Fiscal Imbalances and the Dynamics of Currency Crises

A fiscal imbalance is a current or anticipated future decline, of size Delta, in the present value of the government’s real primary surpluses – defined net of whatever fiscal adjustment the government is able or willing to make – and it is this shock, arriving as news at some date t, that the model traces through to a currency crisis. The paper studies “the dynamics of a currency crisis associated with a fiscal imbalance, defined as current or anticipated future decline in real primary surpluses” (Introduction), explicitly allowing the decline in surpluses to begin either immediately or at some specified future date, so that “we allow for the possibility that fiscal deficits start to deteriorate at a future date” rather than only modeling an instantaneous fiscal shock.

The paper explicitly builds a bridge between Krugman-style (1979) first-generation currency-crisis models, which emphasize money-financed deficits driving a predictable peg collapse, and the fiscal theory of the price level, which treats fiscal and monetary policy symmetrically as joint determinants of the price level. The authors write that their “framework reveals important links between first-generation models of currency crises after Krugman (1979) and the fiscal theory of the price level (FTPL),” noting that the fiscal shock they study “is common in the currency crises literature,” and that their setup relates to the distinction between “active” fiscal policy coupled with “passive” monetary policy from Leeper (1991) and the “non-Ricardian” policy terminology of Woodford (1995, 1996) (Introduction).

With only short-term nominal liabilities, the government’s intertemporal budget constraint at the fixed exchange rate fails the instant the news of the fiscal imbalance arrives, so the exchange rate must jump immediately – there is no way to defend the peg even for a single additional period. As the paper puts it, “upon the arrival of the adverse fiscal news, the government cannot defend the exchange rate parity, not even for one period. If an equilibrium in our economy exists with Delta > 0, it must be the case that the exchange rate jumps up immediately” (Section 3.2); the size of that initial jump is shown to depend negatively on the equilibrium present value of seigniorage and on the outstanding stock of government liabilities, and positively on the size of the fiscal imbalance itself.

The authors show that, in their baseline short-term-debt model, the present value of seigniorage the government could raise by delaying a devaluation is “somewhat surprisingly … identically equal to zero” once the fiscal cost of defending the peg beforehand is netted out. They decompose total seigniorage into revenue collected before the collapse (which is actually negative, since defending the peg entails a costly pre-collapse contraction in money demand as agents anticipate the coming devaluation) and revenue collected after it, and find these two pieces “exactly equal in present value,” so that “an attempt by the government to increase the collection of seigniorage at and after [the collapse date] will be self-defeating”: a larger target for post-collapse seigniorage revenue only produces a larger equilibrium devaluation and a more severe speculative attack beforehand, not more net revenue (Section 3.1).

Long-term, non-indexed government bonds (modeled as perpetuities) create a second channel for transferring resources from the private sector to the government – an immediate fall in the bonds’ market price when bad fiscal news arrives – that does not depend on seigniorage at all, so a large enough stock of such bonds can let the peg survive even when net seigniorage is exactly zero. The paper explains that “an imbalance in the government budget that cannot be matched by seigniorage revenues requires an offsetting wealth transfer from the private to the public sector,” and that “with non-indexed bonds of long maturities, a wealth transfer can occur via an unanticipated jump in bond prices in period t” (Section 4) – because anticipated future inflation and devaluation lower the real value of the fixed nominal coupon the perpetuity promises, and that capital loss to bondholders is a resource gain to the government exactly analogous to the wealth transfer an immediate devaluation would produce.

Government solvency alone bounds the delay: the paper derives a finite upper bound, denoted the date at which a currency crisis must occur, beyond which the government could not remain solvent even by exploiting the maximum feasible fiscal gain from an eventual devaluation. This upper bound depends on the stock of long-term liabilities outstanding, the size of the fiscal imbalance, and the post-collapse money growth rate; the paper notes that a crisis occurring close to this outer bound “corresponds to an extreme scenario in which, while the exchange rate is pegged, the government borrows up to the point in which the maximum fiscal gain from inflation is just enough to guarantee solvency,” and that such a solvency-driven crisis is associated with especially extreme rates of devaluation and inflation (Section 4.2). Solvency alone, however, leaves the exact date of collapse within this window undetermined.

The authors assume the central bank defends the peg for as long as it can keep the domestic nominal interest rate below a fixed threshold – equivalent, via the money-demand equation, to defending a lower bound on the money supply – and show this pins down a unique collapse date by backward induction: the government abandons the peg in the first period in which keeping the exchange rate fixed would push the interest rate to or above that threshold. They are explicit that this interest-rate rule “is closely related to the rule in Krugman (1979),” which instead posits an exogenous rate of domestic credit expansion together with a lower bound on reserves; using perfect foresight, they show individual agents “have no incentive to attack the peg too early” (since an early attack contracts money demand too little to force a collapse) and cannot profitably wait “too long” either, since the interest rate is monotonically rising over time as the collapse date approaches, so the unique equilibrium timing is “the last period in which such an attack is resisted by the authorities” (Section 4.3).

The shadow exchange rate and shadow interest rate are the hypothetical exchange rate and interest rate that would prevail if the peg were abandoned in the current period, calculated at every date leading up to the actual collapse, following the concept introduced by Flood and Garber (1984). The paper’s Figure 3 traces both: when the adverse fiscal news arrives, the shadow exchange rate immediately begins depreciating and continues to do so until it coincides with the actual, pegged exchange rate exactly at the collapse date; correspondingly, the shadow interest rate jumps up on impact but stays below the central bank’s defense threshold until the period just before collapse, when it rises to meet the threshold – so that “the realized interest rate coincides with its shadow value only at” the final period before the peg breaks (Section 4.3).

As long as the government is solvent and has full access to international financial markets, it can borrow reserves as needed to fight speculative attacks up to its solvency limit, so “the size of gross reserves at the central bank is at no time a key indicator of the government’s ability to withstand a speculative attack” (Section 4.4). This changes only once credit constraints bind – for instance statutory limits on debt and deficits, or a coordination failure among the government’s own creditors – at which point the stock of reserves relative to short-term liabilities becomes decisive, since the government can no longer simply borrow its way past a run.

When many small holders of short-term government debt each decide independently whether to roll over their loans or convert them into foreign currency, a coordination failure can produce a self-fulfilling run: if every agent expects others to keep rolling over their debt, no one runs and the peg survives, but if every agent instead expects a run, each individual’s best response is to run too, forcing an immediate devaluation once debt sales exceed the central bank’s reserves. The paper models this “in the presence of a fiscal imbalance,” so that “a run on public debt can force a devaluation as soon as the stock of short-run nominal liabilities of the government is at least as large as the stock of reserves,” with the exact timing of such a run left genuinely indeterminate – “the run prevents the government from borrowing additional reserves for peg defense,” and while the fiscal imbalance still guarantees some collapse by the model’s solvency-driven upper bound, “the exact timing is subject to indeterminacy, and is therefore unpredictable” (Section 4.4).

The authors conclude that fiscal and monetary policy should be viewed symmetrically as joint determinants of the price level and exchange rate in the spirit of the FTPL, that the maturity structure of government liabilities is a first-order determinant of whether and how long a peg can be defended, and that – while their precise zero-net-seigniorage result is a feature of this specific model – the broader lesson that “the costs of peg defense will still constrain feasible seigniorage policies, increasing endogenously the revenue need of the government” is more general. They close by flagging “these costs and their equilibrium implications for policy and exchange rate stability as important topics for future research” (Conclusions), positioning the paper as a foundational step rather than a final word on how fiscal imbalances propagate into currency crises.