I'm really not sure what importance to put on this, but it does seem in line with the idea that demographics, and the life cycle model, have something to say about what is happening in Japan. It seems to fit my picture nicely, which is why I'm cautious. There isn't going to be anything to cheer about here.
Despite a worsening income environment for employees, personal consumption has held up pretty well. Nominal employee compensation in calendar 2002 shrank by 2.4%, but nominal consumption was flat or up 1.5% in real terms. This has resulted in a notable decline in the rate of household savings. The household savings rate fell from the 14% range at the start of the 1990s to around 10% in the late 90s, and to 6.9% in 2001. Figures for 2002 are not yet available, but judging by the relevant data, we estimate that it slipped still further to about 4.3%. The savings rate based on sequential quarterly data (the 4-quarter moving average) has dropped even more startlingly. After dipping below 10% at the start of 1996, declines accelerated from 2001 to 4.7% in January-March 2002, the most recent period that can be calculated. Our estimate is that this figure was down to about 2.5% for October-December 2002, and that a negative savings rate is a possibility soon. This could happen within a few quarters, under our assumptions of an ongoing decline in disposable income and flat consumption.
Normally, the savings rate tends to rise during an economic recession, due to the tendency to suppress consumption amid rising uncertainty. Further, deflation is known to stimulate demand for bank notes and cause expenses to be put off. The classic business cycle theory suggests that the savings rate should now be rising. This makes the actual declining trend all the more alarming. We attribute this decline to (1) growth in the generation that is tapping into assets, a result of the aging of society and the increase in unemployment among older workers (the life cycle hypothesis); (2) reduced incentives to save and a growing preference for durables, due to extremely low interest rates; and (3) due to the difficulty in reining in spending after becoming accustomed to a certain level, spending is cut back, but not by enough to offset the decline in income (the habit-formation hypothesis).
Regarding (1) above, we assume the following age/life-cycle influenced consumption/saving pattern is in effect: consumers borrow when they are young and their incomes are low, save in preparation for old age in mid-life, and live off what they saved in their retirement years. The increase in unemployed older workers has played a substantial role in the decline in the savings rate in recent years. The savings rate for unemployed older workers stands at -26.0%, making it clear that savings are being drawn upon. On top of this, the decline in the savings rate resulting from the increase in unemployed older workers, in addition to being grounded in a long-term trend caused by the aging of the population, has been further exacerbated by earlier retirement age, early retirement programs and the dismissal of older workers. The percentage of households headed by unemployed aged 60 and over is conspicuously trending upward, having increased from around 12% in the early 1990s, to 19.1% in 2000 to 21.0% in 2001 and to 22.0% in 2002. Japan's population is expected to reach a peak in 2006 and then began to decline, but the fact that savings have begun to decline ahead of the population suggests greater urgency.
Regarding (2), the view that extremely low interest rates have destroyed the incentive to save is further supported by recent consumption patterns, which show that the consumption of durable goods has remained relatively strong. One incentive for holding financial assets is the function they provide as a store of value and wealth, but it is conceivable that low interest rates have diluted this "store of value" function of financial assets and thereby increased the preference for real assets. This increased preference for durables, which also function fairly well as a store of value, as opposed to regular consumer goods, could be considered further corroboration of this hypothesis. It is interesting to note that this massive shift of funds in 2002, which was sparked by both implementation of a deposit insurance cap and disappointment in investment trusts with net asset values below initial investment amounts, coincided with the period in which the consumption of durable goods was robust.
Source: Osamu Tanaka, Morgan Stanley Global Economic Forum