(p.221) Annex 4 : Three Generations of Crisis Models
(p.221) Annex 4 : Three Generations of Crisis Models
The three generations of crisis model all have to do with the collapse of a fixed exchange rate.1 What differs across the models are the range of policy response options open to the government, how these are affected by actions taken by investors in the asset markets, and what happens to private sector balance sheets in the real and financial sectors. Fiscal considerations, actual or contingent, play a critical role in all three cases.
In 1979, Paul Krugman published a seminal paper in the Journal of Money, Credit and Banking on balance of payments crises. His so-called first generation model featured a government financing its fiscal deficit by credit from the central bank in the presence of a fixed exchange rate. The rate of inflation is equal to the rate of depreciation of the currency and is therefore zero initially. With the demand for money (which is a function of inflation) fixed to begin with, the credit created to finance the fiscal deficit spills over into current account deficits, depleting the central bank’s foreign exchange reserves. At some point, domestic residents begin to suspect the authorities will run out of foreign currency, forcing the government to abandon the fixed peg and float the currency—the “crisis.” This leads to a jump in the equilibrium inflation rate, which, post-crisis, is simply the rate of depreciation needed for the government’s fiscal deficit to be financed by the inflation tax.
The beauty of Krugman’s model is that the switch to a float happens before reserves actually reach zero. It happens when remaining reserves are exactly equal to the reduction in the demand for money (the home currency) that will occur at the point the float is adopted; the demand for money falls because at that point, the rate of inflation jumps from zero to the level needed to finance the government’s fiscal deficit. Domestic residents simply buy up or “attack” the remaining reserves at the pre-crisis fixed exchange rate, forcing a float of the currency. Therefore, no one suffers a capital loss, since, in this model, domestic residents can calculate exactly when the switch to a float will take place and the exchange rate itself does not jump at that point. The inputs for this calculation include the pre-attack rate of reserve depletion (determined by the size of the fiscal deficit and related credit creation), the function linking money demand to inflation, and the post-float inflation rate needed to finance the deficit by the inflation tax.2
(p.222) Second Generation
In Krugman’s model, the speculative attack on reserves is spurred by inconsistent fundamentals: a fiscal deficit financed by credit is not consistent with a fixed exchange rate (zero inflation). But in a 1994 paper heralding the second generation crisis model, Maurice Obstfeld argued that Krugman’s model could not fully explain the Exchange Rate Mechanism (ERM) crisis of 1992–3 and the experience of Sweden in particular; Box A4.1 contains a brief description of this crisis. Industrial European countries could always augment reserves by borrowing, and “fiscal profligacy” in the shape of a fiscal deficit financed by credit from the central bank may or may not be part of the picture. But other factors come into play, such as the impact of high interest rates and politically unacceptable levels of unemployment, which might create a conflict between the interest-rate defense of a currency peg and bringing unemployment down.
Obstfeld’s central idea is that the government has policy options to deal with actual or perceived unsustainability in macroeconomic policies. Its eventual response is driven by a political calculus in which it balances the pain of holding on to a fixed peg (for example, high interest rates leading to pressures on private sector balance sheets or the exacerbation of high unemployment) with the gain of letting it go (the pain will be avoided without a big effect on the government’s credibility either because other important countries have acted similarly or the move will be seen as “sensible”). The ambiguity about how the government will eventually respond gives rise to “multiple equilibria.” Market expectations, in turn, are molded by perceptions of which policy option the government will choose and the anticipated impact on asset prices. This would then affect positions taken by investors which could force the hand of the government. This circular logic could lead to a crisis as a self-fulfilling prophecy of shifting market sentiment even as fiscal fundamentals remain unchanged. Such models are often referred to as second generation models.
The third generation crisis model was motivated by the East Asian experience of 1997–8. This highlighted balance sheet mismatches on currencies and maturities (short-term dollar debt financing long-term local currency assets), international illiquidity (insufficient reserves), and moral hazard (the private sector makes bad investment decisions but is confident the government will bail it out). As a result, even if the government is not currently running large fiscal deficits, prospective deficits may be high on account of contingent liabilities related to the fiscal costs of bailing out banks and private firms. The possible collapse of the fixed exchange rate is heightened by international illiquidity, manifested in a high ratio of short-term external debt to foreign exchange reserves and/or a high ratio of broad money to reserves. This configuration makes a country susceptible to shifts in market sentiment regarding a devaluation or a reversal of capital flows. If interest rates are raised to make domestic currency assets more attractive, this would put a strain on banks’ balance sheets by increasing nonperforming loans; if the exchange rate is allowed to collapse, the real burden of dollar-denominated debt rises on corporate balance sheets, forcing them into bankruptcy, which hurts the banks as well. This makes it impossible to defend the peg against speculative attacks, forcing its abandonment as well as a costly bailout, as a wholesale bankruptcy of the corporate and financial sectors would be politically unacceptable.3 (p.223)
All three model generations are marked by fixed exchange rates, an open capital account, and low foreign exchange reserves. The second generation brings in market psychology and a conflict between external (maintaining exchange parity) and internal (e.g., keeping unemployment at politically acceptable levels) goals. Not knowing how the government will react brings in the possibility of multiple equilibria: the new equilibrium depends upon which goal the government chooses to pursue. For emerging markets, there is not going to be much of a choice: where (p.224) there is a conflict between external and internal imbalance, the external constraint will typically dominate the internal one. This dominance holds particularly when the balance sheet problems associated with the third generation model are present. I now discuss this.
External versus Internal Balance
Suppose an emerging market has an external financing constraint, which, in extreme form, would be manifested in a sudden stop in capital inflows. The standard way to address this would be to let the real exchange rate depreciate. Although a real depreciation should help relieve the external constraint via both the income effect and price effect (by reducing aggregate demand and shifting it toward domestic goods), it could lead to a costly recession when currency mismatches are present: with assets denominated in local currency and liabilities in foreign currency, the depreciation could lead to an increase in nonperforming loans and a bank credit squeeze. Alternatively, consider an emerging market which lowers its interest rate to deal with an unemployment problem. If the capital account is open, this could lead to an exodus from the domestic currency, forcing a depreciation which could create havoc in the presence of currency mismatches. Textbook prescriptions thus fail and may even backfire when the capital account is open and currency mismatches are present. Hence, the dominance of the external over internal balance is more likely to prevail when the capital account is open and firms and banks have currency mismatches on their balance sheets, as in the third generation model.4
Summing up, the “three generations” have two critical implications for emerging markets: first, it is not enough to look simply at fiscal and growth fundamentals; one also needs to pay attention to the market’s signals on devaluation and default risk. Second, the third generation model has important implications for the government intertemporal budget constraint owing to contingent fiscal liabilities from potential bailouts of the private sector. Its message is clear: watch out for private external debt and currency mismatches.
(1) . The exchange rate does not have to be literally fixed. Managed pegs, crawling pegs, and even inflation-targeting regimes (see Kumhof, Li, and Yan 2007) all fall into the category of regimes vulnerable to speculative attack and collapse.
(3) . Various aspects of the third generation model were inspired or developed by Dooley (2000), Krugman (1999), Chang and Velasco (2000), and Burnside, Eichenbaum, and Rebelo (2001). See also Claessens (2005), Frankel and Wei (2005), and Diaz-Alejandro’s seminal paper (1985).