Households save more in fast growing economies. The respective experiences of the East Asian countries, such as Japan in the 1970s and Korea and Taiwan in the 1980s, attest to this empirical regularity. In the last decade, rising saving in the BRIC countries provides further evidence. However, these observations are difficult to reconcile with the representative agent model, since forward looking consumers with standard preference should save less in a high growth environment as they anticipate a higher level of earnings in the future relative to their present incomes (e.g., Tobin, 1967; Carroll and Summers, 1991). Although income growth can lead to increased saving in the life cycle model (e.g., Modigliani, 1970) or in models with habit formation (e.g., Carroll, Overland and Weil, 2000), the quantitative effect has shown to be small (e.g., Paxson, 1996). This is the household saving puzzle that remains unresolved .
Behind the rise in average saving, there is a second stylized fact regarding life cycle saving profiles. When Taiwan experienced rapid income growth in the period 1976-1990, the saving rates of young households were systematically higher than that of the old households (Deaton and Paxson, 1994). This empirical regularity is also observed in China. From 1992 to 2007, the average saving rate of young households with their heads under age 40 increased by about 11 percentage points (see Figure 3B). The high saving of the young in these two fast growing economies contrasts sharply with the typical hump shaped or relatively flat age saving pro files in developed economies. If growth raises the earnings of all workers, why don't young households consume more with higher future income?
The goal of this paper is to show that the high saving puzzle in fast growing economies can be resolved in a model of intertemporal choice, once we allow for structural shifts in life cycle earning profiles associated with high economic growth. Using unique and comprehensive Urban Household Survey data (UHS) covering the period 1992-2007, we document two striking changes in individual age earning profiles in the fast growing economy of China: (a) there were large upward shifts in the earnings of successive younger worker cohorts, and (b) the age earning profiles have become flattened during this period. These empirical findings are at variance with a stable age earning profile, an assumption made in virtually all models of life cycle saving decisions.
To illuminate how observed changes in earning profiles affect saving decisions through a transparent mechanism, we develop a simple four period overlapping generation (OLG) model with closed form solutions. We find that when the economy begins high income growth, workers of all ages choose a higher rate of saving; and, with the flattening of age earning profiles, young workers raise their saving rate more pronounced than older workers. These results are robust to incorporating within generation heterogeneity in worker types, the inclusion of a pension system, and alternative specifications of perfect foresight and myopic expectations of lifetime earnings. These results provide a novel mechanism for explaining the rise in saving rate particularly among young households in fast growing economies.
The remaining question is quantitative: To what extent can the observed flattening earning profiles with income growth explain the increase in aggregate saving, the rise in the saving rate of the young, and other saving behavior over the life cycle in China? Between 1992 and 2007, China's GDP per capita grew at a remarkable annual rate of 8.5 percent, while the aggregate urban household saving rate increased from 16.6 to 27.6 percent. In addition, consistent with Chamon and Prasad (2010), our national sample of UHS data also show that the age profiles of saving in China exhibit a U-shaped pattern in recent years, with younger and older households having relatively high saving rates (see Figure 3A). There is a more pronounced U-shaped increase in age specific saving rates across the initial (1992-1993) and the ending (2006-2007) periods of the sample (see Figure 3B). While high saving of the young is common in fast growing economies, the high saving of older workers before retirement is a special feature observed in China. To account quantitatively for these observations, we turn to a more sophisticated OLG model in which one period corresponds to one calendar year. Once incorporating the estimates on the structural changes in earning profiles, along with reduced pension benefits during transition, the model with standard parameterization can generate a trend of increasing aggregate saving that is comparable to its empirical counterpart. Moreover, the predicted increases of saving rate over the life cycle fit closely to the U-shaped pattern in the data. Therefore, our quantitative analysis not only provides an explanation for the observed high saving in a typical fast growing economy, but also for the special feature of the Chinese household saving.
Download
Highly Preliminary Draft Life Cycle Earnings and Saving in a Fast-Growing Economy
