Interest Rate Differentials and the Swiss Franc
Punchline: It is often believed the spread between SNB and ECB policy rates is a key determinant of the CHF/euro exchange rate. I explore that claim and find surprisingly little support for it. The franc turns out to be harder to explain than many assume.
Some commentators argued that the SNB should have cut rates by more than 0.25% in June to widen the interest rate differential with the euro area and weaken the Swiss franc. A weaker currency, they contended, would help prevent inflation from falling too low.
This view rests on the belief that the spread between the SNB’s and the ECB’s policy rates is a key determinant of the EUR/CHF exchange rate. The SNB itself appears to believe that the differential matters. In what follows, I examine that claim. My conclusion is that the link between policy rate differentials and the EUR/CHF exchange rate is considerably weaker than often assumed.
Interest parity
According to the interest parity condition, the interest rate differential between Switzerland and the euro area should equal the expected change in the EUR/CHF exchange rate over the maturity of the relevant instruments, plus a risk premium. This implies that what should matter for the exchange rate is not the current level of policy rates, but the expected future path of the policy rate differential and risk premium.
Since expectations of future policy rates influence long bond yields, one might expect the EUR/CHF exchange rate to be more closely related to the spread between long-term bond yields than to the spread between short-term policy rates. While policy rates do affect bond yields, the relationship is indirect and, as the data below suggest, often weak.
Interest rates and the exchange rate
I now turn to the data. The graph below shows the EUR/CHF exchange rate together with the SNB’s and the ECB’s policy rates. The SNB’s rate has averaged about 0.5 percentage points less than the ECB’s, while the franc has appreciated by roughly 2% per year on average.
This co-movement in policy rates reflects the close economic ties between Switzerland and the euro area. Because of the high degree of trade integration, macroeconomic shocks in the euro area tend to spill over to Switzerland, prompting the SNB and ECB to adjust rates in the same direction and at roughly the same time.
Source: SNB and ECB
Because the spread between the SNB’s and the ECB’s policy rates is often cited as a key driver of the exchange rate, I have repeatedly tried to estimate econometric exchange rate models in which it plays a central role. These efforts have consistently failed to uncover a significant relationship. One reason may be that the two rates tend to move closely together, leaving little variation to identify any effect. If anyone is aware of studies that find a strong and robust link, please feel free to share them in the comments.
Of course, the failure to find such an effect does not mean that monetary policy is irrelevant. Rather, it suggests that the spread between the SNB’s and ECB’s policy rates may not be the best metric for capturing the stance of Swiss monetary policy.
Bond yields and interest rate expectations
Given the importance of interest rate expectations for exchange rates, I next turn to long bond yields. The graph below shows 10-year yields in Switzerland and the overall euro area alongside the EUR/CHF exchange rate. Swiss bond yields averaged about 1.6 percentage points less than euro area yields. That matches quite well with the 2% appreciation per year of the Swiss franc.
As was the case with the policy rates, the two bond yields have moved closely together. Comparing the graphs of policy rates and bond yields, it is clear that the link between them is weak.
Source: SNB and FRED
In the case of the euro area, there is no single bond yield. One can use the average for the euro area as a whole, or the yields of individual countries such as Germany and Italy—commonly seen as low- and high-credit-risk benchmarks, respectively. This means a choice must be made when analysing yield differentials.
The graph below shows 10-year yields for the euro area, Germany, and Italy. Risk spreads widened sharply during the Global Financial Crisis in 2008 and again during the fiscal tensions that began around 2011, with German yields falling below and Italian yields rising above the euro area average.
These episodes coincided with upward pressure on the Swiss franc, which appears to have been driven more by safe haven flows and shifts in perceived euro area risk than by the interest rate differential between Switzerland and the euro area.
Source: FRED
Some estimates
To explore this further, I estimated a set of exchange rate models incorporating long bond yields.1 Four results came through consistently across all specifications.
First, euro area risk spreads mattered: when spreads between, for example, Italian and German bond yields widened, the Swiss franc tended to appreciate. This is consistent with capital flowing into Swiss assets during episodes of elevated perceived risk.
Second, the spread between the SNB’s and the ECB’s policy rates was never significant.
Third, the slope of the Swiss yield curve measured as the spread between long-term bond yields and the SNB’s policy rate was significant. When the SNB tightens policy and the yield curve flattens, the franc tends to strengthen.
Fourth, neither the ECB’s policy rate nor the slope of the euro area term structure had much explanatory power. Since it is hard to argue that only Swiss monetary policy matters for the exchange rate, this is a sign that the models I estimated are missing something.
Summing up
While my estimates show that tighter Swiss monetary policy tends to strengthen the franc, the spread between the SNB’s and the ECB’s policy rates is not a reliable measure of the relative tightness of monetary policy in the economies. While the slope of the Swiss yield curve appears to capture the tightness of Swiss monetary policy, how to capture the tightness of ECB’s monetary policy remains unclear. Perhaps some measure of QE needs to be included.
More broadly, I am left to conclude that the determinants of the Swiss franc exchange rate are more complex, and less well understood, than is often claimed. The idea that the SNB can fine-tune inflation by adjusting the exchange rate through small changes in its policy rate appears optimistic. I end by noting that these estimates are preliminary. I may return to this topic in a future post.
The views expressed are my own. The work presented is preliminary and may contain errors. It should not be construed as investment advice. Readers are encouraged to seek professional investment guidance.
I used data starting in either 1999 or 2010 and estimated the model in first differences, with the change in the exchange rate as the dependent variable. The specification included a constant, the lagged change in the exchange rate, and dummy variables for October 2008 and December 2015. I also included the VIX, oil prices, the USD/EUR exchange rate (to capture global financial conditions), US long bond yields, and relative CPI levels and the exchange rate to account for convergence toward purchasing power parity (PPP). The lagged change and the dummy variables were significant and retained, though they did not affect the significance of the main variables of interest. The USD/EUR exchange rate and the PPP-related variables were also significant but had no impact on the key coefficients and were therefore excluded. The remaining variables were insignificant and dropped from the analysis.
Thank you Stefan - have you considered what a similar analysis would show when adjusted for inflation rates, i.e. using real rather than nominal rates?