Chapter 2. Income Inequality and Seriousness of Poverty in Japan

Transcription

1 Chapter 2 Income Inequality and Seriousness of Poverty in Japan Toshiaki Tachibanaki, Kyoto University 1. Introduction People in Japan seem to have become generally conscious of the rising level of inequality within this country's society. One indication of this is the way the terms kachigumi (winners) and makegumi (losers) have become part of everyday conversation. Inequality can of course be considered on different levels. The simplest way of viewing it is in terms of inequality of outcomes, as typified by disparities in income. Equality of opportunity is also an important concept. This has to do with whether everyone is given an equal chance, without bias, in terms of receiving an education, looking for a job, getting promoted, and taking on various challenges. In the following I would like to present some concrete and up-to-date information about the inequality of outcomes in Japan, that is, the gap between rich and poor, using data that have recently become available. Also, I present the situation in Japan for those who find themselves living in poverty. The discussion on poverty continues by investigating how it is serious, and who are suffering from poverty. Finally, the paper examines the effectiveness of various policy tools in order to reduce the number of poor people. 2. Already a Land of Inequality At the end of 2004, the Organization for Economic Cooperation and Development published the results of a survey of income distribution in its member countries. In Table 1 I present some of the results, showing the degree of inequality in income distribution in 20 of 6

2 the surveyed countries as measured by the Gini coefficient. The Gini coefficient is an index N which 0 represents perfect equality and 1 perfect inequality; in other words, the larger the number, the greater the inequality in income distribution, which is to say, the wider the gap between rich and poor. The results were calculated on the basis of household disposable income, adjusted for household size, after deducting taxes and adding transfer payments, such as pensions. Table 1: Income Distribution in Japan and Other Advanced Countries Compared Source: OECD (2004), Income Distribution and Poverty in OECD Countries in the Second Half of the 1990s. Note: An adjustment was made for household size by using an equivalent scale and data on disposable income. For our purposes, the most striking thing about Table 1 is the fact that Japan can be found among the nations with a relatively high degree of inequality in income distribution. If we divided the countries listed into three groups, with group 1 representing nations with a high degree of equality, group 2 those with an intermediate degree of equality, and group 3 those with a high degree of inequality, Japan would clearly fall into group 3. In other words, it has already joined the ranks of nations characterized by a large gap between rich and poor, alongside the United States, Portugal, Italy, New Zealand, and Britain. In an earlier work, I pointed out the growing inequality in Japan's income distribution during the 1980s and the first half of the 1990s. At the time, a number of criticisms were raised in response to this analysis, among them the assertion that, although inequality may indeed be progressing in Japan, it gives little cause for concern because, among the world's industrialized countries, Japan still ranks only a bit above the middle in terms of income inequality. Table 1 shows that this criticism is no longer valid. There are a number of factors contributing to the increasing inequality, including the 7

3 aging of the population, the shift in compensation systems toward merit-based and performance-based pay, the increase in the number of unemployed and low-paid employees as a result of prolonged economic stagnation, and the impact of the government's tax and social welfare policies. But the most important underlying factor is this countries embrace of American-style economic liberalism. The same OECD survey also compared the poverty rates of member countries. Table 2 presents the results for all 27 member countries using the most recent figures available. Measuring the poverty rate is not always easy, and international comparisons are particularly difficult. For one thing, the definition of poverty differs from one country to the next. In addition, one must find ways of compensating for differences in household size and overall national living standards. This OECD analysis defines the poverty level in each country as less than half the median income, with income adjusted for differences in household size. Table 2: Poverty Rates in OECD Countries (%) Source: OECD (2004), Income Distribution and Poverty in OECD Countries in the Second Half of the 1990s. Of course, this definition is adopted expressly for purposes of international comparison and is by no means ideal. Nonetheless, the shocking fact, as seen in Table 2, is that Japan has the fifth-highest poverty rate of all the OECD nations, with a figure of 15.3%, ranking after Mexico, the United States, Turkey, and Ireland. Of the top five countries, however, Turkey and Mexico should probably be excluded from any comparison with Japan, since they are still in the early industrializing phase of development. If we limit the comparison to highly developed countries, then Japan comes in third, after the United States and Ireland. It is also pertinent to note that Japan's poverty rate, 8

4 which stood at 8.1% in 1994, nearly doubled in less than a decade. The dramatic rise in poverty in Japan is manifested also in an increase in the number of people receiving "livelihood protection" assistance, or maybe called income support programme generally. Livelihood protection is a system under which the government uses public funds to help cover the living expenses of people who have very little or no income and are thus unable to support themselves. It is the last resort in terms of poverty relief. The attached Figure 1 shows the dramatic rise in the number of households and individuals receiving livelihood protection assistance over the past decade. The number of households receiving assistance has soared in the last several years in particular and now exceeds 1 million. This is a stunning development, considering that the number stood at only 600,000 households in And in terms of individuals receiving assistance, the figure has topped 1.4 million. This rapid increase in the number of people unable to support themselves is a clear indication that the problem of poverty in Japan is growing ever more serious. Figure 1: The Number of Income Support Recipients (i.e., Livelihood Protection Assistance) 9

5 One thing that needs to be emphasized here is that the number of people in poverty is nowhere near the same as the number of people receiving livelihood protection. Whereas people living in poverty are defined as those whose income falls below a certain level, people on livelihood protection are those who are in fact receiving public assistance. Simply put, not everyone in poverty is receiving livelihood protection, and in fact, the number of people in poverty is much larger than the number receiving livelihood protection. (Using the OECD's poverty rate of 15.3% and Japan's estimated population as of 2004 [ million], we get a figure of 19.5 million Japanese living in poverty.) Inasmuch as everyone living in poverty is by definition enduring economic hardship, it might be suggested that all such people should receive some form of assistance. However, the cost of providing livelihood assistance to all of them would place an enormous burden on public finances. In fact, well under a tenth of Japan's poor are receiving such assistance. Though the figures are somewhat old, a 1993 estimate found that the "catch rate - the percentage of poor people getting livelihood protection benefits - was just 6%. However, if we define poverty as a situation in which people lack the resources to maintain a minimum standard of living, it is difficult to know whether this figure is too low without first ascertaining the actual economic circumstances, state of health, and so forth of those who are not receiving aid. In fact, no country provides assistance to 100% of its poor; in the West, the catch rates range between 20% and 60%. Be that as it may, given the rapid rise in the number of people living in poverty in Japan, it is clear that finding some way of assisting the poor economically is becoming an important issue. 3. Elderly and Polarization of Youth We have seen that Japan has a very large number of people living in poverty. The next question is who these needy people are. To answer it, let us begin by examining the poverty rate by age group. Table 3 presents data on the poverty rate by age group and also indicates what share of the poor each group comprises. The most striking thing about these data is the high rate of poverty among the elderly. Among people aged the poverty rate is 19.5%, and among those 76 and older it is nearly 24%. This state of affairs is underscored by the fact that the 10

6 elderly account for a full 47% of those receiving livelihood protection assistance. In fact, it would be fair to say that the face of poverty in Japan is represented by the elderly. Poverty is especially common among elderly people living alone. Table 3: Poverty Rates by Age Class in Japan (%) Source: OECD (2004) It might be supposed that poverty is uncommon among the elderly, since they are not expected to be able to work and are therefore eligible for pensions and long-term care insurance benefits. In fact, however, a large number of elderly people either do not qualify for pensions or receive very low benefits. Some did not enroll in a public pension plan for whatever reason; others enrolled but paid only a small amount in contributions. If we look at other highly developed countries, there is little evidence that the aging of society has given rise to a swelling population of poor people. The difference is a well-developed pension system. In Japan, the aging of society has increased the number of poor and unquestionably contributed to the unequal distribution of income. Another significant feature of the statistics shown in Table 3 is the high poverty rate among the young: 14.3% among children 17 or younger and 16.6% - second only to the elderly - among young adults years of age. Since young people only rarely have to worry about health problems, the favors underlying their poverty must be assumed to be completely different from those that account for poverty among the elderly. The cause, in a word, is high unemployment stemming from Japan's long economic stagnation. For several years now, unemployment among young people has topped 10%. In addition to those who are out of work, moreover, there are now somewhere between 2 million and 4 million so-called freeters, people without permanent employment, who move from one temporary job to another. The average yearly income of these workers is around 1.4 million, and even among them the situation is polarized, with one large group making less than 1.1 million and 11

7 another making between 2 million and 1.5 million annually. Those in the former group can clearly be said to be living in poverty. Even young people with regular employment rarely earn a high salary, since seniority pay is still the rule, but they can be certain of a steady income. Thus, a characteristic feature of Japanese society today is a growing economic gap within this younger age group, between low-income freeters and the perennially unemployed on the one hand, and the permanently employed on the other. 4. The Plight of Single Mothers There are two other categories of people on public assistance. The first, consisting of the injured and disabled, accounts for 35.1% of the total, second after the elderly. The people in this category are naturally eligible for livelihood protection, being unable to work. With regard to the disabled, the ideal solution would be to provide work and a way to achieve independence, but the feasibility of this depends very much on the individual case. The second category is that of single mothers, who account for 8.8% of the total. These are women raising children on their own as a result of divorce or the death of their husbands, and, less commonly, single women with children and no acknowledged father. It is easy to explain why single-mother households are another typical face of poverty in Japan. To begin with, women generally have a harder time finding a job than men in Japan, and women who do work make less than men. This is because Japan has not developed women workers as a resource in the way that it has men. In addition, in many cases the children in these households are still very young, making it difficult for the mother to work outside the home. Moreover, a woman who suddenly decides to work after spending years as a full-time home- maker is unlikely to find a job that pays decent wages. 5. Estimation of Seriousness in Poverty The previous section showed how the Japan s degree of inequality and poverty rate could be evaluated internationally, and found that it is highly unequal in income distribution, and very high in poverty rate based on the international comparison. This section investigates 12

8 poverty in Japan in depth, by applying more fruitful data. The poverty rate, which was presented previously, is unable to indicate the seriousness (or lightness) of poverty. If the majority of poor people received extremely low levels of incomes, i.e., the distribution of income of poor people were extremely skewed to the very poor level, it would imply that poverty is highly serious. If the majority of poor people received incomes which were somewhat lower than the poverty line, it would imply that poverty is relatively light (or not so serious), although it does not deny that they are still poor. In other words, these two cases are concerned with the distribution of poor people, whether the majority is extremely poor or not. There are several measures, which are able to indicate such seriousness like poverty gap ratio, Watts index, FGT index, CHU index, etc. The poverty gap ratio is the multiple of the head-count ratio (i.e., poverty rate) and the income-gap. The latter is defined by the ratio of the average income of poor people over the poverty line income, which can indicate the degree of the discrepancy between the average income of poor people and the poverty line. The multiple of the income gap ratio with the head-count ratio shows the relative degree of the seriousness (or simply intensity) of poverty. The FGT index can tell the seriousness of poverty more sensitively to a decrease in incomes of poor people who are located in slightly lower levels than the poverty line, while both the Watts index and CHU index can tell the seriousness of poverty more sensitively to a decrease in incomes of extremely poor people. In other words, their income levels are much lower than the poverty line. Since the definition and discussion of these indices or measures are too technical and require many pages to explain, only the empirical result is presented and discussed. Table 4 shows the various measures of poverty index for 1993, 96, 99 and The data source is the Income Re-distribution Survey by the Ministry of Welfare and Labour. The income figure is given by the disposable income which deducted tax payments and social security contributions from initial incomes, and added various social security benefits. We adjust for the scale of a household size, by using the Yagi and Tachibanaki (1996) method. 13

9 Table 4: Various Poverty Measures for All Households Source: Ministry of Welfare and Labour, Income Redistribution Survey. Notes: Poverty line is defined by the 50% figure of the median income. Equivalent incomes are estimated on the basis of Yagi and Tachibanaki (1996). Source: Ministry of Welfare and Labour, Income Redistribution Survey. Notes: Poverty line is defined by the 40% figure of the median income. Equivalent incomes are estimated on the basis of Yagi and Tachibanaki (1996). The definition of poverty, i.e., the poverty line, is always controversial. We adopt the most internationally common method. It is defined by the 50 percent or lower income in comparison with the median income of all samples. The 40 percent of the median income is also used as an alternative in order to obtain a comparative feature between the 50 percent definition and the 40 percent definition. These two tables show several interesting features of poverty in Japan. First, the estimated head-count ratio (i.e., poverty rate) is 17.3% in 2002, the latest figure, for the 50 percent definition. This is somewhat higher than the figure provided by the OECD study, which was explained previously. There are two reasons for the difference. On the one hand, the observation period is different because this study uses the latest data. On the other hand, the estimation method for the equivalent scale is different between the two, although the two studies applied the same data set. Second, the poverty rate increased significantly from 1993 to 2002, by about 1.6% points. During this period the Japanese economy was in a serious recession. It is very likely that it has produced a larger number of poor people. The estimated poverty rate clearly suggests that poverty intensified its seriousness. 14

10 Third, nearly all measures such as poverty gap ratio, Watts, CHU and FGT indicate that the seriousness of poverty had increased significantly, except for the Watts measure from 1999 to Since these measures can tell the implication associated with the distribution of income among poor people, it is possible to conclude that poverty in Japan, regardless of their relative positions, i.e., extremely poor or lightly poor in comparison with the poverty line, increased its seriousness. In other words, both extremely poor people and lightly poor people became poorer and poorer. Fourth, Table 5 is presented to confirm the third point more visually, which is given by the estimated poverty measures based on the various indices under the condition that the poverty line is fixed at the 1996 level. Since the poverty line is fixed at the 1996 level, increases in the values of various poverty measures can be explained almost exclusively by the intrinsic increase in the number of poor people. The reason is as follows; if poverty line were adopted each year as given by Table 4, the estimation of poverty measures would be influenced by each year s value of poverty line. Therefore, fixing poverty line is a worthwhile attempt to obtain the real story of poverty measures. Table 5: Various Poverty Measures When Poverty Line Is Fixed at the 1996 Level Source: Ministry of Welfare and Labour, Income Redistribution Survey. Notes: Poverty line is defined by the 50% figure of the median income. Equivalent incomes are estimated on the basis of Yagi and Tachibanaki (1996). Table 5 shows a more alarming result with respect to poverty in Japan than Table 4 because increases in the absolute values of various poverty indices are larger than the result in Table 4. In particular, the estimated head-count ratio and Watts index are 20.5% and 11.4% respectively which are larger than 20% and 10%. These rates, namely 20.5% and 11.4%, are not socially acceptable. For example, the poverty rate of about one-fifth of total population is equivalent to the poverty rates among developing countries. Since Japan belongs to a group of industrialized, developed and democratic countries, the society has to recognize this fact 15

11 seriously, and make an effort to reduce the number of poor people. Fifth, the difference in the definition of poverty, namely 50 percent of the median income versus 40 percent, is fairly significant to differentiate the absolute values of various measures of poverty. For example, the head-count ratio based on the 50 percent definition is between 15.6% and 17.3%, while it is between 10.4% and 11.8% based on the 40 percent definition. The difference in the head-count ratio between the two definitions is about 5.0% points, which is fairly large. The same story appears also for the other measures of poverty, although the absolute differences in the values between the two definitions differ considerably from measure to measure. The fifth result above provides us with an important issue in the analysis and measurement of poverty. For example, there is a tradition of a controversy between the absolute concept and level of poverty, and the relative concept and level of poverty in many countries. The definition based on either 50 percent of the median income or 40 percent belongs to the relative concept of poverty. Even with the relative concept of poverty the choice of 50 percent or 40 percent is neither so scientific nor persuasive. We adopted 50 percent definition because most poverty studies, in particular international comparative studies, have applied this definition. It should be useful to apply the absolute definition of poverty in order to compare with the estimated result based on the relative definition. Tachibanaki and Urakawa (2006) investigated such a comparison comprehensively. Only the conclusion is described; a comparison between the relative definition of poverty and the absolute one does not provide us with a significantly different result with respect to the estimated poverty measures. Interested readers can refer to Tachibanaki and Urakawa (2006). Thus, it is feasible to conclude that the estimated poverty situation in Japan has worsened considerably during the past ten years, as shown by this paper. One method, which enables us to examine whether various poverty measures have worsened universally from time to time, is to apply the TIP curve (Three indices of poverty). The TIP curve is given both by the horizontal line, i.e., the cumulative rate of population arranged by equivalent household incomes from zero to right (higher incomes), and by the vertical line, i.e., the cumulative sum of poverty gaps per capita. The TIP curve holds the following properties. First, if the TIP curve approached to a horizontal line, the value at the horizontal line would be equivalent to the value of head-count 16

12 ratio, and the value at the vertical line would be equivalent to the value of poverty gap ratio. Second, the degree of concavity at the TIP curve can tell the intrinsic nature of inequality of intensity of poor people. In other words, if a TIP curve in a certain year dominated the curve in another year (i.e., stayed higher in a certain year than another year graphically speaking over all range of the horizontal line of the TIP curve), it would be possible to conclude that various poverty measures given by Tables 4, 5, and 6 satisfy the Zhang (2000) lemma. The Zhang lemma shows that if the TIP curve in one year dominated the curve in another year, it would imply that the degree of poverty would have increased universally and definitely. Thus, the estimation of the TIP curve in different years is useful to examine whether poverty has increased universally and definitely. Figure 2 presents the estimated three years (i.e., 1996, 99, 2002) TIP curves. The result suggests that the TIP curve in 2002 dominates the curve in 1996 over all range, implying that the degree of poverty had increased universally and definitely from 1996 to Japan s poverty intensified significantly during this period. The curve in 1992, however, and the curve in 2002 intersects, graphically speaking. Consequently, it is not feasible to describe that the degree of poverty had increased during this period universally and definitely. Since I take the former finding more seriously because it considers the longer period, I am inclined to propose that the degree of poverty had increased almost universally and definitely. 17

13 Figure 2: The TIP Curve and Three `I's of Poverty (Equivalent Disposable Income Figures) 18

14 The final examination, which used the Income Redistribution Survey, is the estimation of poverty for various demographic groups. The previous section speculated that the majority of poor people would be single mothers, youth and the elderly (in particular the elderly widows). This section, therefore, attempts to confirm whether or not the above speculation is true, by estimating various poverty measures for various demographic groups. Table 6 shows the estimated poverty measures based on various indices for various demographic groups and age classes. There are totally ten demographic groups. (1) a husband and a wife, called nuclear household, with more than 3 children, (2) nuclear household with two children, (3) nuclear household with one child, (4) nuclear household with no children, (5) single member household, (6) single mother household, (7) an elderly husband and an elderly wife, (8) and elderly single household, (9) nuclear household with children and older parents, (10) others. Table 6 suggests the following observations for particular demographic groups. First, the highest poverty rate, i.e., 64.4%, is observed by single mothers. The rate higher than 50% implies that the great majority of households with a mother and children are poor, and more importantly their seriousness of poverty is explained by very high values of both the Watts measure and CHU. The majority of these households appeared because of an increasing number of divorces in recent Japan. These single mothers find difficulty in obtaining jobs with reasonable wage payments, partly because they have not accumulated enough skills and partly because they have to spend their considerable time for child-caring. Although the share of single mothers poverty in terms of the relative size among total populations is low, i.e., 5.6%, it is concluded that the seriousness of their poverty is extremely high. Second, the secondarily serious group of poverty is observed for elderly single households because the poverty share is 21.6%, and the poverty rate is 46.0%. The latter rate, 46.0%, is considerably high because it implies that nearly half of elderly single households are poor. As shown previously, the great majority of these households are female widows who lost their husbands. The principal reason for their serious poverty arises from the fact that their pension payments are considerably low because of the lost public and enterprise pension benefits. 19

15 Table 6: Various Poverty Measures for Several Demographic Groups and Age Classes Notes: Elderly here implies that male ages are over 65, and female ages are over 60. The most important finding in this table regarding elderly single households is their second highest poverty gap ratio, Watts index and CHU index, implying that those elderly single households are extremely poor. This is confirmed by the relatively low estimated poverty measures for households with an elderly household and an elderly wife. In other words, if a husband and a wife are alive, their economic condition is not terribly bad. If a husband dies, however, the economic condition of the survived wife is seriously deteriorated. Third, the third highest rate, if we pay attention to the poverty gap ratio, is observed 20

16 by single member households 9.6%. The seriousness is seen also by the high rate of the head-account ratio, 21.8%. The majority of these people are youth. The figures for age class under 29 years old support this fact because the head-count ratio is high like 16.3%. It is widely recognized that the economic conditions of young people are very bad in views of the fact that the youth unemployment rate is quite high (i.e., about 10 percent), and a large number of young workers are employed as non-regular workers whose wage payments are considerable lower than those of regular employees. Fourth, the lowest poverty measures are occupied by married couples. Their economic conditions are not so bad unlike single member households of both the elderly and the youth. It is an interesting subject to inquire whether the difference between married couples and single member households with respect to their economic conditions is due largely to their marriage status or to other circumstances rather than marriage status. It should be true, nevertheless, that if there were several household members, they would able to commit intra-household members transfers within the family. In other words members can help each other, if one or two members are in an economic trouble. 6. Evaluation of Policies to Reduce Poverty 6.1 Income Support Programme (or Livelihood Protection Assistance) to Poor People The ultimate policy which intends to reduce the number of poor people is the governmental economic support programme as mentioned previously. It is important to evaluate the efficiency of this programme. For examples, is it reducing the number of poor people effectively? Is it supporting to unnecessary people who are not poor? Is it paying excessive amounts higher than the requirement to necessary people who are poor? These three aspects can tell the degree of the efficiency of the programme, and thus it is interesting to estimate the degree of the efficiency empirically. There are two types of efficiency when we evaluate the impact of the programme. Figure 3 presents these two different types of efficiency as shown by Atkinson (1995). Some explanations are provided in order to show the implication of this figure. 21

17 Figure 3: Two Criteria to Evaluate the Efficiency of Income Support Programmes source: Atkinson (1995) p.227. We consider the distribution of income from the lowest income earner to the highest income earner. The horizontal line shows the index of households that is the rank of their income in increasing order. Thus, the index is in [0,1]. Thus, it is the inverse of the cumulative distribution of population which was arranged from the lowest to the highest. The vertical line represents the income figure. OZ is the income figure which is given by the poverty line. Thus, OH is equal to the rate of poor people in total population. It is called the headcount ratio. society. y min shows the income figure who receives the lowest income in the The solid line, i.e., min y N, represents the cumulative distribution function of income. Since it is assumed that the density function of income distribution follows the uniform distribution, it produces the straight line like y min N. The government adopts the income support programme in order to reduce the number of poor people. By adopting such an economic support programme, the cumulative function of income distribution is shifted from the solid line to the dotted line. Figure 3 shows that the income figures of all people who are 22

18 in the range between min y and N (a break-even point) increase by such an economic support programme. Thus, the well-being of these people increases. It is noted that the society had the total amount of poverty represented by the area ZM y min. The income support programme reduces the total amount of poverty by the area A. The area B is the amount of excessive payments higher than the requirement (i.e., OZ) to poor people, and the area C is the amount of unnecessary payments to people who are not poor. The area D represents the total amount of poverty still remained even after the adoption of the income support programme. In other words, the programme was unsuccessful in reducing the number of poor people entirely, but ended up with the remaining poor people. There are two types of efficiency for the income support programme. The first is called horizontal efficiency which is represented by the ratio of A over (A+D). The second is called vertical efficiency, or poverty reduction efficiency, which is represented by the ratio of A over (A+B+C). The former implies the degree of the success in reducing the number of poor people, while the latter implies one minus the degree of ineffectiveness, or of redundancy in reducing the number of poor people. In other words, the higher the latter, the more effective the programme in reducing the degree of ineffectiveness of the programme. It is very likely that there is a trade-off between horizontal and vertical efficiency because raising the former accompanies reducing the latter, and vice versa. We want to examine whether this trade-off is observed in Japan. Table 7 shows the estimated result regarding the two types of efficiency. The statistical source is the Income Redistribution Survey by the Ministry of Welfare and Labour, both 1996 and Needless to say, the target policy is the income (economic) support programme for poor people, called livelihood protection system in Japan. The estimation was made by Tachibanaki and Urakawa (2006). Table 7: The Estimated Efficiencies for Income Support Programme (%) Two different criteria were adopted to define poverty. The first is due to 50 percent of the median income, which is quite common internationally and was adopted by the OECD as 23

19 shown previously. The second is due to 40 percent of the median income. This definition suggests the rate of poor people (i.e., the headcount ratio) is higher because the poverty line is lower than the 50 percent definition. The table shows the following empirical results regarding the degree of efficiency of the income support programme in Japan. First, the horizontal efficiency is quite low, say between the lowest 3.5 percent, and the highest 12.8 percent. It is slightly lower than 10 percent on average. This derives the following conclusion. The income support programme in Japan is quite ineffective in reducing the number of poor people. An extremely large number of people remained as poor even after the income support programme was executed. In other words, the government policy is unsuccessful in reducing the number of poor people. We can point out several reasons for this ineffectiveness such as the expectation of family supports rather than public supports, the severe means-test to be qualified for public support, and the feeling of shyness and stigma of possible candidates for the income support programme. Second, contrary to the first conclusion, the vertical efficiency is quite high, say around 90 percent. It is noted that we support the trade-off relationship between horizontal and vertical efficiency when two efficiencies are compared regarding the effect of the income support programme. The high vertical efficiency implies that the government is successful in minimizing the redundant payment to people who are not poor, or the excessive payment higher than the requirement to people who are poor. In other words, the government management policy regarding the effectiveness of the income support programme is quite efficient to minimize the redundant payment. This is explained largely by the very severe means-test for possible candidates. There are several examples of these severe means-tests. Possible candidates have to write down and report their severe economic conditions to the public authority. Also, their documents are excessively complicated. The second example is given by zero saving amount of possible candidates to be qualified for receiving the income support programme. The third example is given by the condition such that no children of possible candidates can attend senior high schools. Third, we pay attention to the difference in the definition of poverty, namely 50 percent or 40 percent of the median income. It is found that the degree of horizontal efficiency is higher in the 40 percent definition than the 50 percent definition, while the degree of 24

20 vertical efficiency is lower in the former than in the latter. The above result is normal if we understand the meanings of both horizontal and vertical efficiencies correctly. A more useful finding is that the difference between the 50 percent definition and the 40 percent definition is not so large. In other words, the difference in the poverty definition does not matter so much, but the intrinsic nature of the income support programme in Japan is responsible for the large difference in the horizontal efficiency and vertical efficiency. I do not deny the condition of the means-test because the government fund, which has been contributed by the general tax revenues by all people, is limited, and also we have to eliminate the free-rider problem of public policies. The problem in Japan is too severe condition to be qualified for the income support programme. The zero saving condition may be too severe because all people would like to keep some fund to prepare future uncertainty. The Japanese senior high school attendance rate is over 90 percent among the common age group of 15 years old currently. Under this circumstance nearly all parents would hope that their children go to senior high schools. These conditions are the examples of too severe means-test. Let us conclude the usefulness and efficiency of the income (economic) support programme to reduce the number of poor people. It is not so effective to reduce the number of poor people in Japan. In other words, there are a large number of poor people who should be supported economically by the government. It is quite effective, however, to minimize the redundant payment to non-poor people or even to poor people. 6.2 Public Pension Programme The method to estimate the degree of two types of efficiencies is very useful as shown by the example of the income (economic) support programme to poor people. Thus, another public policy is evaluated based on the same method. That is the effectiveness of public pension programme to reduce the degree of poverty. We apply the same estimation technique and the same data set, but to the different sample. Since we are interested in the effect of public pension programme, we restrict the sample within people whose age are 65 years old and over. Table 8 shows the estimated result of two types of efficiency, namely horizontal and vertical. The most impressive observation given by this table is that the degree of the horizontal efficiency is about 80 percent on average. This implies that the public pension 25

21 benefit is very effective to reduce the number of poor old people. If there had been no public pension programme, there would have been an extremely large number of poor old people. The public pension programme is crucially important to eliminate poverty of old people. Table 8: The Estimated Efficiencies for Public Pension Programme to Reduce Poverty (%) Two important sources used to be common to support income of retired people in the past. The first is economic support from their adult children. This kind of intergenerational economic transfers from children to aged parents was quite common. The second is self-support of retired people in the form of personal savings, retirement payments, etc. These two social and economic customs, which support income after retirement from the labour market, have eroded considerably for various reasons. The public pension programme was introduced in order to compensate such erosion. It has developed gradually and gradually during the past 30 years since the year 1973 which is called the beginning of the welfare system in Japan. The public pension programme played a crucial role to minimize the number of poor old people, and thus it worked quite well not to produce poverty among older people. It is noted, nevertheless, that we should emphasize the previous result such that the most important group of poverty is observed by older people. Is this contradictory to the result which is given by the efficient effect of the public pension programme? The answer is No. We have to stress about 20 percent figure of horizontal inefficiency of the public pension programme. The remaining 20 percent figures suggest that there is a non-negligible portion of older people whose public pension benefits are considerably low or who receive no public pension benefits, for various reasons. The degree of the vertical efficiency in Table 8 is about 30~40 percent which is somewhat lower than 50 percent. It is, however, much lower than the estimated vertical inefficiency of the income support programme to poor people. Therefore, it is concluded that some excessive pension benefits or some redundant pension benefits are paid to older pensioners. This is not a bad result because the purpose of the public pension system is not only to reduce the degree of poverty among older people, but also to support the decent level 26

22 of economic life of older people. Thus, it is not surprising to see the fact that the public pension programme pays benefits which are higher than the poverty line. This leads to produce about 30~40 percent of vertical efficiency of the public pension benefit. 7. Concluding Remark The paper showed the most recent situation of income inequality and poverty in Japan. An emphasis was placed on an international comparison, and it was concluded that the degree of income inequality, and in particular of poverty in Japan was one of the highest. The paper investigated the cause of poverty, and estimated how it became serious by showing various poverty measures quantitatively. It was concluded that poverty had worsened universally and definitely. Estimation of poverty degree was made for various demographic groups. The highest seriousness in poverty is observed for single mothers, the youth and the elderly. Finally, several policies which intend to reduce the number of poor people were evaluated. Public assistance payment to poor people is not sufficient at all, while public pension programme is effective to a certain extent to reduce the number of poor people. The result implies the following; first, it is necessary to improve the livelihood protection assistance (or income support programme) to poor people significantly. Second, if there had been no system of public pension, there would have been a larger number of poor people. 27