Between Sept. 2011 and Jan. 2015, the Swiss National Bank intervened in FX markets with the explicit goal of keeping the EURCHF exchange rate above 1.20. This «lower bound» policy had two effects: First, it kept the exchange rate at an artificially inflated level, and second, it reduced its volatility significantly. It was clear from the outset that this policy would be only of a temporary nature, but it was unclear when and why it would be discontinued.
If, as turned out to be the case in Jan. 2015, the reason for discontinuing the policy would be excessive costs and risks because of continued downward pressure, the exchange rate was expected to drop significantly once this policy change was announced, and at the same time, exchange rate volatility was expected to increase. This created the following questions, which I analyzed (together with my colleagues Rolf Poulsen from Copenhagen and Alex Weissensteiner from Bolzano) in two research articles:
(i) Where would the EURCHF exchange rate have been between Sept. 2011 and Jan. 2015 in the absence of the SNB’s «lower bound» policy?
(ii) What were the expectations of market participants regarding the remaining lifetime of the SNB’s policy?
(iii) What were the expectations of market participants regarding the exchange rate volatility after a discontinuation of the SNB’s lower bound policy?
Knowing the answers to these questions was interesting, e.g., for companies considering hedging against changes in the EURCHF rate: If markets expected the SNB to continue its policy for an extended period of time, and both the potential drop in the exchange rate and the increase in volatility following a discontinuation of the policy were estimated to be small, then hedging might be less interesting compared to the expectations of an imminent end of the policy together with a sharp
drop in the exchange rate and a marked increase in exchange rate volatility. Answers to these questions can be found by extracting information from market prices, which aggregate market participants’ expectations.
In this sense, prices contain consensus estimates about future developments, which can be extracted from the prices using suitable models and methods. For the questions described above, we followed two approaches, both of which are based on option pricing theory. The common feature of both approaches is to model the observed EURCHF exchange rate in the presence of the SNB’s policy as the sum of the (hypothetical) exchange rate in the absence of this policy and a put option on this rate with an unknown lifetime. Both approaches estimate the (hypothetical, or latent) exchange rate that would have been observed in the absence of the SNB’s lower bound policy, together with its volatility and the expected remaining lifetime of the policy. The first approach (Journal of Futures Markets 2015, 35(12), 1103-1116) uses exchange rate data together with prices of traded FX options to estimate these quantities.
The second approach (Quantitative Finance 2019, 19(1), 1-11) does not require FX options data and is based solely on the observed exchange rate. The forecasting accuracy of both approaches can be judged using the actual developments after the SNB discontinued its lower bound policy in Jan. 2015. Both approaches work well, with the second approach providing forecasts that were extremely close to the values actually observed. This example shows that the information (the « wisdom of the crowd ») contained in market prices can indeed be extracted using suitable quantitative models.
Blog entry written by Prof. Michael Hanke