U.S. Government Default: Good, Bad, or Just Plain Ugly? Comment on Cowen and Kling
Back in 1993 I gave a lecture at Golden Gate University,"The National Debt: Good, Bad, or Just Plain Ugly?" Toward the end of the lecture (which I have presented many times since), I discussed the problem created by the unfunded liabilities of Social Security and Medicare. Given that politicians have little incentive to fix the problem, I predicted an eventual U.S. government default on Treasury securities, possibly within my lifetime. Even some of my most rabid libertarian listeners thought at the time I was going off the deep end. But over the intervening fifteen years, several commentators have started to take seriously the prospect of a risk premium on Treasuries, including David Romer of Berkeley, the WALL STREET JOURNAL, and Moody's Investors Services. In fact, economist Laurence J. Kotlikoff himself predicts such a default in a recent issue of the ECONOMISTS' VOICE, edited by Joseph Stiglitz.
Now two of the most prominent libertarian economists on the blogosphere have started to discuss government default as well, possibly arising from other causes. Tyler Cowen wrote:"When it comes to the mortgage agencies, there is no real choice but to bail out the debt holders. The alternative is a run on the dollar and collapse of faith in U.S. government securities and the end of the world." This inspired Arnold Kling to consider the potentiality of such a government default in two successive posts: here and here.
There seems to be a certain tension between Tyler's argument for bailouts and some of his other opinions. He wrote in a subsequent post that although"I have very much favored the 'bailouts' to date [,] I don't favor that they were necessary." The statement is difficult to interpret. If it means anything more than that Tyler prefers utopia if possible, it suggests that he wishes the U.S. had a macroeconomic regime in which the alternative to bailouts was not"the end of the world." Yet Tyler has also dismissed free banking and/or a gold standard as providing no significant benefits over the current system of central bank-managed fiat money. But isn't fiat money managed by a central bank a major feature of what, in Tyler's opinion, links the future of the mortgage agencies to the future of the U.S. dollar?
More important, would a U.S. government default indeed be"the end of the world"? Tyler's scenario contains many implicit assumptions that require examination. The first is that saving Freddie Mac and Fannie Mae decreases the probability of U.S. default. As Mark Brady pointed out to me, one could plausibly argue just the opposite. In fact, a firm refusal to bail out the mortgage agencies would establish a strong barrier between U.S. Treasuries and the fortunes of not only the mortgage agencies themselves but also the myriad other institutions that we can imagine receiving similar treatment. Wouldn't that in fact help maintain confidence in U.S. government securities?
A second assumption in Tyler's scenario is that what happens to U.S. Treasuries and what happens to the dollar are inextricably intertwined. It is true that fiat money makes it harder than would a gold standard to separate the fate of a government's money from that of its debt; but it is certainly not impossible to do so. Treasury securities are second-order claims to central bank-issued dollars. Although both may be ultimately backed by the government's power of taxation, that in no way prevents government from discriminating between the priority of the claims. After the American Revolution the U.S. repudiated its paper money and yet successfully honored its debt (in gold). What in theory prevents it from doing the reverse in the future?
That the U.S. in fact would do the reverse is implied by some of the observations (with which I generally agree) that Tyler has made about the recent record of the Fed and other central banks. Seigniorage has become a trivial source of revenue in developed countries with widespread fractional reserve banking. Yet Tyler seems to blithely assume that a major fiscal crisis would lead the U.S. government to resort to inflationary finance. If the double-digit inflation experienced by the U.S. during the 1970s covered no more than 2 percent of central government expenditures, imagine the hyperinflation that would be necessary to cover even a two-thirds increase in total federal expenditures, a conservative projection of what current Social Security and Medicare benefits would require over the next half century. Central bank independence, moreover, permits the Fed to isolate the dollar from such a fiscal crisis. And given a choice between only default on its debt, and both default on its debt and collapse of its currency, I expect the Treasury and the Congress would gratefully acquiesce.
But my major disagreement with Tyler and Arnold is that I believe that a U.S. government default, rather than being "the end of the world," could possibly be a good thing. I even advocated repudiating the national debt in a 1981 issue of CALIBER (the newsletter of the California Libertarian Party), long before predicting a default. My arguments were moral, economic, and political, and I would only soften them slightly today.
The moral argument for repudiation is easiest to follow although by itself says nothing about the practical results. Treasury securities represent a stream of future tax revenues, and investors have no more just claim to those returns than to any investment in a criminal enterprise. I favor total repudiation of all government debt for the same reason I favor abolition of slavery without compensation to slaveholders.
The economic argument depends on whether Ricardian Equivalence holds. Repudiating government debt eliminates future tax liabilities. To the extent that people correctly anticipate those liabilities, the value of private assets (including human capital) should rise over the long run by the same amount that the value of government securities falls. Thus, people will gain or lose depending how closely their wealth is associated with the State. If on the other hand, people underestimate their future tax liabilities, they suffer from a fiscal or "bond illusion" in which Treasury securities make them feel wealthier than they actually are. Debt repudiation will bring their expectations into closer alignment with reality, which should increase saving.
When I survey students in my classes, most of them claim to have no realistic expectation that social security will be there for them when they retire. If they are being honest, and they act on this belief as they earn income, then Ricardian Equivalence should hold for those liabilities, making any default less painful. These of course are all long-run effects, and the current market hysteria about the subprime crisis leaves me far less sanguine than I once was about the short-run financial turmoil following a Treasury default.
The political consequences are the trickiest to analyze. A government default is certainly a balanced-budget amendment with real teeth. Moreover, government defaults in the past, when not obviated by bailouts from other governments, seem to have had positive political consequences. Compare the widespread defaults of American state governments in the 1830s, with their cascading benefits--reluctance of states to set up government-owned railroads the way they had government-owned canals, balanced-budget constitutional amendments at the state level which even today impose lingering constraints, a general state retrenchment in a period of increasing laissez faire, among others--with the baleful consequences of the failure to repudiate the Revolutionary War debt, the most notorious of which was replacement of the Articles of the Confederation with the U.S. Constitution. During the late 1830s, President Martin Van Buren blocked any national bailout of the states, and yet the world did not end, and indeed the U.S. continued on the path of sustained growth that it had only recently started down. Unfortunately nowadays, with the U.S. and assorted international agencies stepping in to stave off sovereign defaults, we don't have good recent comparisons.
The greatest potential political benefit of a future government default would be the end of the democratic welfare state. In fact, nearly all the social democracies seem to be approaching similar fiscal crises at the same time. Even though the U.S. started its pay-as-you-go social insurance later than most of the rest, it has caught up with them because, rather than rationing medical services with national health care, it subsidizes medical services. Nor is it likely that the U.S. government will be able to turn to any combination of explicit taxes and siegniorage to stave off the crisis. I've already explained the futility of relying exclusively on seigniorage. Federal taxes as a percent of GDP have been roughly constant since 1960, bumping up against 20 percent. That is an astonishing statistic when you think about it. There have been numerous tax changes over what is nearly half a century, with rates sometimes rising and sometimes falling, and still the total bite out of the economy has not changed much. The chances that Americans would put up with a doubling of that bite strikes me as close to nil. Look at the problems Governor Arnold Schwarzenegger faces with only a minor tax increase in the face of California's fiscal straits.
So when the default occurs, I expect it to happen very fast, much like the sudden collapse of the Soviet Union. While the Social Security and Hospital Insurance trust funds only provide the verisimilitude of full funding, that is enough to create a seeming firewall between the Treasury debt and social insurance obligations. All that is required for investors to change their expectations is some signal that the firewall is an illusion, such as Congress dipping into general revenues to finance Social Security or Hospital Insurance. (The other parts of Medicare are already financed partly out of general revenues.) Suddenly investors will realize that the enormous $70 trillion unfunded liability (Kotlikoff's estimate) does affect the Treasury's ability to honor its debt. Once a default risk premium appears on Treasuries, with the need to constantly roll over of the debt, things will be over quickly. The Fed won't have the time, much less the ability, to inflate the Treasury out of its problems, even if it wanted to.
The social-democratic welfare State will come to end, just as socialism came to an end. Socialism was doomed by the calculation problem identified by Mises and Hayek. Mises also argued that the mixed economy was unstable, and the dynamics of intervention would inevitably drive it towards socialism or laissez faire. But in this case, he was mistaken; a century of experience has taught us that the client-oriented, power-broker State is the gravity well toward which public choice drives both command and market economies. What will ultimately kill the welfare State is that its centerpiece, government-provided social insurance, is simultaneously above political reproach and beyond economic salvation. Fully funded systems could have survived, but politicians had little incentive to enact fully funded systems, and much less incentive to impose the huge costs of converting from pay-as-you-go.
Whether the inevitable collapse of social democracies will be a good or bad thing depends on what replaces them. The travails of post-Communist countries can be a source of either optimism or pessimism, depending on where you look. But the lesson libertarians should take away from all this is to stop wasting their time on social insurance reforms that have no chance of implementation and small chance of working, even if implemented. Most of them are half-way measures that reduce only a portion of the unfunded liability with tax increases or benefit decreases, and try to cover the remainder with either smoke-and-mirror cuts in other government programs or hand-waving about higher growth rates. Libertarians instead should be helping people to accept the inevitable by explaining why government intervention dooms social insurance and why only voluntary alternatives are viable over the long run. This is the best way to ensure that what comes after the democratic welfare State is something better.
Juro Jordan - 11/29/2008
When US Treasury borrows money, doesn't it have the same effect on economy as when it prints money. And by paying back the debt should not have the countereffect, just as lowering and raising interest rate. It looks like Treasury only backroll every debt without ever paying it off. That tells me that Treasury borrows money from Fed which is a chair board of 12 private central banks. Does Fed borrows money from other countries or print it on its own? This is fishy.
By printing money Treasury would be free from paying interest rate and even from paying back the principal, unless for fighting deflation. Why do we need Fed?
Daniel B Klein - 9/7/2008
Jeff has been saying this for years, and always the Hummel scenario seemed plausible and possible.
Less Antman - 8/22/2008
Polls support your student survey, indicating a solid majority of people under 40 believe they will never collect social security benefits. Interestingly, though, people under 40 still oppose cutting benefits to current retirees or reducing their own social security tax payments.
All this suggests that young people ARE anticipating the long-term failure of these programs, but want to help those who don't have the time to prepare for the collapse (current recipients). I'd say that justifies the optimistic scenario about what will happen when social insurance fails. It also suggests that an immediate abolition of the program would cause a massive voluntary charitable response that would minimize serious hardship.
I agree that it is unproductive to try and come up with reform proposals. The best we can do is to tell the truth about the future of these coercive social insurance programs, warn people about the risks of owning Treasury securities, and then let the market handle the response.
Of course, somebody has to hold the Treasuries at the time of default: I am reminded of Yogi Berra's comment about a restaurant that "nobody goes there anymore because it's too crowded."
Are we really that far off before a risk premium is added to long-term Treasury bonds?
Chris Westley - 8/21/2008
Jeffrey - You should submit this to LRC.