The propensity to consume remained largely unchanged from 2000 to 2012, with the exception of 2003 when the credit card crisis hit. Starting in 2012, however, it fell continuously before slightly rebounding after 2015. This development is distinguished from that seen in 2003 when the decline was immediately followed by a recovery over a short span of time.

In order to assess in detail the characteristics of the recent change in the propensity to consume, this paper breaks down the factors behind it into age groups of heads of household and into income levels, using micro-data from the Survey of Household Finances and Living Conditions and applying the method proposed by Kwon Kyu Ho and Oh Ji Yoon (2014). According to the result of the analysis, the recent decline in the propensity to consume was driven by households headed by those in their 50s and by high-income households. In addition, a look at the factors by spending purposes using the household final consumption expenditure data under the National Account shows that clothing, food and shelter-related basic necessity expenditures led the drop in the overall propensity to consume.

This paper then looks into what impact the characteristics of households, such as the age of the head of household and the income level, has on the propensity to consume through an empirical analysis, using the model of Paiella (2007) and Arrondel et al. (2015). It then interprets the results by applying the permanent income hypothesis. The analysis result shows that the changes in future expected income of households headed by those in their 50s and 60s explain the changes in the overall propensity since 2012 relatively well. In spite of the rapid pace of population ageing in Korea, the elderly do not yet have stable sources of income, resulting in a high level of concern about future income, particularly among households headed by those in their 50s. This sense of anxiousness is presumed to have caused to a significant extent the post-2012 decline in the propensity to consume. However, the government’s measures to strengthen its social-welfare policy led to an improvement in the living conditions and income satisfaction of the elderly, slightly slowing the rapid drop in the propensity to consume. Meanwhile, the result of the empirical analysis on the impact of household assets on the propensity to consume by age and income level shows that the wealth effect tends to decline as the age of the head of household rises and as the income level drops. This implies that the wealth effect could weaken as population ageing advances.

Based on the above empirical analysis, we can draw the following conclusions. First of all, considering that the recent change in the propensity to consume is largely affected by population ageing, which is a structural factor of the Korean economy, it will be difficult to return to the level seen in the early and mid-2000s. Secondly, given that households headed by those in their 50s respond sensitively to changes in future expected income, close attention should be paid when pursuing relevant policies going forward so as to avoid any rapid change in future expected income. Lastly, considering the rapid growth in the share of population aged 60 years or older, whose propensity to consume is relatively insensitive to changes in wealth, policymakers should pay attention to the structural changes taking place in the wealth effect when conducting future monetary policy.