: For some reason, markets, all of a sudden, foresee future possible default, and suddenly stop lending to the U.S. as a result. The U.S. is unable to roll over its debt one week, fails to get aid from other countries, and defaults. How unlikely is that possibility? Again, market prices and rating agencies alike tell us that U.S. default is unlikely. Private investors are unwilling to lend at risk-free rates to individuals likely to default in the next few years; we should believe the same about lending to sovereigns. I simply do not find a sudden-default scenario likely.
Demographics provide another reason why default and inflation are unlikely solutions: Because of its high birthrate, the U.S. is likely to be one of the last rich countries to reach its debt peak. That means that the U.S. will likely watch the default-or-not decision played out time and time again in other prosperous countries. Defaulting lowers a nation’s status in the world pecking order, and after Americans watch a nation or two default on their obligations and become an object of ridicule, they will prefer alternatives to default.
But as the baby boomers age and debt approaches its projected peak, won’t voters rationally support default? If voters thought that their decisions were decisive—if they believed they were each the political scientist’s “pivotal voter”—they might well do so. Once the money has been borrowed to fund and care for the baby boomers, it could well be rational to repudiate the debt, even if it meant no borrowing for decades.
But since individual voters cannot change election outcomes, they will not carefully weigh benefits and costs of default. Instead, they will place massive weight on symbolism and status-group affiliation—they will allow their feelings to abuse the facts of the matter. It is quite unlikely that defaulting nations will be considered high-status, so voters will be reluctant to support politicians who support default. Politicians who support default will likely find themselves turned out of office, a fact that foresighted politicians will keep in mind. Tax increases and spending cuts are acrimonious, but not shameful. Default is shameful.
Finally, crude public choice suggests that default is unlikely: U.S. Treasuries make up well over 10% of U.S. bank assets, a figure that spiked upward in the wake of the financial crisis. And the political influence of the banking sector on U.S. economic policy is beyond doubt, as the bailouts, credit lines, and ring-fences of 2008 and 2009 illustrated. The twin votes on TARP provide further evidence: Ramirez (2011) shows that politicians who switched their votes from “No” to “Yes” on the 2008 bank bailout bill were quite likely to be those who received more campaign donations from the banking industry. Such a politically connected sector would do everything possible to ensure that their investments in U.S. Treasuries pay off.
What Happens in a No-Default World?
If the U.S. avoids both hard and soft default, what is the likely set of outcomes? The most likely is some combination of massive tax increases and massive cuts in projected government health care spending. While it is conceivable that a once-and-for-all deficit deal would last for decades—as the 1983 Social Security reforms did—more likely there will be a series of temporary packages. There will certainly be some “permanent” changes to entitlement programs, but these will always be open for review by future Congresses. This will be true especially on the health-care side, where new innovations or the lack thereof and the relative political power of providers and recipients will drive the politics of health care for decades to come.
The size of the U.S. fiscal gap is well known. The Congressional Budget Office (2011), the IMF (2011), and Auerbach and Gale (2009) have all created estimates that indicate that the federal government must close a annualized long-run fiscal gap that is possibly larger than the size of the Department of Defense.
The two leading proposals for long-term fiscal consolidation—the Simpson/Bowles plan and the Ryan plan—give a sense of what is needed to close the gap. Both rely on cuts to projected health care spending as their biggest source of savings. The Ryan plan entirely avoids tax increases by converting Medicare into a lump-sum insurance premium support payment that grows at the rate of inflation—and since nominal health care costs are expected to rise faster than inflation, the Ryan plan is an ever-growing cut to real expenditures on Medicare coverage.
Both plans are valuable for their candor in illustrating that the U.S. long-term fiscal situation is far from healthy. They also implicitly illustrate how hard it will be politically to achieve balance in a no-default scenario. Since soft or hard default are both unlikely, I anticipate something within the range of the Simpson/Bowles and Ryan plans to be enacted, perhaps in pieces over many years. But cutting health care spending to the elderly, who vote at high rates, will be politically costly, so politicians will look for alternatives. And in a no-default world, the most likely alternative will be tax increases. This brings us to the greatest barrier to solvency: the Republican Party.
A No-New-Taxes Populace: The Bondholder’s Enemy
The strongest argument against my no-default position is that U.S. political Right is exceptionally unwilling to raise taxes: As long as the GOP holds enough sway in at least one branch of government to block a tax increase, the GOP will be a genuine political force. All countries have strong political forces fighting against spending cuts: What makes the United States unique among developed countries is its powerful anti-tax movement. Bondholders will legitimately be worried that the anti-tax movement will become a barrier to repayment, a cause of default.
The most likely form of default is an ostensibly “accidental default”: Politicians bargain for weeks on end and fail to meet a key deadline for rolling over the debt or for raising the debt ceiling. In this environment, markets might decide against rolling over loans to an untrustworthy debtor, forcing a default of at least a few days, a default that would scar the U.S. financial landscape, making a full-blown default likely with a few months.
We saw a training-run version of this in 2011: Some on the anti-tax-increase side argued that the U.S. government could stop all “non-essential” spending or sell off assets to avoid breaching the U.S. debt ceiling. Unusual options became live options. And once some unusual options become live, bondholders reasonably wonder whether the unusual option of default will become the newest live option.
In the long run, bondholders get paid from the gap between government revenues and government spending. In most rich countries, the revenue lever and the spending lever are both tools that politicians can manipulate in order to leave enough to repay the lenders.
In the United States, by contrast, as long as a no-new-taxes GOP holds an effective veto over at least one branch of government, the revenue lever is unresponsive. Bondholders prefer a pliant populace, and they don’t have that in the United States. This is why it’s reasonable (not necessary, but reasonable) for Treasury bondholders to believe that anti-tax, anti-spending movements are bad news from their point of view. Anti-spending movements are quite often cheap talk; anti-tax movements, less so. That may be good news for net taxpayers, but it is bad news for lenders.
“Starve the Beast”: A Barrier to Solvency and Smaller Government?
Does rigorous evidence support the claim that tax fighters increase deficits? Does it support the related claim that deficits tend to increase government spending? Buchanan and Wagner make the second claim:
Debt financing reduces the perceived price of publicly provided goods and services. In response, citizen-taxpayers increase their demands for such goods and services. Preferred budget levels will be higher, and these preferences will be sensed by politicians and translated into political outcomes. (Buchanan and Wagner 1977, 8.9.35)
Gale and Orszag have made a related claim:
The “starve the beast” strategy may simply not work as a political equilibrium. We have in mind that policymakers jointly go through periods of fiscal restraint and fiscal largesse, and the restraint or largesse occurs simultaneously on both the tax and spending sides. (Gale and Orszag 2004, 999)
The latter authors refer to the periods of restraint as “coordinated fiscal discipline” (1000). Politically, periods of diet and exercise alternate with periods of sloth and gluttony. And from a politician’s perspective, a tax cut may feel quite gluttonous.
These positions stand in contrast to the “starve the beast” position made famous by Milton Friedman (2003), George Will (1978) and other limited-government thinkers (see citations in New 2009). In this view, the politically acceptable deficit is close to exogenous, fixed by some combination of political tastes and financial markets, and thus a tax cut causes a spending cut. Under Buchanan and Wagner-style fiscal illusion, by contrast, a tax cut lowers voters’ perceived relative price of government and increases the demand for government spending.
For the federal government, which view is closer to the evidence? Considering the importance of the question, it is surprising that so little empirical work exists to answer it: The empirical question has been left mostly to op-eds and anecdotes. But some data-driven work exists for the United States, and it tentatively favors the fiscal illusion position. Niskanen (2006) uses time-series evidence to show that controlling for the vagaries of the business cycle, times of lower taxes are times of higher, not lower, government spending. Romer and Romer (2009) use a narrative approach to search for exogenous changes in tax policy, finding no evidence that tax cuts starve the beast and weak evidence that tax cuts predict higher spending.
These studies find no evidence that tax restraint curtails government spending at the federal level in the United States. And it is not as though such evidence cannot be found, or cannot be published in academic journals: At the U.S. state level, “starve the beast” is widely studied under the rubric of statutory and constitutional tax limitations, finding modest-to-negligible evidence in its favor (Bails 2006; Mitchell 2010). States and localities are less able to borrow cheaply and readily; this may help explain the modest evidence for the beast-starving below the federal level.
A final form of evidence for federal fiscal illusion comes from survey data: Ura and Socker (2011) find that survey respondents are more supportive of higher domestic spending after a rise in budget deficits, just as Buchanan and Wagner would predict. Their work does not look at the change in demand for government after a change in tax policy, so it is an imperfect test of “starve the beast,” but with so little data on such an important question, every piece of evidence is welcome. At this point, there is no rigorous evidence that tax-limitation efforts at the federal level have curtailed the size of government in past decades, and modest evidence these efforts have increased both the demand for and the supply of government.
A fortiori, this means that federal tax limitation efforts have had even larger effects on the federal debt: a one dollar tax cut would need to cut spending by exactly one dollar to have no effect on the debt, and the evidence to date suggests that a one dollar tax cut, if anything, increases spending. But even if future evidence finds that a dollar of tax cuts causes a 25-cent decline in spending, this is bad news for bondholders. For “starve the beast” to earn even a C among bondholders, it needs an A+ among supporters of smaller government.
Democrats as the Party of Bondholders
The financial services industry donates to both parties, and politicians of both parties have “retired” to lucrative jobs at Treasury-holding investment banks, so raw political interest suggests some incentive for both parties to please their supporters. But which of the two parties will actually vote the way these supporters would like?
Political scientists have tracked how Republican and Democratic members of Congress differ. Polsby and Schickler (2002, 175) summarize a classic finding of Mayhew (1966) as follows: “Democrats…always vote to sustain their constituent interests; Republicans tend to vote on principled grounds even against their constituents and allies.” In recent decades, party unity has massively increased in both political parties (Sides 2011), but the GOP’s reputation as a barrier to revenues is salient. The no-tax-increase principle dominant in today’s GOP is likely on the minds of Treasury bondholders around the world. Bondholders’ best hope is that the GOP will lose power from time to time, opening brief windows for pro-solvency tax increases.
With the GOP reluctant to raise taxes to meet bond obligations, and with both parties reluctant to cut health care spending for the elderly, bondholders will pay particular attention to years when the GOP—especially the Congressional GOP—is out of power. These will be the times bondholders can hope that government will use its tax lever. If tax increases are not forthcoming in years of Democratic ascendancy, this will bode ill for bondholders. Ironically, crises may be most likely in years when the tax-hiking party is in power, because Democratic inaction will send a stronger signal to bondholders than Republican inaction.
Default is unappealing to voters, unlikely to rescue our fiscal position, and unforetold in market prices. Hence, I predict that both soft and hard defaults are extremely unlikely in the United States.
The caveats to that position lie in the no-new-taxes position of the modern