The Swiss National Bank (SNB) recently announced (PDF) that it would be pursuing a price floor on the Swiss franc of EUR/CHF 1.20. Moreover, reports indicate that this intervention will be unsterilized, which suggests a significant commitment to altering the actual franc supply dynamics, rather than simply relying on signaling effects from sterilized intervention (which both research (subscription required) and experience suggest are generally ineffectual beyond the short run). And to top it all off, the SNB heavily sold forward volatility options,[*] effectively seeking to dampen market expectations around the fluctuations in the euro-Swissie rate.
This action came on the back of almost a decade-long strengthening of the franc against the dollar, a two-and-a-half year strengthening against the euro, and---since mid-2010---a sharp increase in both the rate of Swissie appreciation coupled with an increase in the volatility of the currency (see figure). The appreciation has also meant that the SNB has accumulated an unprecedented amount of international reserves, swelling its holdings of foreign exchange on its balance sheet even before the announcement.
How credible is the Swiss commitment to maintaining the 1.20 exchange rate? One argument is that it is entirely credible. After all, unlike the case where a central bank is defending a (potentially) overvalued currency, the logic of a standard currency crisis model does not actually imply a steady erosion of foreign exchange reserves and the inevitable crash; after all, a central bank has the ability to print unlimited amounts of its domestic currency. Consequently, the process of converting foreign currency to the franc at the 1.20 floor can, in theory, go on forever.
Well... maybe. While long-run inflation (and hyperinflation) may be always and everywhere a monetary phenomenon, more modest inflation, especially in the short run, is affected by a host of other factors, including demand slackness, inflationary expectations, and the actual size of the monetary base. For example, were foreign holders of francs not spend their newly-acquired francs on actual goods and services in Switzerland (but instead hoard them in Swiss bank accounts), inflation expectations may well remain in check, especially when coupled with relatively stable domestic demand.
Does this then mean that the Swiss commitment is entirely credible? Unfortunately, no. So long as the SNB cares a little bit about both inflation and output, it will face the circumstance in the future where, were market participants to react positively to its 1.20 commitment, it would have an incentive to renege on its earlier exchange rate target. The logic is entirely analogous to the standard argument of time-inconsistent central banking.[†]
So, what do the travails of a small, rich economy have to do with developing countries? One only needs to look around the world to see that the currencies of emerging economies are now being increasingly viewed as safe havens akin to the Swiss franc (see figure). The Banco Centro do Brasil has been battling real appreciation ever since the (while simultaneously struggling with the one against inflation); many Asian currencies are seeing highs not experienced since the Asian financial crisis; and there is even market talk about shifting portfolios toward fairly thinly-traded Eastern European currencies.
Source: World Bank staff calculations, from BIS effective exchange rate indices.
This leaves developing countries that are currently experiencing a surge in capital inflows (and concomitantly strong exchange rates) in the uncomfortable position of needing to navigate the flood with some compromise of the impossible trinity: relinquish monetary policy independence with a hard peg, impose (leaky) capital controls, or accept currency strength but introduce some limited policy offset via (tighter) fiscal policy and (stricter) regulation.
†. More formally, let the loss function of the central bank at time t be given by a fairly standard Barro-Gordon type Lt = (πt - π*t)2 + a (yt - y*t)2, where π is inflation, y is output, asterisks indicate targets of each respective variable (so that the first term is the deviation of inflation from the central bank's target, while the second is the output gap), and a is the weight on the output gap. Using the purchasing power parity relation et pft = pt (where e is the spot exchange rate, and p and pf are home and foreign price levels, respectively. Using the definition of inflation, it is possible to rewrite the above in terms of the exchange rate target: Lt = (πt - E π*f)2 + a (yt - y*t)2, where E is the target. If foreign inflation is given (as it would be in a small open economy), then the expression E π*f effectively contains no control variables. For any central bank that cares about inflation, there will exist a rate of inflation where the optimal solution under discretion and following a rule differ; this implies that the fixed rate E cannot be credible.