Chart 3 shows that the increase in the saving rate between the beginning of 2023 and the end of 2024 (of 1.5 percentage points) is mainly due to the composition effect of income. The portion saved from wages even initially decreased before returning at the end of 2024 to its level of the first quarter of 2023. In other words, the increase in the saving rate observed since 2023 appears to stem mainly from the shift in the composition of income towards financial income, rather than from a change in the underlying ratio between consumption and labour income.
An econometric approach to quantify the composition effect
The impact of the income composition effect presented above is purely a matter of accounting, insofar as no assumptions are made about the economic behaviour of households vis-à-vis the shift in the income structure. Furthermore, the contribution of the component linked to the change in the share of the wages in income calculated using the accounting approach tends to overestimate the composition effect because it is based on the implicit assumption that all income excluding wages is fully saved. In order to quantify this effect more precisely in the current rise in the saving rate, we use an econometric consumption equation similar to that presented in Ouvrard and Thubin (2020). This model takes into account the share of a number of components of household income in order to capture structural effects. In particular, we look at the share of income from assets (income from property and financial assets), taxes and social security contributions; and the gross operating surplus of individual entrepreneurs in household income. According to this equation, a shift in the structure of income from income from assets (1 percentage point increase in the GDI of income from assets, with total income remaining unchanged) would lead to a rise in the saving rate of around 0.4 of a percentage point.
We then carry out a simulation exercise for the 2023-24 period, in which we compare two scenarios. In the first scenario, the shares of the different components of income change in line with the data published by INSEE. In the second scenario, these shares remain at their levels of the first quarter of 2023, with an identical change in total income. We then interpret the difference in saving rates between these two scenarios as the contribution of the income composition effect to the change in the saving rate over this period.
The results of these simulations show that the income composition effect appear to have contributed 0.7 percentage point to the 1.5-percentage point increase in the saving rate in 2023-24. This effect appears to stem entirely from the increase in income from assets (Chart 1).
In 2025, we expect the saving rate to start falling, with a change in the composition of household income that would be more favourable to consumption than to savings. Indeed, the share of financial income in household gross disposable income is expected to fall in line with the decline in interest rates. At the same time, the delayed adjustment of nominal wages to the past rise in inflation should support the purchasing power of labour income, and therefore also consumption.