The Lost Decade (Japan) PDF
The Lost Decade (Japan) PDF
THE CAUSES OF JAPANS LOST DECADE: THE ROLE OF HOUSEHOLD CONSUMPTION Charles Yuji Horioka Working Paper 12142 http://www.nber.org/papers/w12142 NATIONAL BUREAU OF ECONOMIC RESEARCH 1050 Massachusetts Avenue Cambridge, MA 02138 March 2006
Forthcoming in Japan and the World Economy, vol.18, no. 4 (December 2006). The views expressed herein are those of the author(s) and do not necessarily reflect the views of the National Bureau of Economic Research. 2006 by Charles Yuji Horioka. All rights reserved. Short sections of text, not to exceed two paragraphs, may be quoted without explicit permission provided that full credit, including notice, is given to the source.
The Causes of Japans Lost Decade: The Role of Household Consumption Charles Yuji Horioka NBER Working Paper No. 12142 March 2006 JEL No. D12, E21, E32, O47, O53 ABSTRACT In this paper, I analyze the causes of the prolonged slowdown of the Japanese economy in the 1990s and find that the stagnation of investment, especially private fixed investment, was the primary culprit. I then investigate the causes of the stagnation of household consumption during the 1990s and find that the stagnation of household disposable income, the decline in household wealth, and increased uncertainty about the future are among the contributing factors. Finally, I consider whether demand side factors or supply side factors were more important as causes of the prolonged slowdown of the Japanese economy in the 1990s and conclude that the former (especially misguided government policies) were probably more important. Charles Yuji Horioka Institute of Social and Economic Research Osaka University 6-1, Mihogaoka Ibaraki, Osaka 567-0047 JAPAN and NBER [email protected]
1. Introduction Japans economy has been stagnant more or less continuously for more than a decade (Japans so-called Lost Decade), and Japans growth rate during this period has been the lowest among the major industrialized countries of the world. During the
1995-2002 period, for example, the annualized growth rate of Japans real gross domestic product (GDP) averaged only 1.2%, which is lower than all of the other G7 countriesCanada (3.4%), the United States (3.2%), the United Kingdom (2.7%), France (2.3%), Italy (1.8%), and Germany (1.4%)--as well as the Euro area average (2.2%) and less than half of all of the other larger OECD countriesKorea (5.3%), Australia (3.8%), Spain (3.3%), the Netherlands (2.9%), and Mexico (2.6%)--as well as the OECD-wide average (2.7%). Figure 1 shows trends over time in the annual growth rate of real GDP and household consumption during the 1980-2003 period and, as can be seen from this figure, GDP growth was relatively high during the 1980-1991 period (the so-called bubble period) and much lower during the 1991-2003 period (the so-called post-bubble period). This is
verified by the first column of Table 1, which shows that GDP growth averaged a full 3.89% during the 1980-91 period but only 1.14% (less than a third of the 1980-1991 level) during the 1991-2003 period. What caused the sharp decline in GDP growth, and what caused it to persist for more than a decade? In this paper, I attempt to shed light on the causes of the prolonged
slowdown of the Japanese economy in the 1990s by analyzing demand side data on the sources of growth of GDP and the sources of growth of household consumption in conjunction with similar data on the immediately preceding 1980-1991 period. By so doing, I attempt to assess whether, and to what extent, the stagnation of household consumption is responsible for the prolonged slowdown of the Japanese economy. I then
speculate about the causes of the stagnation of household consumption, and after devoting
1
the bulk of the paper to demand side factors, I turn finally to a consideration of the relative importance of demand side and supply side factors as causes of the prolonged slowdown of the Japanese economy. This paper is organized as follows: section 2 considers the sources of growth of GDP, section 3 considers the sources of growth of household consumption, section 4 considers possible causes of the stagnation of household consumption, section 5 considers whether demand side or supply side factors contributed more to the prolonged slowdown
of the Japanese economy, and section 6 is a brief concluding section. To preview the main findings of this paper, I find that the stagnation of investment, especially private fixed investment, was the major culprit of the prolonged slowdown of the Japanese economy in the 1990s and that the stagnation of spending on clothing and footwear, transport, and to a lesser extent, miscellaneous goods and services, education, and food and non-alcoholic beverages were the main culprits of the stagnation of household consumption. By contrast, I find that the main factors holding up GDP growth were
household consumption, government consumption, and net exports and that the main factors holding up household consumption growth were spending on health, communication, and to a lesser extent, restaurants and hotels and housing-related expenditures. I also find that the stagnation of household consumption was due primarily
to the stagnation of household disposable income, the decline in household wealth (which in turn was due primarily to the collapse of land and equity prices), and to a lesser extent, increased uncertainty about the future (especially about old age in general and public old-age pensions in particular), the deterioration of future prospects, etc. Finally, I
consider whether demand side factors or supply side factors were more important as causes of the prolonged slowdown of the Japanese economy in the 1990s and conclude that the former (especially misguided government policies) were probably more important.
2. An Analysis of the Sources of GDP Growth In this section, I analyze the sources of GDP growth during the 1991-2003 period and consider the extent to which the stagnation of household consumption was responsible for the prolonged slowdown of the Japanese economy. For the purposes of this analysis, I decompose gross domestic product (GDP) into the following six components: actual final consumption of households (hereafter referred to as household consumption), government actual final consumption (government consumption), private gross domestic fixed capital formation (private fixed investment), public gross domestic fixed capital formation (government fixed investment), changes in inventories (inventory investment), and net exports of goods and services (net exports). Household consumption differs from the standard concept in that it includes social transfers in kind, which consist of social benefits in kind (health insurance and nursing care insurance benefits, which are used to finance the consumption of medical and nursing care services by households) and transfers of individual non-market goods and services from the government (such as subsidies for textbooks, public day care centers, etc.) and from private non-profit institutions serving households (such as subsidies for private nursery schools, art museums, zoos, etc.). Since both types of consumption are ultimately for the benefit of households, it seems preferable to include them in household consumption. Conversely, government consumption differs from the standard concept in
that it excludes social transfers in kind from the government to households. Looking first at Figure 1, it can be seen from this figure that the growth rate of household consumption exceeded that of GDP in eight out of the twelve years during the 1991-2003 period. This suggests that household consumption did not act as a drag on the
economy and rather that it prevented it from stagnating further. The first two columns of Table 1 show the average annualized real growth rate (hereafter referred to as growth rate) of GDP and the various components thereof during
3
the 1980-91 and 1991-2003 periods. Looking first at the 1991-2003 period, the growth rate of GDP during this period was a mere 1.14%, whereas the growth rate of household consumption was 1.56%. Thus, the growth rate of household consumption was somewhat
higher than that of GDP during the 1991-2003 period, which provides further corroboration that household consumption did not act as a drag on the economy and in fact raised rather than lowering the growth rate of GDP. Note, however, that the growth rate of household
consumption was not very high in absolute terms, that it was not nearly as high as the growth rate of certain other components of GDP such as net exports (6.39%) and government consumption (3.22%), and that it ranked only third among the six components of GDP. What did act as a drag on the economy was investment--private and government fixed investment as well as inventory investment, all of which showed slower growth than GDP and in fact all of which showed negative growth during the 1991-2003 period: -0.24% in the case of government fixed investment, -0.59% in the case of private fixed investment, and very negative but not calculable in the case of inventory investment (because it was negative in the terminal year). A breakdown of private fixed investment shows that
private housing (dwelling) investment declined especially sharply (-2.48% as opposed to -0.14% in the case of plant and equipment investment), which suggests that sluggish private housing investment might have been the primary culprit of the prolonged slowdown of the Japanese economy. The third and fourth columns of Table 1 show the contribution to real GDP growth of each component thereof during the 1980-91 and 1991-2003 periods. Looking first at the 1991-2003 period, household consumption made by far the largest contribution to real GDP growth (85.40%). Government consumption made the second largest contribution (18.64%), and net exports made the third largest contribution (13.07%). Some have
claimed that strong export growth prevented the prolonged slowdown of the Japanese
4
economy from becoming even worse, but the contribution of net exports to real GDP growth was relatively modest and it ranked only third. The contributions of government
fixed investment (-1.26%), inventory investment (-4.35%), and private fixed investment (-11.49%) to real GDP growth were all negative, with the contribution of private fixed investment being especially large in absolute magnitude. A breakdown of private fixed investment shows that private housing investment was responsible for 81% of the negative contribution of private fixed investment, suggesting again that it was the primary culprit of the prolonged slowdown of the Japanese economy. Some of the foregoing results concerning the growth rate of each component of GDP and its contribution to real GDP growth are seemingly at odds with one other, but the differences can be explained by the share of each component in total GDP. For example,
household consumption made by far the largest contribution to real GDP growth even though it ranks only third with respect to growth rates simply because it is by far the largest component of GDP with a share of more than 60% (see Table 2). Conversely, the
contribution of net exports to real GDP growth was only third highest even though its growth rate was by far the highest of any component of GDP simply because its share of GDP is so small (3% or less, which puts it in fifth place among the six components of GDP). Next, I compare the sources of growth of the Japanese economy during the 1991-2003 period to those during the immediately preceding 1980-91 period. first at the Looking
growth rates of GDP and its components (compare the first and second growth rate of GDP during the 1991-2003 period was less than a Household
third of what it was during the 1980-91 period (1.14% vs. 3.89%).
consumption showed the third highest growth rate of any component of GDP in both time periods, but its growth rate was much lower in absolute terms during the 1991-2003 period than it was during the 1980-91 period (1.56% vs. 3.59%).
5
household consumption exceeded that of GDP during the 1991-2003 period whereas it fell slightly short of that of GDP during the 1980-91 period. If we compare the growth rates of the remaining components of GDP during the two time periods, the growth rate of net exports was much higher during the 1991-2003 period than it was during the 1980-91 period, and its rank was also much higher (first vs. sixth). Similarly, the growth rate of government consumption was only slightly lower
during the 1991-2003 period than it was during the 1980-91 period, and its rank was quite a bit higher (second vs. fourth). By contrast, the growth rates of the investment-related
components of GDP (private fixed investment, government fixed investment, and inventory investment) were all negative during the 1991-2003 period even though they were all positive and inventory investment and private fixed investment ranked first and second, respectively, during the 1980-91 period. Turning next to a comparison of the contribution to real GDP growth of each component of GDP during the two time periods (compare the third and fourth columns of Table 1), and looking first at household consumption, household consumption made by far the largest contribution to real GDP growth during both time periods, but its contribution was much larger during the 1991-2003 period than during the 1980-1991 period (85.40% vs. 57.24%) even though its growth rate was much lower than during the 1980-91 period and even though its growth rate was so low in absolute terms because the growth rate of GDP was even lower and because its share of GDP is by far the largest of any component of GDP. Government consumption and net exports made the second and third largest contributions to real GDP growth during the 1991-2003 period even though their contributions were only the third and sixth largest during the 1980-91 period, and the magnitudes of their contributions were also much larger during the 1991-2003 period than during the 1980-91 period. By contrast, private fixed investment, government fixed
6
investment, and inventory investment (especially private fixed investment) all made negative contributions to real GDP growth during the 1991-2003 period even though their contributions were positive during the 1980-91 period. The most striking case is that of
private fixed investment, whose contribution to real GDP growth was a full 34.09% during the 1980-91 period but -11.49% during the 1991-2003 period, a decline of 45.58 percentage points! This indicates how volatile investment has been and suggests that investment
(especially private fixed investment) may have been the main culprit of the slowdown of the Japanese economy after 1991. To examine this contention more directly, I next analyze the change in the average annualized real growth rate of each component of GDP and the contribution of each component of GDP to the decline in real GDP growth between 1980-91 and 1991-2003. Looking first at the change in growth rates (see the first column of Table 3), the growth rate of consumption declined by 2.03 percentage points between 1980-91 and 1991-2003, which is considerably less than the decline in the growth rate of GDP (2.75 percentage points). This confirms my earlier contention that household consumption did not act as a drag on the economy. Looking at the change in the growth rates of the other components of GDP, net exports were the only component showing an increase in its growth rate (by 5.60 percentage points). The other components of GDP all showed declines in their growth
rates: 0.32 percentage points in the case of government consumption, 1.09 percentage points in the case of government fixed investment, 6.70 percentage points in the case of private fixed investment, and not calculable but large in the case of inventory investment. A closer look at private fixed investment shows that the decline in the growth rate of private plant and equipment investment (-7.59%) was much larger in absolute magnitude than the decline in the growth rate of private housing investment (-4.92%), which appears to contradict my earlier contention that the decline in private housing investment was the
7
primary cause of the slowdown of the Japanese economy in the 1990s. Turning finally to the contribution of each component of GDP to the decline in real GDP growth between 1980-91 and 1991-2003 (see the second column of Table 3), private fixed investment made by far the largest contribution to the decline in real GDP growth (67.91%), followed by household consumption in second place (36.34%), inventory investment in third place (5.22%), and government fixed investment in fourth place (3.84%). The contributions of the remaining components of GDP (government
consumption and net exports) to the decline in real GDP growth were negative, meaning that they held up real GDP growth and prevented it from falling further. A closer look at
private fixed investment shows that the contribution of private plant and equipment investment to the decline in real GDP growth (54.86%) was much larger than the contribution of private housing investment (13.05%), which casts further doubt on my earlier contention that the decline in private housing investment was the primary cause of the prolonged slowdown of the Japanese economy in the 1990s. These results confirm that private fixed investment and, to a lesser extent, government fixed investment and inventory investment were the main culprits of the prolonged slowdown of the Japanese economy in the 1990s, that net exports and government consumption prevented the slowdown from becoming even worse, and that household consumption was somewhere in the middle, contributing the most to real GDP growth but, at the same time, contributing substantially to the decline in real GDP growth. A detailed analysis of the reasons why private fixed investment was the main culprit of the prolonged slowdown of the Japanese economy during the 1990s is beyond the scope of this paper, but previous studies have attributed it to a combination of factors including (1) the sharp curtailment of bank lending (the so-called credit crunch) and the increase in systemic risk, both of which were caused by the financial crisis and the proliferation of non-performing loans, which in turn were caused by the collapse of the
8
bubble economy of the late 1980s and the subsequent decline in asset (land and equity) prices, (2) a further curtailment of bank lending due to the untimely introduction of the Basel guidelines for capital adequacy in 1993, (3) the inadequacy of government actions aimed at resolving the financial crisis and the non-performing loan problem, (4) the inadequacy of aggregate demand due in large part to the inadequacy of monetary and fiscal stimuli, (5) increased uncertainty about future prospects for the Japanese economy as well as increased volatility, and (6) massive overinvestment in corporate plant and equipment during the bubble years (due in large part to overly expansionary monetary policy), which induced firms to sharply curtail fixed investment during the post-bubble years as a way of reducing excess capacity in the corporate sector (see section 5 for more details). Saito (2000) does a similar analysis for the earlier postwar period and finds that household consumption made the largest contribution to real GDP growth during the earlier postwar period as well but that its contribution (and also the contribution of government consumption) were not as large as they were after 1991. Thus, the role played by
household (and government) consumption during the post-1991 period was large not only in absolute terms but also relative to the earlier postwar period.
3. An Analysis of the Sources of Consumption Growth I turn next to an analysis of the sources of growth of household consumption. I
break household consumption down into twelve components using the same classification scheme used in the National Accounts of Japan. In this section, I use the same concept of household consumption used in the previous section except that I exclude transfers of individual non-market goods and services from the government and from private non-profit institutions serving households to households because a detailed breakdown thereof by purpose is not available for all years. However, I do include social benefits in kind in household consumption and, in particular, I include them in the health
9
component of household consumption because they consist of health insurance and nursing care insurance benefits, which are used to finance the consumption of medical and
nursing care services by households. The first two columns of Table 4 show data on the average annualized real growth rate of each component of household consumption during the 1980-91 and 1991-2003 periods, and looking first at the 1991-2003 period, communication (11.69%) showed by far the most rapid growth during this period (because of the rapid diffusion of cell phones), and health (3.66%), recreation and culture (2.99%), housing, electricity, gas and water supply (2.19%), and restaurants and hotels (1.88%) also grew faster than overall consumption (1.54%), thereby boosting it. By contrast, furnishings, household
equipment and household services (1.12%), miscellaneous goods and services (0.66%), transport (0.53%), food and non-alcoholic beverages (-0.02%), alcoholic beverages and tobacco (-0.17%), education (-1.15%), and clothing and footwear (-2.73%) grew less slowly than overall consumption, thereby serving as a drag thereon. The third and fourth columns of Table 4 show the contribution of each component of household consumption to real household consumption growth during the 1980-91 and 1991-2003 periods, respectively, and looking first at the 1991-2003 period, housing, electricity, gas and water supply made the largest contribution (30.11%) because it showed the fourth highest growth rate and because it has by far the largest share (see Table 5), while health made the second largest contribution (26.80%) because it showed the second highest growth rate and because it has one of the largest shares. The relatively large
contribution of housing-related expenditures was due primarily to the increase in the imputed rent on owner-occupied housing (imputed services of owner-occupied dwellings). Imputed rent grew at a rate of 2.34% (vs. 2.19% in the case of housing-related expenditures as a whole), its share of housing-related expenditures is about 70%, and its contribution to real household consumption growth was 22.65% (vs. 30.11% in the case of housing-related
10
expenditures as a whole).
mismeasured because it is not a market transaction but rather is imputed using various assumptions, and I doubt that it grew as rapidly as the official figures suggest during a period when land prices were declining steadily. Thus, the fact that imputed rent on
owner-occupied housing was the single largest source of growth of household consumption during the 1991-2003 period, according to the official figures, suggests that there is considerable uncertainty about exactly how much consumption grew during this period. Recreation and culture made the third largest contribution (21.76%) because its growth rate was the third highest and because its share is the third or fourth largest, while communication made the fourth largest contribution (15.32%) even though its share is one of the smallest because its growth rate was by far the highest. Restaurants and
hotels made the fifth largest contribution (7.36%), followed by miscellaneous goods and services, furnishings, household equipment and household services, and transport. By contrast. food and non-alcoholic beverages, alcoholic beverages and tobacco, education, and clothing all made negative contributions, with the negative contribution of clothing being the largest in absolute magnitude (-9.83%). Next, I compare the sources of growth of household consumption during the 1991-2003 period to those during the immediately preceding 1980-91 period. Looking
first at the growth rates of each component of household consumption in the two periods (compare the first two columns in Table 4), the growth rate of household consumption as a whole declined from 3.61% in 1980-91 to 1.54% in 1991-2003. Communication
showed the highest growth rate during both periods, but its growth rate increased sharply. Health increased from sixth to second, housing, electricity, gas and water supply from seventh to fourth, and restaurants and hotels from ninth to fifth, even though the growth rates of the first two of these components declined. By contrast, furnishings, household equipment and household services declined from third to sixth, miscellaneous goods and
11
services from third to seventh, transport from fifth to eighth, and clothing and footwear from eighth to twelfth (last place), with the growth rates of these components declining sharply. Finally, recreation and culture, food and non-alcoholic beverages,
alcoholic beverages and tobacco, and education showed little change in their ranks but their growth rates declined considerably. Looking next at the contribution of each component to the growth of real household consumption in the two periods (compare the third and fourth columns of Table 4), the contribution of housing, electricity, gas and water supply increased sharply, allowing it to maintain its number one ranking, the contribution of recreation and culture increased somewhat but its rank fell from second to third nonetheless, and the contribution of health more than doubled, allowing it to increase its rank from fifth to second. Communication and restaurants and hotels showed sharp increases in their ranks as well as their contributions, whereas miscellaneous goods and services, furnishings, household equipment and household services, transport, food and non-alcoholic beverages, and clothing and footwear showed declines in their ranks as well as their contributions. Finally, alcoholic beverages and tobacco and education showed little
change in either their ranks or their contributions. I next analyze the change in the average annualized real growth rate of each component of household consumption and the contribution of each component to the decline in the real growth of household consumption between 1980-91 and 1991-2003. Looking first at the change in growth rates (see the first column of Table 6), the growth rate of household consumption as a whole declined by 2.06 percentage points between 1980-91 and 1991-2003, and the change in the growth rates of each component should be compared against this standard. The only components of household consumption showing increases in their growth rates between the two time periods were communication (an increase of 3.77 percentage points) and restaurants and hotels (an increase of 0.10 percentage points),
12
but health, food and non-alcoholic beverages, alcoholic beverages and tobacco, and housing, electricity, gas and water supply showed smaller declines in their growth rates (in absolute magnitude) than overall consumption, thereby propping up consumption,
whereas education, furnishings, household equipment and household services, clothing and footwear, recreation and culture, transport, and miscellaneous goods and services showed sharper declines in their growth rates (in absolute magnitude) than overall consumption, thereby contributing to its decline. Turning finally to the contribution of each component to the decline in the real growth of household consumption between the two periods (see the second column of Table 6), transport contributed the most to the decline in real household consumption growth (31.18%), followed by miscellaneous goods and services, clothing and footwear, food and non-alcoholic beverages, furnishings, household equipment and household services, recreation and culture, housing, electricity, gas and water supply, education, and alcoholic beverages and tobacco. Restaurants and hotels, health, and
communication made negative contributions to the decline in household consumption (i.e., propped up real growth in household consumption), with the contributions of communication and health being especially large in absolute magnitude. To sum up, the relative importance of the various components of household consumption differs depending on which criterion is used to rank them, but the components that rank relatively high with respect to virtually all criteria include communication and health and, to a lesser extent, restaurants and hotels and housing, electricity, gas and water supply, while the components that rank relatively low with respect to virtually all criteria include clothing and footwear, transport, and, to a lesser extent, miscellaneous goods and services, education., and food and non-alcoholic beverages. With the exception of housing-related expenditures, necessities (such as clothing and footwear, and food and non-alcoholic beverages) were the most stagnant and
13
contributed the most to the stagnation of household consumption during the 1991-2003 period, whereas with the exception of transport and education, luxuries (such as health, recreation and culture, communication, and restaurants and hotels) showed the strongest growth and contributed the most to holding up household consumption during this period. Thus, somewhat surprisingly, consumption patterns became more affluent during the 1991-2003 period despite the stagnation of household income and wealth, suggesting that the prolonged slowdown of the Japanese economy was not severe enough to impoverish Japanese households. However, it could be that the increasing affluence of consumption patterns is attributable to an increase in income and wealth disparities among Japanese households, and if this explanation is the correct one, it implies that the poor became even poorer. Changes in relative prices seem to be able to explain some of the observed patterns in household consumption. For example, the average annualized inflation rates of
communication and recreation and culture were negative and large in absolute magnitude (-3.78% and -3.39%, respectively), and thus, assuming that the demand for these components is relatively price-elastic, the negative inflation rates can help explain why the consumption of these components increased relatively rapidly. By contrast, the inflation
rate of education was positive and large in absolute magnitude (2.34%), and thus, assuming that the demand for education is relatively price-elastic, the high inflation rate thereof can help explain why the consumption thereof declined absolutely. Another possible influence on consumption patterns is demographic trends. The sharp decline in the birth rate has reduced the share of the young in the total population, and this in turn will reduce the demand for education, and conversely, sharp increases in life expectancy have increased the share of the aged in the total population, and this in turn will increase the demand for health-related expenditures. Yet another possible influence on consumption patterns is technological change.
14
For example, there were rapid improvements in cell phone and internet technology during the 1991-2003 period, which in turn led to sharp price declines and the introduction of new products and features like cell phones with email, internet, camera, and video capabilities. The rapid increase in communication-related expenditures is presumably due in large part to the price declines and new products and features made possible by rapid advances in cell phone and internet technology.
4. The Causes of the Stagnation of Household Consumption In section 2, we found that household consumption did not cause the prolonged slowdown in the Japanese economy, but it is nonetheless true that household consumption was relative stagnant during this period. In this section, we analyze the causes of the
stagnation of household consumption during the 1991-2003 period (see Bank of Japan Research Department (1998) for more details). Economic theory predicts that household consumption will be influenced by the following factors, among others: (1) household disposable income, (2) household wealth, (3) uncertainty about the future (for example, about income, employment, retirement, public old-age pensions, etc.), and (4) future prospects (for example, about income, employment, etc.). These factors may have contributed to the stagnation of household consumption during the 1991-2003 period if (1) household disposable income had declined or been stagnant, (2) household wealth had declined, (3) uncertainty about the future had increased, or (4) future prospects had deteriorated during this period. factors in turn. We look at each of these
(1) The Stagnation of Household Disposable Income The average annualized real growth rate of household disposable income was only
15
0.98% during the 1991-2003 period, which is far less than the average annualized real growth rate of household consumption during the same period (1.56%) and also far less than the average annualized real growth rate of household disposable income during the 1980-91 period (3.32%). This suggests that the stagnation of household disposable
income was a major cause of the stagnation of household consumption and that it would have caused household consumption to be even more stagnant had it not been for other factors operating in the opposite direction.
(2) The Decline in Household Wealth Household wealth (net worth) declined during the 1991-2003 period as a whole, due largely to the sharp decline in land and equity prices, and the average annualized real rate of decline of household wealth during this period was 0.39%. Thus, it is quite
possible that the stagnation of household consumption during this period was due at least partly to the decline in household wealth (a reverse wealth effect).
(3) Increased Uncertainty about the Future If the stagnation of household consumption were due to increased uncertainty about the future, we would expect the household saving rate to have increased, but in fact it declined steadily and sharply during the 1991-2003 period (except during the 1996-98 period)--from 15.1% in 1991 to 6.4% in 2002 in the case of the unadjusted rate and from 13.3% in 1991 to 5.4% in 2002 in the case of the adjusted rate, a decrease of more than 50% in both cases (see Table 7)! This suggests that the stagnation of household
consumption is not due to increased uncertainty about the future, except possibly during the 1996-98 period, when the household saving rate increased from 9.9% to 11.1% in the case of the unadjusted rate and from 8.5% to 9.6% in the case of the adjusted rate. It is not
surprising to find that the household saving rate showed a temporary upturn during the
16
1996-98 period because it is during this period that a spate of bankruptcies (most notably the bankruptcies of Yamaichi Securities and Hokkaido Takushoku Bank in November 1997) occurred in the financial sector. These bankruptcies caused increased uncertainty not only
about the health of the financial sector but also about the employment situation because the bankruptcies entailed large-scale layoffs of workers. Additional verification of the importance of uncertainty about the future can be obtained from data on saving motives, etc., from the Public Opinion Survey on Financial Assets and Liabilities, conducted annually by the Central Council for Financial Services Information. As Table 8 shows, the proportion of respondents saving for illness and
unforeseen emergencies has not shown a clear trend over time, and the proportion saving for peace of mind increased only moderately and only until 1999, which provides further corroboration that increased uncertainty about the future is not a major cause of the stagnation of household consumption. However, Table 8 also shows that the proportion of
respondents saving for old age has increased sharply over time (from 50.5% in 1991 to 60.4% in 2003), and moreover, Table 9 shows that the proportion of under-60 respondents who are worried about old age has increased sharply over time, from 63.7% in 1992 to 87.9% in 2003 and that the proportion of these respondents who are worried about old age because pensions and insurance are not adequate increased from 55.5% to 72.2% over the same time period, making it the reason that increased the most in importance. This
suggests that increased uncertainty about old age and about old-age pensions did contribute toward the stagnation of household consumption. In a related line of research, Horioka and Watanabe (1997) and Horioka, et al. (2000) calculate the amount of saving for each motive and find that the retirement and precautionary motives are by far the most important motives for saving and that they are far more important in Japan than they are in the United States. Similarly, Horioka, Murakami, and Kohara (2002) and Horioka, Kohara, and Murakami (2004) find that dissaving is the
17
most common way in which the Japanese deal with unforeseen emergencies, which corroborates the importance of the precautionary motive as a motive for saving in Japan from the other side of the ledger. A number of studies have analyzed the importance in Japan of precautionary saving arising from various types of uncertainty. For example, Ginama (1988), Ogawa
(1991), Doi (2001), and Zhou (2003) analyze the importance of precautionary saving arising from income risk. Ginama (1988), Ogawa (1991), and Zhou (2003) find that
precautionary saving arising from income risk is relatively unimportant except at the time of the first oil crisis and except for the self-employed and farmers, and Doi (2003) finds that it is of some importance in the case of salaried worker households but that employment risk is not important (but see also the next subsection). Murata (2003a, 2003b) looks at the importance of precautionary saving arising from uncertainty about overall economic conditions and from uncertainty about public old-age pensions and finds that the former is not important but that the latter is. Nakagawa (1999) analyzes the importance of different types of uncertainty on the household saving rate in Japan by age and income and finds that different types of uncertainties are important for different age and income groups: he finds that income risk is important for the low- to middle-income, that employment risk is important for the middle-aged and aged low-income, that uncertainty about public old-age pensions is important for the young, and that the risk of becoming bedridden is important for the aged. Finally, Saito and Shiratsuka (2003a, 2003b) analyze the impact of various
types of uncertainty on the household saving rate and find that uncertainty about employment and price deflation and, to a lesser extent, overall uncertainty and uncertainty about income exert upward pressure on the household saving rate. Thus, the available evidence is not always consistent, but it suggests that precautionary saving arising from income risk or employment risk has generally not been all that important nor has it increased over time but that precautionary saving arising from
18
uncertainty about old age in general and about public old-age pensions in particular is important and has increased over time and hence that it may have contributed to the stagnation of household consumption during the 1990s.
(4) The Deterioration of Future Prospects In the previous subsection, we discussed the impact of increased uncertainty concerning the future, but a closely related factor is the deterioration of future prospects. If household expectations concerning future incomes, future employment prospects, etc., deteriorate, this should cause them to reduce their current consumption. In addition to
looking at the impact of income and employment uncertainty, Doi (2001, 2003) also looks at the impact of the deterioration of income and employment prospects on the household saving rate in Japan and finds that reduced employment prospects have had a positive and significant impact on Japans household saving rate but that the impact of reduced income prospects is marginal at best. Japans unemployment rate has increased steadily
throughout the 1990s, reaching its highest level ever (5.5%) in August 2002 and declining only moderately thereafter. Thus, Dois finding that reduced employment prospects have
induced Japanese households to save more and consume less is not at all surprising.
(5) Other Factors Turning finally to other factors that may have influenced the level of household consumption, (1) the profits of individual proprietors have been stagnant during the current slowdown, putting a damper on the consumption of individual proprietors and (2) deflationary expectations concerning consumer prices may have depressed household consumption because consumer prices have been falling since the mid-1990s (with the exception of 1997) and price deflation means that the longer one waits, the cheaper one can buy a given item. By contrast, (3) price deflation might actually stimulate household
19
consumption because it increases the real value of household asset holdings; (4) near-zero (nominal) interest rates may have depressed household saving and boosted household consumption to the extent that the interest elasticity of saving is positive, (5) the rapid aging of the population may also have boosted household consumption because the elderly typically finance their living expenses in large part by decumulating their previously accumulated savings, as a result of which their propensity to consume is typically higher than that of the working-age population, and (6) the introduction of a public nursing care insurance program in 2000 may have weakened the perceived need to save, thereby boosting consumption. Thus, there are factors working in both directions, and the factors
that exert upward pressure on household consumption have presumably prevented the stagnation of household consumption from becoming worse than it would have been otherwise.
(6) Summary In sum, the evidence suggests that the stagnation of household disposable income and the decline in household wealth (the latter of which was due primarily to the collapse of land and equity prices) appear to have been the main causes of the stagnation of household consumption during the 1990s. Increased uncertainty about the future does not appear to
have been a major cause of the stagnation of household consumption during the decade as a whole, but it does appear to have been of some importance during the 1996-98 period, and increased uncertainty about old age in general and about public old-age pensions in particular may have contributed to the stagnation of household consumption during the period as a whole.
5. Demand Side vs. Supply Side Factors Thus far, I have focused on the demand side in my search for the causes of the
20
prolonged slowdown of the Japanese economy in the 1990s, and many other authors including Bayoumi (2001), Harada and Iwata (2002), Noguchi (2002, 2004), Posen (1998), and Takemori (2002) also emphasize the importance of demand side factors. However,
other authors including Hayashi (2003), Hayashi and Prescott (2002), Kawamoto (2004), Kobayashi and Inada (2005), Kobayashi and Katou (2001), Miyagawa (2003, 2004), and Ogawa (2003) emphasize the importance of supply side factors (see Miyao (2006) for a concise and useful survey). Authors who find that demand side factors are more important as explanations of the prolonged slowdown of the Japanese economy emphasize the importance of such factors as (1) the sharp curtailment in bank lending (the so-called credit crunch) and the increase in systemic risk, both of which were caused by the financial crisis and the proliferation of non-performing loans, which in turn were caused by the collapse of the bubble economy of the late 1980s and the subsequent decline in asset (land and equity) prices, (2) a further curtailment of bank lending due to the untimely introduction of the Basel guidelines for capital adequacy in 1993, (3) the inadequacy of government actions aimed at resolving the financial crisis and the non-performing loan problem, (4) inadequate monetary and fiscal stimulus leading to inadequate aggregate demand, (5) increased uncertainty about future prospects for the Japanese economy as well as increased volatility, and (6) massive overinvestment in corporate plant and equipment during the bubble years (due in large part to overly expansionary monetary policy), which induced firms to sharply curtail fixed investment during the post-bubble years as a way of reducing excess capacity in the corporate capital stock. Turning to authors who emphasize the importance of supply side factors, Hayashi and Prescott (2002) find that the prolonged slowdown of the Japanese economy during the 1990s was due primarily to the decline in the growth rate of total factor productivity and
to the reduction in working hours from 44 hours per week to 40 hours per week during the
21
1988-93 period pursuant to the revision of the Labor Standards Law in 1988 and that it was not due to the breakdown of the financial system (except during the 1996-98 period). By
contrast, Kawamoto (2004) finds little or no decline in the pace of technological change during the 1990s and attributes the measured slowdown in productivity growth (and economic growth) to cyclical fluctuations in the utilization of capital and labor and in the reallocation of inputs across sectors. As another example, Kobayashi and Inada (2005)
finds that the economic slowdown during the early 1990s was due to the downward rigidity of nominal wages, which increased real wages and induced companies to cut back on employment. Finally, Miyao (2006) finds that, at least since 1993, persistent negative
productivity shocks have caused the GDP gap to widen and that they also had feedbacks effects on aggregate demand by causing long-term growth prospects to deteriorate. This discussion has shown that the prolonged slowdown of the Japanese economy in the 1990s was due to both demand side and supply side factors and, moreover, that the two are often closely interrelated. Thus, there is no easy answer to the question of
whether demand side factors or supply side factors were more important, but my reading of the data and of the literature convinces me that demand side factors were probably more important and that the single most important cause of the prolonged slowdown of the Japanese economy in the 1990s was the stagnation of private fixed investment, which in turn was caused by overinvestment in plant and equipment during the bubble economy of the late 1980s, the collapse of asset prices during the post-bubble period, and an inadequate policy response to these events. In particular, I feel that policy mistakes during the bubble period (e.g., overly expansionary monetary policies) as well as during the post-bubble period (e.g., overly contractionary fiscal and monetary policies and the inadequacy of government actions aimed at resolving the financial crisis and the non-performing loan problem) are largely to blame.
22
6. Conclusion In this paper, I analyzed the causes of the prolonged slowdown of the Japanese economy in the 1990s and found that the stagnation of investment, especially private fixed investment, was the major culprit of the prolonged slowdown and that the stagnation of spending on clothing and footwear, transport, and to a lesser extent, miscellaneous goods and services, education, and food and non-alcoholic beverages were the main culprits of the stagnation of household consumption. By contrast, I found that the main factors holding
up GDP growth were household consumption, government consumption, and net exports and that the main factors holding up household consumption growth were spending on health, communication, and to a lesser extent, restaurants and hotels and housing-related expenditures. I also found that the stagnation of household consumption was due
primarily to the stagnation of household disposable income, the decline in household wealth (which in turn was due primarily to the collapse of land and equity prices), and to a lesser extent, increased uncertainty about the future (especially about old age in general and public old-age pensions in particular), the deterioration of future prospects, etc. Finally, I
considered whether demand side factors or supply side factors were more important as causes of the prolonged slowdown of the Japanese economy in the 1990s and concluded that the former (especially misguided government policies) were probably more important. Turning to the policy implications of my analysis, my findings suggest that the economic policies of the Japanese government were largely misguided during the bubble period of the late 1980s as well as during the post-bubble period of the 1990s. For
example, monetary policy should have been tightened sooner during the bubble period to prevent the persistence of such a pronounced bubble, and conversely, monetary policy should have been loosened sooner, more fiscal stimulus should have been provided, and the government should have acted sooner to resolve the financial crisis and the non-performing loan problem during the post-bubble period. In particular, the government should have
23
done more to stimulate private investment in housing and plant and equipment as well as household consumption, and in my opinion, the best way of doing so would have been to introduce temporary and targeted tax breaks for housing and plant and equipment investment, household consumption, etc. There are those who oppose tax breaks of any
kind because Japan already has the highest government debt-to-GDP ratio of any major industrialized nation in the world, but I feel that such temporary and targeted tax breaks should have been implemented for the following reasons: (1) the Japanese economy required further stimulus to recover more quickly, and in the absence of such stimulus, tax revenues declined even further, thereby causing the government debt to increase even further, (2) temporary and targeted tax breaks would have increased the government debt far less than more permanent and/or broad-based tax cuts, and (3) the tax breaks would
have been more effective if they had been temporary because temporary tax breaks would have induced firms and consumers to accelerate their purchases of the goods and services being targeted in order to take advantage of the tax breaks before they expired. A
temporary tax break for housing investment was tried and proved to be successful, suggesting that similar temporary tax breaks for investment in plant and equipment, research and development, and consumption would also have been effective. My analysis suggests that an alternative way of stimulating household consumption would have been to reduce uncertainty about the future, especially about old age in general and about public old-age pensions in particular--for example by fundamentally reforming the public old-age pension system to make it solvent as well as equitable. In 2003, there was a sharp increase not only in the proportion of people saving for old age but also in the proportion of people who are worried about old age because of the inadequacy of pensions and insurance (see Tables 6-7). Thus, there was an urgent
need to allay peoples fears about public old-age pensions, and doing so would have conferred the added benefit of stimulating household consumption.
24
Turning finally to policy recommendations that target the supply side, Hayashi and Prescotts (2002) findings imply that subsidies to inefficient firms and declining industries should have been discontinued since they presumably lower the overall rate of productivity growth. Such a policy would have brought about an improvement in government finances in addition to enhancing productivity growth and hence would have killed two birds with one stone. Now that the Japanese economic is staging a modest recovery, there may no longer be any need for stimulative fiscal and monetary policies, but at the very least, the Japanese government should be very cautious about tightening either fiscal or monetary policy until the economy has fully recovered to prevent the economy from slipping back into recession. Acknowledgements I am grateful to an anonymous referee, Anton Braun, Robert Dekle, Sheldon Garon, Fumio Hayashi, Takeo Hoshi, Michael Hutchison, Nazrul Islam, Douglas Joines, Akira Nagae, Tae Okada, Keunkwan Ryu, Shinpei Sano, Shizuka Sekita, Dong Shen, Dariusz Stanko, Midori Wakabayashi, Junmin Wan, Frank Westermann, Keiko Yoshida, Yanfei Zhou, participants of the CESifo Venice Summer Institute (VSI) Workshop on Economic Stagnation in Japan, which was held July 25-26, 2003, in Venice, Italy, and especially Bernadette Andreosso-OCallaghan (the designated discussant) for their helpful and insightful comments. The author is also grateful to the Ministry of Education, Culture, Sports,
Science and Technology of the Japanese Government for Grant-in-Aid for Scientific Research number 12124207, which supported this research.
25
References
Ando, Albert (2002), Missing Household Saving and Valuation of Corporations: Inquiry into Japanese National Accounts I, Journal of the Japanese and International Economies, vol. 16, no. 2 (June), pp. 147-176.
Ando, Albert; Christelis, Dimitrios; and Miyagawa, Tsutomu (2003), Inefficiency of Corporate Investment and Distortion of Savings Behavior in Japan, in Magnus Blomstrom, Jennifer Corbett, Fumio Hayashi, and Anil Kashyap, eds., Structural Impediments to Growth in Japan (Chicago and London: The University of Chicago Press), pp. 155-190.
Bank of Japan Research Department (1998), Saikin no Kojin Shouhi Doukou ni tsuite (On Recent Trends in Personal Consumption), Nippon Ginkou Chousa Geppou (Bank of Japan Monthly Research Report) (June), pp. 111-139 (in Japanese).
Bayoumi, Tamim (2001), The Morning After: Explaining the Slowdown in Japanese Growth in the 1990s, Journal of International Economics, vol. 53, no. 2 (April), pp. 241-259.
Bernanke, Ben S. (2000), Japanese Monetary Policy: A Case of Self-Induced Paralysis? in Mikitani and Posen (2000), pp. 149-166.
Blomstrom, Magnus; Corbett, Jennifer; Hayashi, Fumio; and Kashyap, Anil, eds. (2003), Structural Impediments to Growth in Japan (National Bureau of Economic Research Conference Report). Chicago, Illinois, U.S.A.: The University of Chicago Press.
26
Callen, Tim, and Ostry, Jonathan D., eds. (2003), Japans Lost Decade: Policies for Economic Revival. Washington, D.C., U.S.A.: International Monetary Fund.
Doi, Takero (2001), Chochiku-ritsu Kansuu ni motozuku Yobi-teki Chochiku Kasetsu no Kenshou (A Test of the Precautionary Saving Hypothesis using a Saving Rate Function), ESRI Discussion Paper no. 1, Economic and Social Research Institute, Cabinet Office, Government of Japan (in Japanese).
Doi, Takero (2003), Por que ahorran los hogares japoneses? Caidas en la renta y situacion del Mercado de trabajo (Precautionary saving and employment risk in the 1990s), Informacion Comercial Espanola, no. 807 (October), pp. 57-81 (in Spanish).
Ginama, Isamu (1988), Estimation of Precautionary Savings in the U.S. and Japan, unpublished Ph.D. dissertation, University of California, San Diego.
Hamada, Kouichi; Horiuchi, Akiyoshi; and Naikaku-fu Keizai Shakai Sougou Kenkyuusho (Economic and Social Research Institute) (2004), ed. Ronsou: Ushinawareta 10-nen (Debate: The Lost Decade). Tokyo: Nihon Keizai Shinpousha (in Japanese).
Harada, Yutaka, and Iwata, Kikuo, eds. (2002), Defure Fukyou no Jisshou Bunseki (Empirical Analysis of the Deflationary Recession). Tokyo: Touyou Keizai Shinpousha (in Japanese).
Hayashi, Fumio (2003), Kouzou Kaikaku nakushite Seichou nashi (No Growth without Structural Reform), in Iwata and Miyagawa (2003), pp. 1-20 (in Japanese).
27
Hayashi, Fumio, and Prescott, Edward C. (2002), The 1990s in Japan: A Lost Decade, Review of Economic Dynamics, vol. 5, no, 1 (January), pp. 206-235.
Horioka, Charles Yuji (2001), Japans Public Pension System in the Twenty-First Century, in Magnus Blomstrom, Byron Gangnes, and Sumner La Croix, eds., Japans New Economy: Continuity and Change in the Twenty-First Century (New York: Oxford University Press Inc.), pp. 99-119.
Horioka, Charles Yuji; Fujisaki, Hideki; Watanabe, Wako; and Kouno, Takatsugu (2000), Are Americans More Altruistic than the Japanese? A U.S.-Japan Comparison of Saving
and Bequest Motives, International Economic Journal, vol. 14, no. 1 (Spring), pp. 1-31.
Horioka, Charles Yuji; Kohara, Miki; and Murakami, Akane (2004), Defure Jidai ni okeru Nihonjin no Risuku Taisho-hou (The Risk Coping Mechanisms of the Japanese in a Deflationary Age), in Yoshio Higuchi, Kiyoshi Ohta, and the Institute for Research on Household
Economics, eds., Josei-tachi no Heisei Fukyoo (The Heisei Recession and Women) (Tokyo: Nihon Keizai Shinbunsha), pp. 235-260 (in Japanese).
Horioka, Charles Yuji; Murakami, Akane; and Kohara, Miki (2002), How Do the Japanese Cope with Risk? Seoul Journal of Economics, vol. 15, no. 1 (Spring), pp. 1-30.
Horioka, Charles Yuji, and Watanabe, Wako (1997), Why Do People Save?
Micro-Analysis of Motives for Household Saving in Japan, Economic Journal, vol. 107, no. 442 (May), pp. 537-552.
Ito, Takatoshi (2004), Retrospective on the Bubble Period and Its Relationship to
28
Iwata, Kikuo, and Miyagawa, Tsutomu (2003), Ekonomikusu Shiri-zu: Ushinawareta 10-nen no Shinin ha Nanika? (Economics Series: What Are the Real Causes of the Lost Decade?), Tokyo: Touyou Keizai Shinpousha (in Japanese).
Kinyuu Kouhou Chuuou Iinkai (The Central Council for Financial Services Information), ed., Kakei no Kinyuu Shisan ni kansuru Seron Chousa (Public Opinion Survey on Household Financial Assets and Liabilities), 2005 edition (and earlier editions of the same). Tokyo: Kinyuu Kouhou Chuuou Iinkai (The Central Council for Financial Services Information) (in Japanese).
Kobayashi, Keiichirou, and Inada, Masaru (2005), Business Cycle Accounting for the Japanese Economy, mimeo., The Research Institute of Economy, Trade and Industry, Tokyo, Japan.
Kobayashi, Keiichirou, and Katou, Souta, Nihon Keizai no Wana (The Trap of the Japanese Economy). Tokyo: Nihon Keizai Shinbunsha (in Japanese).
Kawamoto, Takuji (2004), What Do the Purified Solow Residuals Tell Us About Japans Lost Decade? IMES (Institute of Monetary and Economic Studies) Discussion Paper No. 2004-E-5, Bank of Japan (April).
Mikitani, Ryoichi, and Posen, Adam S., eds. (2000), Japans Financial Crisis and Its Parallels to U.S. Experience (Special Report 13). International Economics.
29
Miyagawa, Tsutomu (2003), Ushinawareta Juu-nen to Sangyou Kouzou no Tenkan (The Lost Decade and the Conversion of the Industrial Structure), in Iwata and Miyagawa (2003), pp. 39-61 (in Japanese).
Miyagawa, Tsutomu (2004), Nihon Keizai no Chouki Teitai to Kyoukyuu Saido (The Long-term Stagnation of the Japanese Economy and the Supply Side), in Hamada et al. (2004), pp. 3-27 (in Japanese).
Miyao, Ryuzo (2006), Nihon Keizai no Hendou Youin: Seisansei Shokku no Yakuwari (Causes of Fluctuations in the Japanese Economy: The Role of Productivity Shocks), Bank of Japan Working Paper Series No. 06-J-1 (January) (in Japanese).
Murata, Keiko (2003a), Mikuro De-ta ni yoru Kakei Koudou Bunseki: Shourai Fuan to Yobi-teki Chochiku (An Analysis of Household Behavior based on Micro Data: Uncertainty about the Future and Precautionary Saving), Kinkyuu Kenkyuu (Institute for Monetary and Economic Studies, Bank of Japan), vol. 22, no. 3 (September), pp. 23-58 (in Japanese).
Murata, Keiko (2003b), Precautionary Savings and Income Uncertainty: Evidence from Japanese Micro Data, IMES Discussion Paper No. 2003-E-9, Institute for Monetary and Economic Studies, Bank of Japan.
Naikakufu Keizai Shakai Sougou Kenkyuusho (Economic and Social Research Institute, Cabinet Office, Government of Japan) (2005), Kokumin Keizai Keisan Nenpou (Annual Report on National Accounts), 2005 edition (and earlier editions of the same). Dokuritsu Gyousei Houjin Kokuritsu Insatsu-kyoku (Government Printing Bureau).
30
Tokyo:
Nakagawa, Shinobu (1999), 90-nen-dai-irigo mo Nihon no Kakei Chochiku-ritsu ha Naze Takai-ka? (Why Is Japans Household Saving Rate High Even After Entering the 1990s?), Nippon Ginkou Chousa Geppou (Bank of Japan Monthly Research Report) (April), pp. 69-101 (in Japanese).
Noguchi, Akira (2002), Kouzou Mondai-setsu no Hihanteki Kaimei (A Critique of the Structural Problem Hypothesis), in Harada and Iwata (2002), pp. 9-39 (in Japanese).
Noguchi, Akira (2004), Nihon Keizai no Chouki Teitai ha Kouzou Mondai ga Genin ka? Sangyou Kouzou Chousei Furyou-setsu no Hihanteki Kentou (Are Structural Problems the Cause of the Long-term Stagnation of the Japanese Economy? A Critical Analysis of the
Inadequate Industrial Structural Adjustment Hypothesis), in Hamada et al. (2004), pp. 29-60 (in Japanese).
Ogawa, Kazuo (1991), Shotoku Risuku to Yobi-teki Chochiku (Income Risk and Precautionary Saving), Keizai Kenkyuu, vol. 42, no. 2 (April), pp. 139-152 (in Japanese).
Ogawa, Kazuo (2003), Daifukyou no Keizai Bunseki (Economic Analysis of the Great Recession). Tokyo: Nihon Keizai Shinbunsha (in Japanese).
Ohtake, Fumio (2005), Nihon no Fubyoudou (Inequality in Japan). Shinbunsha (in Japanese).
Posen, Adam S. (1998), Restoring Japans Economic Growth. for International Economics.
31
Saito, Makoto, and Shiratsuka, Shigenori (2003a), Precautionary Motives versus Waiting Options: Evidence from Aggregate Household Saving in Japan, IMES Discussion Paper No. 2003-E-2, Institute for Monetary and Economic Studies, Bank of Japan.
Saito, Makoto, and Shiratsuka, Shigenori (2003b), Yobi-teki Douki to Machi no Opushon: Wagakuni no Makuro Kakei Chochiku De-ta ni yoru Kenshou (Precautionary Motives versus Waiting Options: A Test based on Japanese Aggregate Household Saving Data),
Kinyuu Kenkyuu (Institute for Monetary and Economic Studies, Bank of Japan), vol. 22, no. 3 (September), pp. 1-18 (in Japanese).
Saito, Mitsuo (2000), The Japanese Economy (Economic Ideas Leading to the 21st Century, vol. 1). River Edge, N.J.: World Scientific Publishing Co., Inc.
Saxonhouse, Gary R., and Stern, Robert M., eds. (2004), Japans Lost Decade: Origins, Consequences and Prospects for Recovery. Malden, Massachusetts, U.S.A.: Blackwell Publishing.
Takemori, Shunpei (2002), Keizai Ronsen ha Yomigaeru (Economic Debate Is Revived). Tokyo: Touyou Keizai Shinpousha (in Japanese).
Zhou, Yanfei (2003), Precautionary Saving and Earnings Uncertainty in Japan: A Household-level Analysis, Journal of the Japanese and International Economies, vol. 17, no. 2 (June), pp. 192-212.
32
Figure 1: Trends over Time in GDP and Consumption Growth 8.0 7.0 6.0 5.0 4.0 Percent 3.0 2.0 1.0 0.0 -1.0 -2.0 GDP Consumption
1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2.9 2.8 1.6 3.1 5.1 3.0 3.8 6.8 5.3 5.2 3.4 1.0 0.2 1.1 1.9 3.4 1.9 -1.1 0.1 2.9 0.4 -0.5 2.5 1.8 4.4 3.2 2.5 3.5 3.4 4.0 4.7 4.6 4.4 3.0 2.6 1.6 2.7 2.1 2.6 0.9 -0.0 0.7 1.7 1.9 0.8 1.2 Calendar Year
Note: See Table 1 for the definition of consumption. Source: The same as Table 1.
33
Table 1: The Average Annualized Real Growth Rate and the Contribution to Real GDP Growth of Each Component of GDP, 1980-2003 Average Annual Real Growth Rate (percent) 1980-91 1991-2003 3.59 3 1.56 3 3.54 4 3.22 2 6.11 2 -0.59 5 (2.44) (-2.48) (7.44) (-0.14) 0.85 5 -0.24 4 (-0.53) (-0.89) (-2.85) (-1.06) (2.45) (0.02) 9.56 1 na 6 0.78 6 6.39 1 (4.61) (4.26) (5.94) (3.63) 3.89 1.14 Contribution to Real Growth Rate of GDP 1980-91 1991-2003 57.24 1 85.40 1 5.45 3 18.64 2 34.09 2 -11.49 6 (3.54) (-9.27) (30.54) (-2.22) 1.67 4 -1.26 4 (-0.04) (-0.15) (-1.56) (-1.18) (3.26) (0.07) 1.14 5 -4.35 5 0.42 6 13.07 3 (9.68) (37.78) (9.26) (24.71) 100.00 100.00
Component of GDP Actual final consumption of households Government actual final consumption Private gross domestic fixed capital formation Housing Plant and equipment Public gross domestic fixed capital formation Dwellings Plant and equipment General government Changes in inventories (inventory investment) Net exports of goods and services Exports of goods and services Imports of goods and services Gross domestic expenditure (product)
Notes: The average annual real growth rate of component X between year t1 and year t2 was calculated as [(X(t2)-X(t1))**(1/(t2-t1)) - 1]*100. The contribution of component X to real GDP growth was calculated as [X(t2) - X(t1)]*100/[GDP(t2)GDP(t1)].
The "actual final consumption expenditure of households" is the sum of the final consumption expenditure of households and social transfers in kind received. The Japanese government switched from the fixed base year method to the chain linking method in 2004, but all data shown in this paper are based on the previous fixed base year method because data based on the chain linking method are available only since 1994 and are available only for the main aggregates. The figures to the right of each figure indicate the rank of that figure. Source: Naikakufu Keizai Shakai Sougou Kenkyuusho (Economic and Social Research Institute, Cabinet Office, Government of Japan), ed., Kokumin Keizai Keisan Nenpou (Annual Report on National Accounts) , 2005 edition (Tokyo: Media-Rando Kabushiki Kaisha, 2005).
34
Table 2: The Composition of GDP, 1980-2003 Component of GDP Actual final consumption of households Government actual final consumption Private gross domestic fixed capital formation Housing Plant and equipment Public gross domestic fixed capital formation Dwellings Plant and equipment General government Changes in inventories (inventory investment) Net exports of goods and services Exports of goods and services Imports of goods and services Gross domestic expenditure (product) The Share of Each Component in Real GDP 1980 1991 2003 62.91 1 60.97 1 64.08 6.10 4 5.88 4 7.50 19.31 2 24.38 2 19.81 (6.08) (5.21) (3.36) (13.24) (19.17) (16.44) 8.91 3 6.43 3 5.45 (0.36) (0.22) (0.17) (2.99) (1.43) (1.10) (5.57) (4.78) (4.18) 0.34 6 0.62 6 -0.02 2.42 5 1.73 5 3.18 (7.87) (8.49) (12.22) (5.45) (6.76) (9.04) 100.00 100.00 100.00
1 3 2
6 5
Note: The figures show the share of each component of real GDP in real GDP (in percent). Also see the notes to Table 1. Source: The same as Table 1.
35
Table 3: The Decline in the Growth Rate and the Contribution to the Decline in Real GDP Growth of Each Component of GDP, 1980-2003 Contribution to the Decline in Real GDP Growth, 1980-91 vs. 1991-2003 (percent) 36.34 -4.33 67.91 (13.05) (54.86) 3.84 (0.05) (-1.84) (5.63) 5.22 -8.98 (-11.18) (-2.20) 100.00
Component of GDP Actual final consumption of households Government actual final consumption Private gross domestic fixed capital formation Housing Plant and equipment Public gross domestic fixed capital formation Dwellings Plant and equipment General government Changes in inventories (inventory investment) Net exports of goods and services Exports of goods and services Imports of goods and services Gross domestic expenditure (product)
Change in the Growth Rate, 1980-91 vs. 19912003 (percent) -2.03 4 -0.32 2 -6.70 5 (-4.92) (-7.59) -1.09 3 (-0.36) (1.79) (-2.44) na 6 5.60 1 (-0.35) (-0.32) -2.75
2 5 1
3 6
Notes: The contribution of each component to the decline in real GDP growth between 1980-91 and 19912003 was calculated as the ratio between the decline in the change in that component between 1980-91 and 1991-2003 and the decline in the change in GDP between 1980-91 and 1991-2003 (in percent). "na" denotes "not available." Also see the notes to Table 1. Source: The same as Table 1.
36
Table 4: The Average Annualized Real Growth Rate and the Contribution to the Real Growth of Household Consumption of Each Component of Household Consumption, 1980-2003 Contribution to the Real Growth of Household Annual Annualized Real Consumption (percent) Growth Rate (percent) Purpose of Expenditure 1980-91 1991-2003 1991-2003 1980-91 Food and non-alcoholic beverages 1.21 10 -0.02 9 5.83 7 -0.21 9 Alcoholic beverages and tobacco 1.06 11 -0.17 10 1.03 11 -0.31 10 Clothing and footwear 1.85 8 -2.73 12 3.98 8 -9.83 12 Housing, electricity, gas and water supply 3.76 7 2.19 4 21.15 1 30.11 1 Imputed service of owneroccupied dwellings (4.05) (2.34) (15.68) (22.65) Furnishings, household equipment and household services 5.60 3 1.12 6 6.53 6 3.32 7 Health 3.82 6 3.66 2 10.49 5 26.80 2 Out-of-pocket (2.04) (2.92) (1.81) (5.98) Social benefits in kind received (4.67) (3.95) (8.68) (20.81) Transport 5.29 5 0.53 8 13.74 3 3.29 8 Communication 7.93 1 11.69 1 1.96 10 15.32 4 Recreation and culture 7.72 2 2.99 3 17.90 2 21.76 3 Education 0.99 12 -1.15 11 0.75 12 -1.52 11 Restaurants and hotels 1.78 9 1.88 5 3.25 9 7.36 5 Miscellaneous goods and services 5.60 3 0.66 7 13.38 4 3.90 6 Domestic final consumption expenditure of households 3.61 1.54 100.00 100.00 Notes: The average annual real growth rate of component X between year t1 and year t2 was calculated as [(X(t2)-X(t1))**(1/(t2-t1)) - 1]*100. The contribution of component X to the real growth of household consumption C was calculated as [X(t2) X(t1)]*100/[C(t2)-C(t1)]. Household consumption is the sum of the final consumption expenditure of households plus social benefits in kind received, and the entire amount of social benefits in kind received are included in "health." The figure for total consumption in this table do not exactly match the figure for household consumption in Tables 1-3 because the figures in this table refer to "domestic final consumption expenditure of households" whereas the figures in Table 1-3 refer to "final consumption expenditure of households," with the difference between the two being that the former exclude "direct purchases abroad by resident households" and "direct purchases in the domestic market by non-resident households" whereas the latter include both, and because the figures in Tables 1-3 include "transfers of individual non-market goods and services" from the government and from private non-profit institutions serving households whereas the figures in this table do not. The figures to the right of each figure indicate the rank of that figure. Source: The same as Table 1.
37
Table 5: The Composition of Household Consumption, 1980-2003 The Share of Each Component in Household Consumption (percent) 1980 1991 Purpose of Expenditure 2003 Food and non-alcoholic beverages 19.64 2 15.18 2 12.60 3 Alcoholic beverages and tobacco 3.97 9 3.02 10 2.46 11 Clothing and footwear 8.48 5 7.03 7 4.20 9 Housing, electricity, gas and water supply 20.14 1 20.47 1 22.09 1 Imputed service of owneroccupied dwellings (13.64) (14.30) (15.70) Furnishings, household equipment and household services 3.79 10 4.67 9 4.45 8 Health 9.80 3 10.02 5 12.84 2 Out-of-pocket (3.46) (2.93) (3.44) Social benefits in kind received (6.34) (7.10) (9.40) Transport 8.58 4 10.25 4 9.08 5 Communication 0.71 12 1.12 12 3.50 10 Recreation and culture 6.74 8 10.34 3 12.26 4 Education 3.13 11 2.36 11 1.71 12 Restaurants and hotels 7.24 7 5.95 8 6.19 7 Miscellaneous goods and services 7.78 6 9.59 6 8.63 6 Domestic final consumption expenditure of households 100.00 100.00 100.00 Note: See the notes to Table 4. Source: The same as Table 1.
38
Table 6: The Decline in the Growth Rate and the Contribution to the Decline in the Real Growth of Household Consumption of Each Component of Household Consumption, 19801991 vs. 1991-2003 Contribution to the Decline in the Real Change in the Growth Growth of Household Rate, 1980-91 vs. 1991- Consumption, 1980-91 2003 (percent) vs. 1991-2003 (percent) -1.23 4 15.91 4 -1.23 4 3.25 9 -4.58 9 27.04 3 -1.57 6 6.19 7 (-1.71) -4.48 -0.16 (0.88) (-0.72) -4.77 3.77 -4.73 -2.15 0.10 -4.94 -2.06 8 3 (4.04) 11.89 -16.74 (-5.17) (-11.57) 31.18 -20.32 11.46 4.55 -3.61 29.21 100.00 5 11
Purpose of Expenditure Food and non-alcoholic beverages Alcoholic beverages and tobacco Clothing and footwear Housing, electricity, gas and water supply Imputed service of owneroccupied dwellings Furnishings, household equipment and household services Health Out-of-pocket Social benefits in kind received Transport Communication Recreation and culture Education Restaurants and hotels Miscellaneous goods and services Domestic final consumption expenditure of households
11 1 10 7 2 12
1 12 6 8 10 2
Notes: The contribution of each component to the decline in the real growth of household consumption between 1980-91 and 1991-2003 was calculated as the ratio of the decline in the change in that component between 1980-91 and 1991-2003 to the decline in the change in household consumption between 1980-91 and 1991-2003 (in percent). Also see the notes to Table 4. Source: The same as Table 1.
39
Table 7: Trends in the Household Saving Rate, 1980-2003 Calendar Year 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Unadjusted Household Saving Rate 17.3 18.2 16.8 16.2 16.1 15.5 14.8 13.1 13.5 13.6 13.9 15.1 14.2 13.7 12.6 11.9 9.9 10.0 11.2 10.8 9.6 6.7 7.3 7.5 Adjusted Household Saving Rate 15.4 16.2 14.9 14.3 14.3 13.7 13.0 11.5 11.9 12.0 12.3 13.3 12.5 12.0 11.1 10.4 8.5 8.7 9.6 9.3 8.2 5.7 6.1 6.3
Notes: The household saving rate was calculated as the ratio of net household saving to the sum of net household disposable income and "changes in pension reserves in pension funds, receivable." The unadjusted rate does not include "social transfers in kind" in the denominator, whereas the adjusted rate does. Source: The same as Table 1.
40
Table 8: The Proportion of Respondents Who Are Saving for Each Motive, 1984-2005 Saving Motive Calendar Illness and Other Living Expenses during Old Year Unforeseen Emergencies Peace of Mind Age 1984 75.0 25.7 42.1 1985 77.2 26.4 42.5 1986 75.0 25.3 42.5 1987 76.4 26.1 46.1 1988 77.1 28.0 50.2 1989 80.5 28.7 51.5 1990 74.3 25.7 52.4 1991 73.3 23.7 50.5 1992 68.3 23.0 48.2 1993 70.9 23.5 50.1 1994 69.4 24.2 51.6 1995 71.2 25.2 52.9 1996 69.7 25.9 53.9 1997 69.1 24.9 53.2 1998 73.3 24.5 55.3 1999 71.9 27.5 56.7 2000 67.5 27.1 55.9 2001 69.4 26.2 58.6 2002 69.1 26.9 56.9 2003 73.3 25.4 60.4 2004 65.9 26.0 57.4 2005 66.8 25.3 58.7 Note: The figures show the proportion of respondents saving for each motive (in percent). Source: Kinyuu Kouhou Chuuou Iinkai (The Central Council for Financial Services Information), ed., Kakei no Kinyuu Shisan ni kansuru Yoron Chousa (Public Opinion Survey on Household Financial Assets and Liabilities) , 2005 edition (Tokyo: Kinyuu Kouhou Chuuou Iinkai (The Central Council for Financial Services Information), 2005), and earlier editions of the same.
41
Table 9: The Proportion of Respondents Who Are Worried about Old Age, 1984-2005 The Proportion of Under-60 The Proportion of Under-60 Respondents Who Are Worried about Calendar Respondents Who Are Worried Old Age Because Pensions and Year about Old Age (percent) Insurance Are Inadequate (percent) 1984 na 49.1 1985 na 60.6 1986 na 64.5 1987 na 63.5 1988 na 69.6 1989 na 71.0 1990 na 68.0 1991 na 60.8 1992 63.7 55.5 1993 62.0 59.5 1994 69.9 59.1 1995 71.6 56.9 1996 71.3 59.0 1997 78.8 63.1 1998 85.5 67.1 1999 84.1 66.9 2000 84.7 68.1 2001 84.3 66.5 2002 86.6 66.7 2003 87.9 72.2 2004 86.1 66.4 2005 84.4 68.1 Notes: The denominator in the right-hand column is the number of under-60 respondents who are worried about old age. "na" denotes "not available." Source: The same as Table 8.
42
Endnotes Social benefits in kind, the largest component of social transfers in kind, were
included in household consumption under the older 1968 System of National Accounts to which Japan adhered until 2000, and social transfers in kind are included in their entirety in actual household consumption (an alternate consumption concept) under the newer 1993 System of National Accounts to which Japan has adhered since 2000.
This conclusion ignores second- and higher-order effects. For example, the stagnation
of consumption might have induced firms to cut back on their investment spending. I am indebted to Keunkwan Ryu for this point.
Since social benefits in kind are available only in nominal terms, I converted them into
real terms using the price deflator for the health component of household consumption.
The data appear to show that income inequality increased in Japan during the 1990s, but
Ohtake (2005) argues that the apparent trend toward greater inequality is largely a statistical artifact (except in the case of young cohorts) arising from the aging of the population and the decline in average household size.
All of these figures denote the average annualized real rate of growth of household
disposable income (inclusive of social transfers in kind) deflated by the price deflator for actual household consumption (inclusive of social transfers in kind).
43
The average annualized real rate of decline of household wealth was calculated by Since
deflating household wealth by the price deflator for actual household consumption.
household wealth is evaluated at the end of the each calendar year, I calculated the end-of-year price deflator by averaging the price deflators for the fourth quarter of the current year and the first quarter of the following year.
Ando (2002) and Ando, et al. (2003) attribute the stagnation of household wealth in large
part to the low dividend-payout ratio of Japanese corporations, which is another way of saying that Japanese corporations overinvested in physical assets, at least during the bubble period. The high growth rates of private fixed investment during the 1980-91 period in Table 1 corroborate Andos contention.
The difference between the two rates is that the latter includes social transfers in kind
in the denominator whereas the former does not. See the second paragraph of section 2 for a definition of social transfers in kind. .
It is, of course, possible that households were not able to save more despite their desire
This is not surprising because Japans public pension system has been periodically
reformed to keep it solvent in the face of rapid population agingwith contribution rates being increased, benefit levels being reduced, and the pensionable age being increased over
44
Saito and Shiratsuka (2003a, 2003b) distinguish between precautionary saving (which
depends on the magnitude of risks) and saving as a waiting option (which depends on how long it takes for uncertainties about the future to be resolved). They find that
precautionary saving has been more important since the 1980s but that there is some evidence of saving as a waiting option in the 1990s.
Also see the papers in Blomstrom, et al. (2003), Callen and Ostrey (2003), Hamada, et
al. (2004), Harada and Iwata (2002), Iwata and Mikitani (2003), Mikitani and Posen (2000), and Saxonhouse and Stern (2004).
For example, as Bernanke (2000), Ito (2004), and Posen (1998) argue, the Bank of
Japan discount rate should not have been cut in February 1987 and should have been raised in August 1988 when the Federal Reserve and Bundesbank raised interest rates, monetary policy should have been eased more in the early 1990s, fiscal policy should not have been tightened in April 1997, the Japanese Government mismanaged the financial crisis of 1997-98, and the Bank of Japan should not have ended its zero interest rate policy in August 2000..
45