Guest Contribution: ” Does monetary policy respond to temperature shocks?”

Today, we are pleased to present a guest contribution written by Filippo Natoli of the Directorate General for Economics, Statistics and Research of the Bank of Italy. The views presented in this note represent those of the author and not necessarily reflect those of the Bank of Italy.

Central banks are debating on how best to make their contribution to the global fight against climate change. I offer insights on how US monetary policy responded to unfavorable economic shocks coming from temperature fluctuations during the last 50 years.

 Climatologists and economists all agree: climate change is a threat for future economic performance. The literature has indeed shown that, in recent decades, rising and volatile temperatures had significantly affected countries’ GDP. Although there are significant differences across economies as described in this Econbrowser post, advanced countries like the United States are no exception.

This poses challenges for policymakers. Though monetary authorities are not as central as governments in fighting climate change, there is a lively discussion on what central banks can do using their policy toolkit for mitigating its economic effects. However, the extent to which the conduct of monetary policy might itself be affected by the ongoing climatic trends is still unclear. This lack of evidence rests on the fact that the joint effects of temperature fluctuations on economic output and consumer prices, as well as the correct monetary policy response, are an open issue. Extreme temperatures can have supply-side effects – e.g., by decreasing labor productivity – and demand-side ones – e.g., via rising energy expenses for households and firms – so it is not clear a priori whether and how monetary policy should respond to temperature shocks.

I take up this issue by quantifying the impact of temperature oscillations on the US economy.  I study how consumption, investment, and ultimately GDP are affected, how the CPI index responds and, in turn, how these effects propagate to short and long-term interest rates on government bonds.

For this purpose, I propose a new way to identify unpredictable changes in temperatures that fit with the notion of a shock in macroeconomics. Using average daily temperatures for each US county since the 1970s, I compute quarterly county-level “surprises” as the difference between the number of very high and low temperature days within the quarter and those observed on average during the same quarter of the past five years. The underlying idea is that agents learn about the distribution of temperatures and, based on what they experienced in the recent past, form their beliefs on the highest and lowest temperatures to expect in current season, with these beliefs being updated every year. County-level surprises are aggregated to obtain a US-wide “temperature shock”. By focusing on the size of the shock and by identifying exogenous variations with respect to the most recent temperature data, my approach stresses the idea that, by surprising agents, exceptionally hot and cold weather is what matters in the short run. It therefore overcomes flaws of other methods proposed in the literature based on positive vs. negative temperature variations – which can potentially mix good and bad economic shocks – and fixed-effects panel estimates – where temperature variations with respect to long-run averages can be predictable, as climate change continuously increases the incidence of temperature extremes over time.

Figure 1 displays the incidence of surprises by county (panel a) and the evolution over time of the US-wide temperature shocks (panel b). The first picture reveals that, between 1975 and 2019, surprises have been largest in southern counties; the second one shows that, at the national level, adjustments in the shape of the temperature distribution have been greatest – inducing bigger shocks – in the early part of the sample than in recent times. This does not imply that temperature fluctuations have reduced in size, but simply that extreme temperatures have become more the normality in recent times, so they are somewhat less surprising than in the past.

Figure 1 – County-level temperature surprises and US-wide temperature shocks

Source: Natoli, F. “Temperature surprise shocks”, MPRA Working paper n. 112568, March 2022. [Latest version here]

I then use the constructed US-wide shocks to study the response of key economic variables using local projections. Panel (a) of Figure 2 shows the result for GDP and the consumer price index: while the negative impact on economic activity is significant, that on the CPI index is more muted, suggesting that demand- and supply-side effects may balance out on average. The shock induces a significant reaction by the Federal Reserve, displayed in Panel (b): in line with the response of GDP, some months after the shock the Fed’s economic nowcast (produced within the set of Greenbook Forecasts) is revised down due to the shock. This induces an expansionary monetary policy reaction, as short rates fall at the same horizon. While the behavior of short rates is not by itself guarantee that the Federal Reserve correctly identified the source of the downturn, some evidence points to an increase in the Fed’s attention to temperature fluctuations right after the shock. Indeed, the occurrence of temperature-related wording in the transcripts of each FOMC meeting slightly increases after adverse temperature shocks (last picture of panel b).

Figure 2 – Response to adverse temperature shocks

Source: Natoli, F. “Temperature surprise shocks”, MPRA Working paper n. 112568, March 2022. [Latest version here]

All in all, these findings suggest that climate-related shocks have concurred to determine the conduct of US monetary policy during the last 50 years, adding another piece of evidence to the debate on the role central banks can play to counteract the economic effects of climate change.

This post written by Filippo Natoli.

You may also like...