VOLATILE CONSUMER SPENDING: New evidence of the key impact of precautionary saving against the risks of unemployment

The risk of losing your job becomes stronger or weaker over the course of the business cycle – and this has a big impact on consumer spending as people change their preferred level of ''precautionary savings''. That is the central conclusion of research by Edouard Challe and Xavier Ragot, published in the February 2016 issue of the Economic Journal.

Their study shows that in a recession, when unemployment risks are higher, precautionary savings rise significantly, which naturally leads to a considerable fall in aggregate consumption, thereby magnifying the economic downturn.

The researchers note that consumption expenditures represent about a third of aggregate demand, but the sources of its volatility and their impact on growth and employment are not well understood.

In particular, economists'' basic model of consumption behaviour, equipped with standard preferences, implies that individuals smooth consumption quite effectively in the face of income ''shocks''. But embedded in a macroeconomic model hit by plausible shocks, the basic consumption choice model generates a volatility of aggregate consumption expenditures that is at best half as large as actual historical volatility.

While several plausible alternatives have been proposed to achieve a better understanding of the volatility of aggregate consumption expenditures, the new study argues that an important contributor is time-varying precautionary savings against unemployment risk.

At each point in time in the analysis, individuals wish to hold a buffer stock of precautionary wealth in case they are hit by a bad income shock in the near future, and their preferred buffer stock varies over the business cycle as the risks of a shock change. The researchers focus on income shocks arising from unemployment, which generate a drop in individual income (from a full wage to unemployment benefits) of about 40% in the United States.

The desired amount of precautionary wealth is cyclical: it rises in a typical recession, when unemployment risk is high and hence the likelihood of experiencing the income drop associated with unemployment is large. The increase in actual precautionary wealth towards desired precautionary wealth is then achieved by cutting consumption.

Conversely, desired and therefore actual precautionary wealth falls in booms (that is, when unemployment risk is low), which frees individuals'' resources for more consumption. Because time-varying precautionary saving behaviour tends to take consumption down during recessions (when unemployment risk is high) and up during booms (when unemployment risk is low), it effectively magnifies the consumption fluctuation predicted by the basic consumption choice model.

So at least from a theoretical point of view, this channel may increase the predicted volatility of consumption and take it closer to its historical counterpart. But then the question is: by how much and is that enough to explain the historical volatility of aggregate consumption?

The quantitative analysis that the researchers use suggests that it does. Their model incorporates the two key features of unemployment risk and imperfect insurance against this risk (so that an unemployment shock actually materialises into a fall in individual income).

The model is fed with shocks to unemployment risk with size and persistence that are directly estimated from the data (the main indicator of unemployment risk is the quarterly job loss rate – that is, the probability of losing one''s job in the coming quarter). It turns out that given these shocks (which have plausible magnitude by construction), the analysis produces a volatility of aggregate consumption expenditures that is in line with the data.

''Precautionary Saving Over the Business Cycle'' by Edouard Challe and Xavier Ragot is published in the February 2016 issue of the Economic Journal. Edouard Challe is at the Ecole Polytechnique and CREST. Xavier Ragot is at the Paris School of Economics.