William Davidson Institute Working Paper 402
assumption that such an institution would almost eliminate the risk of currency
attacks. In the next section we take a critical look to this issue.
4. Speculative attacks and currency boards
4.1 First generation models: arguments pro currency boards
A well-known explanation to currency crises under a fixed exchange rate
regime was worked out by Salant and Henderson (1978), Krugman (1979) and Flood
and Garber (1984). The law of one price governs the price level in the small country,
while the price level of the foreign country is given.8 The uncovered interest parity
also holds. These “first-generation models” postulate an exogenous increase in the
central bank domestic credit (Flood and Marion, 1999). The subsequent increase in
money supply puts downward pressure on interest rates and entails a capital flight. In
order to defend the parity, the central bank must sell reserves such as the total amount
of high-powered money is kept constant. At some moment in time, its reserves will be
totally depleted and the fixed rate regime must be abandoned. Holders of domestic
currency realise that the policy is unsustainable and force the exit from the fixed
exchange rate arrangement before the moment when the reserves are completely
9
depleted.
In the case of a currency board, the monetary base has no other counterpart
than the reference foreign currency. Thus, such a monetary institution cannot create
money by increasing credit to banks or to the government. In the line of the former
analysis, many experts suggest that this institutional arrangement could completely
rule out the risk of a currency attack. This certitude should be twice qualified.
4.2 Currency boards, the Phillips curve and output cycles
It may be argued that in the short and medium-run, prices in transitional
economies are driven by the wage costs given the stance of the labour market. Indeed,
in these economies, the productive sector is dominated by large firms, which often
hold a monopolistic position. Such firms may form prices simply by adding a profit
margin to unit costs. Then, for constant productivity, there is a straight relationship
between wage and price inflation. At some moment in time, for a given level of
unemployment, prices and wages would chase each other upward. Inflation is
therefore positive if unemployment falls below its structural value.10 This alternative
logic may explain a loss in competitiveness that does not build on the assumption of
exogenous increase in central bank domestic credit.
Starting from an initial situation with positive inflation, at some moment in
time, domestic prices will exceed the world prices and net exports will become
negative. As a consequence, aggregate demand diminishes and investment
profitability deteriorates. Private capital inflows decline or are reversed. The reserves
of the currency board begin to fall; simultaneously, the money stock is reduced. The
8 Some authors part with the law of one price and write an equilibrium condition in the money market
that establishes a positive relationship between the money stock and the domestic price level.
9 The optimal timing requires no discrete jump in the price level (i.e. an infinite inflation rate).
Therefore, the currency attack should occur when the actual fixed rate and the implicit exchange rate
that would clear the market under a flexible scheme (the “shadow” exchange rate) are equal (Flood and
Garber, 1984).
10 This is quite a traditional explanation, which can be traced back to the paper by Samuelson and
Solow (1960). See for a modern reference Layard, Nickell and Jackman (1991).