We study the role of precautionary savings in the recent rise in French household savings. Using different approaches, we show that elevated uncertainty contributed to the surge in savings in 2020, notably in the first lockdown. For the whole year, however, uncertainty effects were small compared to the contribution of administrative spending constraints.
During the first Covid-induced lockdown in France, household savings increased strongly, with the savings ratio peaking at just above 27% in the second quarter of 2020 (Figure 1). After a temporary decline during the summer, it increased again significantly in 2020Q4 to 22%. What explains this strong increase in 2020? A significant part of these “excess savings” can be attributed to “forced savings”: while household disposable income has been largely protected by fiscal transfers, private consumption has been limited by administrative restrictions on travel, restaurants, shops and most recreational activities. However, the rise in economic and financial uncertainty since the outbreak of the Covid-19 pandemic (Figure 2) points to a potential role for precautionary motives. In this blog, we provide evidence that a small though non-negligible share of the Covid-related increase in household savings can be attributed to precautionary motives.
There are several reasons why it is important to determine whether precautionary motives play a role in households’ current and future saving and consumption plans. First, disentangling forced from precautionary savings makes it possible to quantify the share of the consumption decline directly attributable to restrictions due to lockdown policies. Second, high and persistent levels of uncertainty can mitigate the decline of excess savings even after lockdown measures are lifted, with adverse effects for aggregate demand. Third, high precautionary savings can limit the effectiveness of economic policies. For instance, if households prefer saving over spending when facing uncertainty about their future income, the transmission of expansionary policies intended to foster aggregate demand may be dampened.
We first discuss the empirical link between uncertainty and household consumption and its importance in the current context. To this end, we estimate an equation for household consumption that is closely aligned with the specification in our semi-structural model for France (EN-BDF), but also includes uncertainty indicators that affect short-run consumption. Based on a variety of tested indicators, our baseline model features three uncertainty measures. While the change in the unemployment rate serves as a direct measure of households’ precautionary savings motive, we also include a measure based on the within-quarter variation in the “Country-Level Index of Financial Stress (CLIFS)” for France (an index provided by the ECB based on financial stress measures in different market segments). Moreover, we use the “major purchase intention” sub-index of the INSEE consumer survey to capture broad real and financial uncertainty.
Figure 3 shows the cumulated dynamic contribution to household consumption implied by our equation. Throughout the year, forced savings effects captured in the residual explain a large share of the variation in household consumption. Still, about one quarter (i.e. -3 ppt.) of the fall in consumption growth during the first lockdown in 2020Q2 can be attributed to economic and financial uncertainty (Figure 3). In particular, uncertainty related to a rise in the unemployment rate appears to have negatively affected household consumption in the spring (-1.3 ppt.). These effects were in part transitory, as household purchase intentions almost returned to their pre-crisis level in June (Figure 2). Thus, our results do not indicate a comparably strong role for precautionary savings in 2020Q3 and Q4.
In a second exercise, we examine the role of households’ expectations in simulations with FR-BDF. In the model, household consumption is affected, inter alia, via the link between the unemployment rate and expected disposable income, independently of any direct effect that changes in unemployment may have on current disposable income. Thus, the effect of the expectation channel can be isolated from any potential direct income channels through which unemployment might also affect consumption in the short run.
We examine the impact of a permanent increase in the unemployment rate of 2 ppt compared with the rate in a baseline scenario up to end-2023. By only allowing for consumption changes related to the link between unemployment and expected future income in the alternative scenario, we ensure that the comparison of the two scenarios reflects only the expectation channel.
With a permanently higher unemployment rate, households’ expectations of future disposable income are lower. In turn, their incentive to accumulate savings is larger. This effect translates into cumulative “surplus” savings of about EUR 3 billion a year, i.e. an increase of around 0.2 ppt in the saving ratio. While this difference seems quantitatively moderate, it should be noted that it reflects only the isolated precautionary savings channel effect described above.
Overall, this blog indicates that precautionary motives can negatively affect household consumption in the current context. Even though forced savings likely explain the majority of the excess savings accumulated during the Covid-19 pandemic, precautionary motives seem to have played a significant role during the first lockdown. Should employment-related uncertainty increase again, precautionary motives could again contribute to excess savings accumulation. On the contrary, if pandemic related uncertainties were to decline markedly, this could have a positive effect on household consumption.