This photo from Feb. 26, 2013 shows dry cracked land near a water reservoir in Kiwitahi, New Zealand. (AP Photo/New Zealand Herald, Christine Cornege, source : phys.org)

During the past few weeks, with my co-author Gauthier Vermandel, we have redrafted our working paper (available on HAL) untitled “Weather Shocks” that looks at the role of the weather in the generation of business cycles. In this blog post Gauthier and I provide a summary of our work.

Instead of looking at what happens in the long-run, which is what is mainly done in the macroeconomic literature related to climate change, we focus in our paper on what happens in the short-run. We look at how weather shocks can generate business cycles and induce welfare costs.

Why such a paper?

The measure of weather-driven business cycles is of crucial interest in a context of climate change. The climate projections point to a drastic increase in both the frequency and the variance of weather disasters, yet there is no macroeconomic framework that can gauge their potential effects on economic activity.

Our methodology, in a nutshell

Since there is little empirical evidence on how the economy responds to a weather shock, we propose a theoretical approach strongly rooted in empirical observations. We estimate a Real Business Cycle model to analyze the short-term effects of weather shocks on macroeconomic fluctuations. This model also makes it possible to study the long-term effects of climate change

More specifically, we first employ a Vector Autoregression Model (VAR) to document the transmission mechanism of a weather shock – as depicted by a drought – using quarterly data from New Zealand.

Based on the evidence provided by this empirical model, we construct a theoretical one and estimate with the same data as in the VAR, using Bayesian techniques.
In a nutshell, the model characterizes a small-open economy composed of two sectors, including an agricultural sector, and a non-agricultural one. The former is sensitive to drought-related weather shocks, which affect the productivity of agricultural land.

Once estimated, the model can be employed to assess the short and long terms implications of the weather.

  • We use it to tackle the propagation of weather shocks and assess their contribution to the business cycles of an economy.
  • The framework is also amenable for gauging the quantitative potential of climate change in terms of welfare and macroeconomic volatility.

Our findings

  • First our model shows that a weather shock reduces land productivity damaging the agricultural sector. The weather leaks positively to the non-agricultural through an increased demand from farmers for non-agricultural goods offsetting the drought, combined with a depreciation of the real exchange rate. This spillover across sectors partially dampen the recessionary effects of the weather, but not sufficiently to avoid the recession.

The key mechanism of the model, illustrated:

A more quantitative illustration:

  • In addition, weather shocks are also proving to be an important sources of business cycles in New Zealand over the sample period. This volatility has welfare implications. A comparison of our model with another in which weather fluctuations are absent provides a way to quantify these impacts. The results indicate that households would be willing to give up 0.19% of their unconditional consumption to live in a world free of weather shocks. This 0.19% measure of well-being is not to be underestimated. By contrast, the cost of a standard productivity shock is only 0.05%
  • Under most of the climate scenarios tested, the variance of climate shocks increases (see the Figure below) and generates increased volatility in macroeconomic series. As a result, the cost in terms of welfare increases considerably. It is multiplied by 1.5 in the worst case scenario, relative to the historical situation.

Projections of welfare cost depending on the variance of the weather shock

In the paper, we estimate changes in the variance of drought shocks under four climate scenarios. One of the characteristics of these scenarios is the date on which the peak of CO2 emissions is reached. After that date, CO2 emissions decrease.

Let us assume the very simplifying assumption that the year in which the CO2 peak is reached is the major determinant of the variability of the drought shock. Under this assumption, it is possible to estimate, for different years, the corresponding value of the variation in the variance of the weather shock. To this end, we use the carbon peak dates associated with each of the four scenarios on which we rely, as well as the estimated variances of the weather shock, to interpolate the variances of this shock at different dates, ranging from 2020 to 2100.

For each date, we thus have an estimated value of the variance of the drought shock in New Zealand. It is then possible to compare the evolution of the variance of the weather shock with that observed historically, over the period 1989-2014. It is also possible to compare, in a similar way, the change in the variance of some macro variables. For simplicity, we normalize the historical variance of each variable to 100. We can compare the welfare measure under historical conditions with that under modified weather conditions. All these comparisons are shown in the Figure below. By default, it displays the estimates for a carbon emission peak in 2020. For such a date, our estimates give a decrease in the variance of the weather shock of 8% (the y-axis value of the climate shock indicates 91.97). This leads to a decrease in the variance of GDP by 2.5%. The decrease in the variance of all macroeconomic variables results in an increase in well-being, which increases by 0.175% compared to the historical value.

By dragging the slider to the right, it is possible to display the results of our simulations for different peak carbon emission dates (unfortunately, the “play” button does not work properly).

Weather shocks are therefore a major source of business cycles that entails a non-trivial welfare cost. Most of climate projections for the end of this century predict a rise in the variance of weather shocks in New Zealand that would result in our model in a drastic increase of the welfare cost, accompanied by an increased macroeconomic volatility.

The contribution of our research to the literature

  • The economic literature on the effects of the weather on the economy mainly takes a long-term perspective, exemplified by the Integrated Assessment Models. Our paper contributes to this literature connecting the weather to the economy by providing a short-term perspective on the role of weather shocks as a generator of economic cycles.
  • A first preliminary contribution of the paper remains in the measurement of the weather at a macro level. We collect soil moisture deficits data from weather stations in New Zealand and then engineer a quarterly weather index that provides an accurate measure of land productivity for agricultural activities.
  • A second contribution we make lies in the evidence provided by the VAR model on the interaction between the weather index and other macroeconomic time series. In particular, an adverse weather shock induces an important contraction of economic activity, mainly driven by the agricultural sector, combined with a depreciation of the domestic currency. More interestingly, the effects of a weather shock persist longer than the weather shock itself, thus highlighting an unusual persistence mechanism that we refer to as the weather hysteresis effect.
  • A third contribution concerns the enrichment of an otherwise standard RBC model with a weather dependent sector. More formally, we incorporate an endogenous land productivity subject to exogenous changes in the weather that captures the weather hysteresis effect.

Today, weather shocks have significant consequences for the economy, especially for countries that are heavily dependent on their agricultural sector. Within a few decades, as indicated by climate projections, an increase in the variability of climate shocks is expected. These shocks negatively impact the economies of the countries affected by them, in a persistent manner. It is important to document how our systems respond to such shocks. A better understanding of the processes at work can help to develop mechanisms in future research to minimize the harmful effects of weather shocks. Our article aims to carry out this work of analyzing and documenting the effects of weather shocks on the economy.

Reference

Ewen Gallic, Gauthier Vermandel. Weather Shocks. 2019. ⟨halshs-02127846⟩

Leave a Reply

Your email address will not be published. Required fields are marked *

Time limit is exhausted. Please reload CAPTCHA.