Fiscal Consequences for Mexico of Adopting the Dollar

“Fiat government debt – debt that promises to pay only government-issued paper – is much more closely analogous to equity issued by private firms than to debt issued by private firms. Indexed government debt, or government debt denominated in foreign currency, is analogous to privately issued debt” (Abstract, p. 1). Just as a firm’s equity value can absorb bad news about future earnings through a lower share price with no default, a government’s fiat liabilities can absorb bad fiscal news through a change in the price level with no formal default; dollar-denominated debt, by contrast, promises a fixed real payment and so behaves like ordinary corporate debt, which must be serviced in full or explicitly defaulted on. Sims frames dollarization as, in this light, a decision “involving many of the same considerations that arise in corporate finance when a firm decides between equity and debt finance” (p. 1) – converting the government’s entire balance sheet from a mix of “equity” and “debt” to pure “debt.”

With both fiat debt B_t and dollar debt F_t outstanding, the real value of fiat debt equals the discounted present value of future primary surpluses net of the real cost of servicing dollar debt: “the current real value of fiat debt is a discounted present value of future primary surpluses, net of the real value of foreign debt service” (Section II, eq. 2, p. 3). Sims derives a simplified special case (constant real interest rate, surplus, dollar value, and foreign debt): since B_t cannot jump discontinuously, “increases in taxes tau or domestic real rate rho-bar are deflationary; increases in spending G, foreign interest rate R*, or the real value of the dollar e/P are inflationary” (Section II, eq. 4, p. 4). He is explicit this simplified equation is a useful benchmark “with roughly the same status as the Mv = PT equation of the quantity theory of money,” not a literal empirical model (p. 4).

Because a government with fiat liabilities can respond to a sudden fiscal shock (war, disaster, a bailout) either by raising taxes, by borrowing while credibly promising future tax increases, or, crucially, by simply not changing taxes at all and letting the price level rise to reequilibrate the valuation equation – an option that “is possible only because of the existence of fiat government liabilities. If there were only dollar liabilities, this option would not exist” (Section II, p. 4). Sims is careful to note the value of this option does not depend on believing governments should or would inflate away debt routinely – debt holders who anticipate steady inflation demand compensating higher nominal rates – but rather on the option’s use as an occasional, unanticipated shock absorber; he states plainly that his argument does not dispute the concern that short-lived governments may be “tempted… into a high-inflation equilibrium,” only that this risk must be weighed against “definite losses from giving up the option of using the price level as a fiscal shock absorber” (Section II, p. 4).

Sims shows the first-order conditions reduce, in ordinary fiscal states, to the simple result tau_t = tau_(t-1) – a stronger conclusion than Barro’s own tau_t = E_t[tau_t+1] – meaning the optimal policy under time-consistent commitment keeps tax rates essentially constant, using unanticipated deflation in good states (to preserve the real value of debt) and unanticipated inflation in bad states (to erode it) so long as the fiscal shock stays within an ordinary range, with outright debt repudiation reserved for the single worst possible fiscal state (Section III, pp. 5-7). Under a Markov, persistent spending process, there is a threshold level of spending above which existing debt is repudiated; below it, taxes stay fixed at their prevailing level and the price level absorbs the shock. Sims contrasts this with the all-dollar-debt case, which collapses to Barro’s original random-walk tax rate – a policy that “puts no limits on the range of variation in b[debt],” is “clearly worse,” and would eventually violate realistic bounds on tax rates (such as 100%), whereas the fiat-debt-enabled optimal policy “keeps b[debt], in a fixed range so long as the range of G is fixed” (Section III, pp. 7-8).

No – Sims argues the theory predicts precisely the opposite of what dollarization’s supporters might infer: real outstanding public debt is theoretically lowest right after a period of major fiscal stress (having just been partly inflated away or repudiated), and it is exactly those episodes and the fiscal reforms they induce that lay the foundation for later debt expansion and stabilized public credit (Section III, p. 9). “Fiscal stress and low debt are therefore not in themselves arguments for dollarizing all public finances.” He illustrates the broader mechanism with England’s debt expansion during the wars of the late 17th and early 18th centuries, after which public debt became markedly more secure – roughly consistent, he argues, with a country building a larger “cushion” of debt as a shock absorber over time, given sufficient policy credibility (Section III, p. 9).

Because the interest-rate gap observed before dollarization reflects a Modigliani-Miller-style allocation of fiscal risk across the government’s whole portfolio, not a property of dollar debt in isolation: “so long as there remains a nontrivial amount of fiat debt outstanding, uncertainties about the stream of future primary surpluses are absorbed by the fiat debt, allowing the dollar debt to be insulated from these uncertainties and maintain a lower expected return. But as soon as dollar debt becomes the entire stock of debt, its return must reflect the full range of uncertainty about future primary surpluses” (Section IV, pp. 9-10). Just as a firm cannot lower its overall cost of capital merely by shifting its liability structure from equity toward risk-free-seeming debt – doing so raises the risk premium on the remaining equity in exact offset – Sims shows “the real value of a country’s outstanding liabilities… is invariant to changes in the composition of those liabilities… and consequently the overall interest burden of the debt is unaffected by such shifts in composition” (Section IV, p. 10).

No – Sims argues that in practice a dollarized government retains an effective substitute for default: “suspending dollar payment” on its debts during a crisis, exactly as governments historically “suspended specie payment” under the gold standard, creating a discounted “Mexican debt dollar” that trades at a non-unit exchange rate with the actual dollar (Section IV, pp. 10-11). He further argues the political barriers to this kind of quasi-default may actually be lower after dollarization than the barriers to inflating away fiat debt, since ordinary inflation imposes costs on every domestic cash user and is correspondingly unpopular, whereas suspending payments to dollar-bondholders – especially if many are foreigners – imposes “no direct consequences for anyone except holders of the debt,” and the short-term political cost of doing so “may be small or even negative” (Section IV, p. 11).

Sims computes annual unanticipated real returns on U.S. government debt from 1950-1989 and finds them “not large relative to the Federal deficit – on the order of $40 billion as the maximum annual capital loss in the 1970s” – but “not negligible either,” with a pattern of negative unanticipated returns concentrated in 1973-1980 (consistent with an optimizing government offsetting the fiscal shock of the oil crises) followed by positive unanticipated returns in the 1980s as inflation was brought under control (Section V, pp. 11-13). He is careful to flag the limits of this evidence – some of the unanticipated variation could reflect private-sector real-rate shocks or term-premium effects rather than deliberate fiscal-risk absorption – but notes the negative correlation of unanticipated returns with both interest-rate and price-level innovations is consistent with capital losses on the (by then roughly 7-year average maturity) long-term debt dominating the pattern, rather than simple short-rate surprises (Section V, p. 13).

Because a government that retains the ability to issue fiat debt can always borrow to supply crisis liquidity so long as its discounted expected future surpluses are positive, whereas a government restricted to dollar-only borrowing cannot expand its borrowing without first raising the present value of its primary surpluses – so “the likelihood of widespread runs may rise with dollarization, rather than decline. In effect, speculation against the prospect of some form of Mexican government bankruptcy would be a source of much the same kind of instability that we see with fixed exchange rate regimes” (Section VI, pp. 14-15). He is skeptical of the two most commonly proposed fixes – a US-backed crisis line of credit (“unwise” given the US’s own inconsistent record on international financial commitments) and a large pre-funded dollar reserve, which he notes is costly and, per the paper’s own optimal-policy logic, generally not itself optimal fiscal policy (Section VI, pp. 15-16).