Skip to content

Lesson M21.L03: Currency Crises and Speculative Attacks

Module: M21: Open Economy Macroeconomics Part II Level: intermediate Duration: 30 minutes Learning Objective: Distinguish first-generation and second-generation currency crisis models using the 1997 Asian financial crisis. Data as of: 2024 Provenance: BIS Working Papers on Currency Crises | IMF World Economic Outlook Historical Data

Explanation

A currency crisis occurs when a country's fixed exchange rate collapses under market pressure — either through forced devaluation or the exhaustion of foreign reserves. Two major theoretical frameworks explain how and why crises occur.

First-Generation Model (Krugman, 1979)

Paul Krugman's model explains crises as fundamental-driven. The mechanism:

  1. A government runs persistent fiscal deficits financed by money creation (seigniorage).
  2. Money growth causes domestic inflation and erodes the real exchange rate.
  3. Under a fixed exchange rate, the central bank must sell foreign reserves to defend the peg.
  4. Reserves deplete at a predictable rate: R₀ = total reserves; μ = reserve loss per period.
  5. Rational speculators attack before reserves hit zero — at the critical threshold T* when it becomes profitable to attack.

The key formula for the reserve depletion timeline is:

\[T^* = \frac{R_0 - R_{\min}}{\mu}\]

where R₀ = initial reserves, R_min = minimum viable reserve level, and μ = rate of reserve loss per period.

Asian Crisis (Thailand, 1997): The Thai baht was a classic first-generation crisis. Thailand ran large CA deficits (−8% GDP), maintained a dollar peg, and had a fragile banking sector with unhedged foreign borrowing. The Bank of Thailand's reserves depleted rapidly as it defended the peg. In July 1997, the baht collapsed, triggering contagion across Indonesia, South Korea, Malaysia, and the Philippines.

Second-Generation Model (Obstfeld, 1994)

Maurice Obstfeld's model explains crises as self-fulfilling. The key insight: a government faces a trade-off between defending a peg (at cost: high interest rates, recession) and abandoning it (at cost: loss of credibility). If markets believe devaluation is likely, they raise interest rates, making defence more costly — which validates the expectation of devaluation. Multiple equilibria exist:

Belief Government action Outcome
Markets expect peg holds Interest rates stay low → defence cheap Peg holds ✓
Markets expect devaluation Interest rates rise → defence costly → abandon peg Devaluation ✓

Both outcomes are self-consistent. A "sunspot" can shift the economy from the good to the bad equilibrium.

UK ERM Crisis (1992): George Soros bet against the pound. The UK had joined the European Exchange Rate Mechanism (ERM) at an overvalued rate. High German interest rates (post-reunification) forced the UK to raise rates too — into a recession. The Bank of England spent £3.3bn defending the peg before capitulating on 16 September 1992 ("Black Wednesday"). This fits second-generation: the peg could theoretically have held, but the political cost of high interest rates made defence untenable once markets coordinated on devaluation.

Australia: Australia floated the AUD in December 1983 and has maintained a free float ever since. With no fixed exchange rate to defend, Australia has no speculative attack risk in either framework. The RBA does not target a specific exchange rate level, and APRA's prudential regulation reduces the currency mismatches (unhedged foreign-currency borrowing) that amplified the Asian crisis.

Worked Example

Reserve depletion timeline (First-Generation model)

Given: - Thailand's usable reserves: R₀ = USD 30 billion - Minimum viable reserve level: R_min = USD 5 billion - Rate of reserve loss (CA deficit drain + forward market interventions): μ = USD 2 billion per month

Step 1: Calculate the reserve buffer above the minimum.

\[R_0 - R_{\min} = 30 - 5 = 25 \text{ billion USD}\]

Step 2: Calculate the critical threshold time T*.

\[T^* = \frac{R_0 - R_{\min}}{\mu} = \frac{25}{2} = 12.5 \text{ months}\]

Step 3: Interpret. Without a speculative attack, reserves would last 12.5 months. However, rational speculators attack before T*, anticipating reserve exhaustion. An attack is profitable when the expected devaluation gain exceeds the carry cost of borrowing domestic currency to short it.

Step 4: Speculative attack condition. An attack occurs when the "shadow exchange rate" (what the rate would be if the peg collapsed today) exceeds the fixed rate:

\[e_{\text{shadow}} > e_{\text{fixed}}\]

In Thailand's case, the shadow rate showed the baht was overvalued by approximately 15–20%, making attack profitable at any reasonable borrowing cost.

Step 5: The actual July 1997 attack occurred after Thailand exhausted reserves through off-balance-sheet forward market intervention — reserves were far lower than publicly reported, making the fundamentals worse than they appeared and T* effectively zero.

Common Misconception

Misconception: Currency crises are always caused by irresponsible government policy (excessive money printing or fiscal deficits).

Correction: Second-generation models show that crises can be entirely self-fulfilling, even when a government's fundamentals are sound. The UK in 1992 had no excessive money growth — its crisis arose from the political cost of defending an overvalued peg in a recessionary environment. Self-fulfilling crises mean that no amount of reserve accumulation guarantees protection if markets coordinate on the bad equilibrium. This is why moving to a floating exchange rate (as Australia did in 1983) eliminates the speculative attack mechanism entirely — there is no fixed rate to attack.

Practice Prompts

  1. Conceptual: Explain why rational speculators attack a currency before reserves reach zero, rather than waiting until the peg actually breaks. → Answer: A speculative attack is a profitable option: borrow (short) domestic currency at the fixed rate, then buy it back after devaluation at a lower rate. If speculators wait until reserves hit zero, the devaluation has already occurred and the profit opportunity is gone. By attacking once the shadow exchange rate exceeds the fixed rate, speculators convert a near-certain future profit into an immediate one. Early attack is a dominant strategy for rational speculators, which is why crises appear sudden even when reserve exhaustion was predicted months in advance.

  2. Numerical: A country has R₀ = $50bn, R_min = $8bn, and μ = $3bn per month. (a) Calculate T. (b) If reserves are doubled to R₀ = $100bn (R_min and μ unchanged), recalculate T. (c) What does this imply about reserve accumulation as a defence strategy? → Answer: (a) T = (50 − 8) / 3 = 42/3 = 14 months (b) T = (100 − 8) / 3 = 92/3 = 30.7 months (c) Doubling reserves more than doubles the timeline (14 → 31 months), buying more time for adjustment. However, it does not eliminate the attack if fundamentals remain weak (μ > 0). As long as the underlying fiscal deficit forces reserve losses, the attack will eventually come. Reserve accumulation buys time for adjustment but is not a permanent solution to weak fundamentals.

  3. Application: Compare the 1997 Thai baht crisis (first-generation) and the 1992 UK ERM crisis (second-generation). For each, identify (i) the primary mechanism, (ii) whether the crisis was avoidable, and (iii) the role of market expectations. → Answer: Thailand (first-gen): (i) Fundamental deterioration — CA deficits, reserve depletion, unhedged banking sector. (ii) Avoidable with earlier adjustment: earlier float, reduced CA deficit, or bank restructuring before crisis. (iii) Expectations confirmed fundamentals — once reserve decline was detected, attack was inevitable. UK (second-gen): (i) Self-fulfilling — political cost of high rates in recession made defence conditional on market beliefs. (ii) Potentially avoidable if German rates had fallen faster or if UK had entered ERM at a more competitive rate. (iii) Expectations were central — Soros's short position itself was the catalyst. Crisis happened not because fundamentals demanded it, but because coordinated selling made defence too costly. This illustrates that second-gen crises can strike countries with sound fundamentals if the peg is politically untenable.

Further Resources