SECO’s flash estimate of GDP for 2024Q2 and a Swiss Phillips curve
2024-13
With the flash estimate for real GDP in 2024Q2 published on 15 August, here I update the estimate of the output gap discussed in my earlier post. I also fit a simple Phillips curve for Switzerland.
According to SECO’s press release, real GDP rose 0.5% in Q2. The revised output gap is small and virtually identical to the earlier estimates that ended in 2024Q1. Given the trend growth rate of the economy, the output gap for 2024Q2 is trivially smaller (in absolute value) than the output gap in Q1. All-in-all, the flash estimate suggests that the state of the Swiss business cycle has not changed much.
Source: my estimates using data from SECO.
Using the model for the output gap, I forecast the level of GDP until the end of 2025. The table below shows the forecasted growth rate over four quarters and over one quarter, and the predicted output gap. These forecasts are uncertain and subject to large standard errors so caution should be used in judging them.
Since the output gap in Q1 was small, growth is forecasted to be similar to trend and the output gap is expected to close slowly in the coming quarters.
Source: my estimates using data from SECO.
I go on to estimate a Swiss Phillips curve. Since preliminary estimates showed that the output gap affected the domestic inflation rate more strongly that headline inflation (which also comprises imported goods whose prices presumably are largely determined in international markets), I focus on domestic inflation here.
With inflation subject to a strong seasonal factor, the dependent variable is the four-quarter change in domestic prices, I regress it on two of its own lags and the output gap, using data for the period 2000-2024Q2 (but drop the data point for 2023Q1 which is an outlier). The results are (R-sq = 0.92):
Source: my estimates using data from SECO.
I end with three suggestions for further work, which I plan to pursue in the future.
Since the output gap is estimated, the regression results above are subject to generated-regressors bias. Estimating the output gap jointly with the Phillips curve seems desirable.
It would be useful to incorporate additional variables such as sentiment indicators in the estimation of the output gap.
The Phillips curve is rudimentary and could be expanded by considering additional variables that may impact on the cost structure of the domestic sector, such as energy cost.
Interesting analysis. What measure of domestic inflation do you use?
I suppose another thing to consider from the academic standpoint is that when using y/y data for inflation (rather than q/q SA) is that one can get complicated MA structures in the errors. Wondering if these could result in biases in a model with lagged dependent variables.