What People Think Backs U.S. Government Debt

In 1790, Alexander Hamilton argued that public credit depends on opinion as well as reality. The distinction remains central today. Treasury prices and yields show the demand absorbing government debt, but not what investors believe gives that debt value. The same high debt value, low yield, and stable rollover can reflect expected future budget surpluses, demand for a safe asset, confidence that the government can keep refinancing, or Federal Reserve support.

Those explanations are not interchangeable. They can support the same price in normal times but imply different resilience when fiscal conditions deteriorate. To judge sustainability, we need to know not only how much demand exists today, but which sources of support would remain under stress.

My research with Ricardo Delao measures these valuation frameworks directly. We first ask an open-ended question about what ultimately backs U.S. government debt. The answers produce five recurring sources; we add financial repression (policies and regulations that steer savings toward government debt) from the literature. We then ask 985 screened government-bond owners, 1,001 registered voters, and 247 people with graduate training in economics or finance to allocate perceived debt value across the six sources.

Global safe-asset demand is the largest perceived source in all three samples, receiving 26 to 30 points out of 100 on average. Future primary surpluses, the textbook fiscal anchor, receive only 14 to 16 points. Safe-asset demand and permanent rollover (the government's ability to issue new debt to repay old debt) together receive 43 to 46 points. That is close to half of perceived value and roughly three times the surplus share.

The averages conceal substantial disagreement. Respondents tend to concentrate weight on a dominant source; their largest individual allocation averages about 60 points. The sample means therefore combine distinct valuation frameworks rather than describe one representative view. Formal training changes the ranking very little: even the economics and finance group assigns only 16 points to future surpluses.

The result is not an artifact of presenting respondents with a menu. In a separate open-ended survey, the same broad channels appear in respondents' own words, and safe-asset demand remains prominent. The ranking also differs sharply from national newspaper coverage, where primary surpluses dominate valuation-relevant discussion while permanent rollover receives little attention.

The model explains why these beliefs matter. An investor who would stop lending once repayment is doubted and one who would continue lending can hold the same Treasury at the same normal-time price. Prices reveal the total demand absorbing Treasury supply, but not whether that demand comes from safety motives, persistent rollover, or investors who would withdraw under stress. The same observed demand can therefore imply different refinancing capacity in a crisis.

Public backing has a similar hidden composition. Future surpluses depend on future policy and cannot readily be transferred to creditors. Real government assets can be sold, while central-bank resources can be deployed without issuing another fiscal promise. Two economies with the same debt and normal-time prices can therefore have different resources available under stress. The survey disciplines these hidden compositions, but it does not structurally identify market depth or refinancing capacity.

The paper then turns from valuation to sustainability. Across all three samples, respondents put the average probability of a U.S. debt crisis within ten years near 50 percent. Their answers are broadly consistent with two separately elicited measures: the most likely crisis date and the maximum debt-to-GDP ratio they believe the government can sustain. The concern is therefore not confined to one survey question.

Yet concern rarely becomes decisive stated action. Among respondents expressing concern, 91 percent of voters say debt has not been decisive in their voting decisions, and 72 percent of investors report no concrete past portfolio change. Valuation frameworks also explain little of the cross-sectional variation in concern or action. Believing that debt rests on future surpluses does not make a respondent more likely to act.

A randomized information treatment sharpens this gap. Showing bond investors the current debt level and the Congressional Budget Office's long-run projection raises their stated ten-year crisis probability by 14.9 percentage points. The share planning to reduce Treasury or bond-fund holdings rises by an imprecisely estimated 4.2 points. Information can materially change perceived risk without producing a comparable change in intended exposure.

This does not show that people would remain passive in an actual crisis. It shows that beliefs about fiscal risk and beliefs about what makes Treasuries valuable are distinct, and that high perceived risk need not dominate choices under normal conditions. The evidence measures stated valuation frameworks, beliefs, and intentions, not realized prices, trades, votes, or the beliefs of the marginal investor setting the market price.