High temperature shocks create monetary policy trade-offs. This column uses a new monthly data set to study how 14 European central banks have responded to abnormally high temperatures in a centennial perspective (1920-2019). High temperature shocks are negative supply shocks to which central banks have reacted by lowering their interest rates. The macroeconomic impact of individual high temperature shocks on inflation has diminished over time, but the size and the frequency of such shocks has increased.
Climate change is the challenge of our time. Its consequences have been discernible since the 1980s and are becoming more pronounced, including rising numbers of extreme weather events. Among them, periods of abnormally high temperatures loom large. Not only are average temperatures increasing, but so are the size and the frequency of heatwaves (IPCC 2023).
What are the macroeconomic effects of such high temperature shocks and what are the consequences for macroeconomic policy? Central bankers in particular have expressed concern over whether the increased frequency and amplitude of climate shocks create new trade-offs for monetary policy (Cœuré 2018). Yet the demand for research into climate change related trade-offs is met only partially by academic supply (Natoli 2022). Economics research typically deals with long-term impacts on economic growth and standards of living (e.g. Bilal and Känzig 2024, Aerts et al. 2024).
The literature on the impact of temperature shocks has identified two channels through which they affect the economy (see Baleyte et al. 2024 for a review of this extensive literature). The first effect is felt in the agricultural sector: weather shocks affect crop yields, harvests, and agricultural production. The second channel operates through labour productivity, which might decline during heatwaves. However, no study in this literature has systematically examined the short-term macroeconomic impact of these shocks.
This is where our research comes in (Baleyte et al. 2024). We investigate short-term fluctuations of output and inflation induced by high-temperature shocks in the summer months. While periods of abnormally high temperatures have become more frequent of late, high temperature shocks of smaller size and lower frequency have always existed as part of random weather patterns. We can analyse such anomalies in a conventional monetary policy framework: are they supply or demand shocks, and how have central banks responded to them?
Data and methodology
We cover 14 European countries from 1920 to 2019. Global coverage would be preferable but is impossible to achieve given the need for monthly data and fully established central banks (central banks originated in 19th century Europe and spread globally only after WWII). Monthly temperature data are taken from the Berkeley Earth project (Rohde and Hausfather 2020) and are aggregated at the national level. Monthly data on industrial production, consumer prices, and interest rates were collected for a companion project on historical central bank policies (Bazot et al. 2024). Industrial production proxies for output, in line with similar studies examining the short-term impact of monetary policy.
We focus on the summer months – June, July, August, and September – because our research assesses the potentially short-term destabilising effects of the highest temperature shocks, rather than the overall effect of global warming throughout the year (Faccia et al. 2021). Monthly temperatures are used to calculate average temperatures and temperature anomalies. A temperature anomaly is defined as deviation from the 1951-1980 average (global warming became discernible in the 1980s).
Table 1 Frequency of monthly temperate anomalies exceeding +1°C and +2°C for 14 European countries, four different periods
Table 1 confirms that the frequency of high temperature shocks has increased over time. Yet it also shows that such shocks were by no means uncommon before 1980. A frequency of one means that a temperature anomaly occurs on average once per year. Given our focus on the summer months, the frequencies range between zero and four. We require countries with sufficient exposure to temperature shocks. The West/Central and South European countries lend themselves to this purpose. For example, they are the only ones in which a summer anomaly above 1°C (‘moderate hot temperature shock’) has occurred at least twice per year since 2000, and where a summer anomaly above 2°C has occurred at least once per year over the same period. We choose the nine countries with the highest frequency of summer anomalies in history and recent times as baseline scenario (this excludes Britain and the Nordic countries).
Non-linear effects are important (Kotz et al. 2024). Minor deviations from the average temperature are unlikely to have noticeable effects. By contrast, large temperature shocks might force central banks into taking strong action. We take non-linearity into account by using state-dependent local projections with a smooth threshold, following Auerbach and Gorodnichenko (2012). Observations just below the threshold are incorporated but given lower weight. Conversely, larger anomalies are given greater weight. This method has the advantage of increasing the number of observations in the state-dependent estimate and making the results less sensitive to the choice of a fixed threshold.
High temperature shocks are supply shocks
Figure 1 estimates the effects of a moderate hot temperature shock for our baseline scenario of nine European countries from 1920-2019, where the size of the shock is normalised to +1°C. The effects on output and prices are typical of a supply shock: lower output growth and higher inflation. The year-on-year variation in industrial production declines by 0.9 percentage points and reverts to normal in nine months. For example, if the year-on-year output growth rate is 2% in normal times, it declines to 1.1% after the shock and returns to 2% after three quarters. There is no rebound effect involving positive growth; consequently, the temporary effect on output growth causes a permanent loss to the output level. Year-on-year inflation increases by 0.4 percentage points and the effect is more long-lasting.
Figure 1 Effect of a hot temperature shock (above 1°C threshold) on the central bank interest rate, the growth of production, and the inflation rate
Notes: Main sample (nine countries), 1920-2019.
The effects are typical of a supply shock, to which the central bank responds by lowering the interest rate by 40 basis points. Production is prioritised over inflation, presumably because the immediate impact is larger on the former. The interest rate reaction is also consistent with the more persistent impact on the inflation rate: the decrease in the interest rate returns production growth to normal after nine months while inflationary pressures remain.
How do our results compare with standard estimates of the impact of monetary policy shocks? Recent studies on monetary policy shocks in the US and the Euro area (Bauer and Swanson 2023) find an immediate negative response of monthly industrial production of 1.2 percentage points and consumer prices of 0.3 percentage points to a 100 basis points rise in the interest rate. Put differently, the effects of a +1°C temperature shock on output and prices are in the same ballpark as a 100 basis points rise in the official interest rate. This helps explain why central banks reacted to high-temperature shocks long before climate change increased the size and frequency of such shocks.
Evolution over time
We extend our analysis in different directions. They confirm our main results but also give a fuller picture of what climate change means for central banks. First, when we confine our sample to more recent periods (Figure 2), we find some evidence that the effects of moderate hot temperature shocks (1°C threshold) are lower today than in the past. This holds true in particular for inflation. The smaller impact might be explained by a decrease of the agricultural sector as share of total GDP and/or of food as part of total consumption. Alternatively, the reduced impact might reflect technological improvements such as air conditioning which make today’s workforce more resilient to high temperature shocks. Please note that results for 1950-2019 and 1990-2019 remain statistically significant. Crucially, we return to impact sizes for the full period if we re-run estimations for the more recent periods but for anomalies exceeding the 2° Cthreshold (Figure 3). Central banks might simply reserve their firepower for larger temperature shocks today.
Figure 2 Effect of a hot temperature shock (above 1°C threshold) on the central bank interest rate, the growth of production, and the inflation rate
Notes: Main sample (nine countries), three different periods.
Figure 3 Effect of a hot temperature shock (above 2°C threshold) on the central bank interest rate, the growth of production, and the inflation rate
Notes: Main sample (nine countries), four different periods.
We find similar results and caveats for Britain and the Nordic countries with their lower and less frequent temperature anomalies. If we include them for the entire period based on a 1°C threshold, some of the responses are no longer statistically significant. Yet if we apply the 2°C threshold, our main finding – that hot temperature shocks are negative supply shocks to which central banks have consistently reacted – comes back in all its clarity.
Conclusions
Based on monthly data for 14 European countries spanning a full century, we document that hot-temperature shocks are supply shocks: inflation rises and production falls. Central banks reacted to fluctuations in output caused by high-temperature shocks (as part of random weather patterns) long before the effects of climate change became discernible in the 1980s.
Yet the increased size and frequency of high-temperature shocks means that central banks will find themselves today more often in a situation where output is falling and inflation rising. If inflation is already above target for other reasons (as it has been in recent years), this trade-off will grow even stronger. Central banks are no longer able to side-step the issue of climate change.
Source : VOXeu
