Policy Update 055

Does a Growth Strategy Based on an Agglomeration Economy Conflict with Reviving the Fertility Rate?

KONDO Keisuke
Fellow, RIETI

Introduction

The debate over local revitalization in Japan has heated up again recently. The government's 2014 Basic Policies for Economic and Fiscal Policy Management and Structural Reform lists maintaining a stable population of about 100 million people 50 years from now as a basic policy. The government has also set up an office under the Shinzo Abe administration to prepare for the establishment of a revitalization headquarters for the city, people, and creating employment, which will take steps to reinvigorate rural areas and overcome the problem of declining population. Driving this is a uniquely Japanese problem: a population that is rapidly declining as the fertility rate falls and the population ages.

Perhaps one development that greatly affected the government's judgment was a list of municipalities that are "in danger of vanishing," as announced by the Japan Policy Council several months earlier (Note 1). A similar discussion took place in July 2014 at the National Governors' Association meeting in Saga prefecture, where the participants adopted a "declaration of declining fertility state of emergency," which stated that declining fertility countermeasures should be addressed as a national issue.

Various government ministries and agencies, such as the Ministry of Internal Affairs and Communications and the Ministry of Land, Infrastructure, Transport and Tourism, have begun taking measures with the future in mind. For its part, the Ministry of Economy, Trade and Industry considers the existence of two economic zones: one global and one local. Based on this idea, it has begun taking measures toward the revitalization of local economies as well as the globalization of Japan's economy. One thing that is common to the policies of the various ministries and agencies is the idea of forming self-supporting and efficient economic activity by dispersing functions among different regions and at the same time consolidating those functions compactly within each region. In such an economic structure, businesses would take advantage of local agglomeration and concurrently depend on each other through networks. This is an area in which spatial economics specializes. This report will ask why there are different fertility rates in different regions, guided by the perspectives of "space" and "agglomeration." The report will then consider what kind of growth strategy and declining-fertility countermeasures should be taken (Note 2).

Where is the fertility rate low?

Population data shows the regions in which the total fertility rate (TFR) is low (Note 3). Figure 1 plots the TFR by municipality on a map of Japan. Broadly speaking, the TFR is higher in prefectures in the west and lower in prefectures in the east. The TFR is higher on average in Kyushu and Okinawa than in Hokkaido and Tohoku. Additionally, the TFR is lower in urban prefectures. Figure 2 plots the population density in Japan by municipality, showing the highest population densities in the major urban areas including the Big Three: Tokyo, Osaka, and Nagoya (Note 4). These two figures suggest that the greater the population density of a region is, the lower is its TFR, as shown in Figure 3.

Figure 1: TFR by Municipality (2008-2012)
Figure 1: TFR by municipality (2008-2012)
Note: Municipalities are categorized into six quantile levels.
Source: Figure created by the author based on 2008-2012 Specified Report of Vital Statistics by Public Health Center and Municipality.
*All of the ranges are half-open intervals that are closed on the right.
Figure 2: Population Density by Municipality
Figure 2: Population density by municipality
Note: Municipalities are categorized into six quantile levels. Calculations are of spatially-smoothed population densities. For details, see footnote 4.
Source: Figure created by author based on 2010 National Census.
*All of the ranges are half-open intervals that are closed on the right.
Figure 3: Relationship of TFR and Population Density by Municipality
Figure 3: Relationship of TFR and population density by municipality
Source: Figure created by the author; TFR based on 2008-2012 Specified Report of Vital Statistics by Public Health Center and Municipality; data for population density (population age 15 and up) are based on 2010 National Census.

Why is the fertility rate low in regions with high population densities?

So far we have seen only a simple correlation, but it strongly suggests that there are factors in regions with high population densities that discourage having children. Let us consider the possibility of regional factors that might explain why the TFR is low in regions with high population densities. The Ministry of Education, Culture, Sports, Science and Technology reports in its Survey of Household Expenditure for Children's Education that the cost of non-school activities increases with the greater population of an area, as shown in Figure 4. Looking at the Regional Difference Index of Prices as per the National Survey of Prices of the Ministry of Internal Affairs and Communications, the supplementary education cost index is higher in urban areas, as shown in Figure 5. Finally, the National Fertility Survey of the National Institute of Population and Social Security Research reports that when it compared people's ideal number of children, the high cost of education was the main reason given for why that number of children was not reached. Thus, we can surmise that in more urban areas, costs such as children's education are higher, and as a result people tend to have fewer children in these areas (Note 5).

Figure 4: Cost of Non-school Activities of Students Attending Public School, by Population Size
Figure 4: Cost of non-school activities of students attending public school, by population size
Source: Figure created by author based on Ministry of Education, Culture, Sports, Science and Technology, Survey of Household Expenditure for Children's Education.
Figure 5: Population Density and Supplementary Education Cost Index by Municipality
Figure 5: Population density and supplementary education cost index by municipality
Source: Figure created by author; supplementary education cost index (national average = 100) based on Regional Difference Index of Prices of 2007 National Survey of Prices; population density (total population) based on the average of the 2005 and 2010 National Census.

The price effect in which rising child-rearing costs cause households to have fewer children was analyzed long ago by Becker (1960, 1992) (Note 6). Moreover, we can predict from Sato (2007), who incorporated the concepts of region and agglomeration, that population density creates geographical differences in child-rearing costs, which results in different fertility rates in different regions. Furthermore, the fact that there is a greater opportunity cost of raising children (that is, from quitting work to raise a child) in urban areas also could be a factor in the declining fertility (Note 7). Willis (1973) analyzed the effect of this opportunity cost on declining fertility. Only from the above, we might conclude that agglomeration has a great negative impact, but it also seems likely that, on the whole, the benefits of agglomeration outweigh the costs, and, as a result, people are moving into the cities.

Does agglomeration really lower fertility rates?

Let's take the above discussion a little further. How definitively can we say that the declining fertility in urban areas is a result of agglomeration increasing child-rearing costs (including the opportunity cost)? In fact, there are many other factors at work, so a variety of other hypotheses must be tested at the same time. The following are also possibilities:

  • People who live in cities may prefer simply not to have children after they marry.
  • Rural areas may place a greater emphasis on children as security for their parents in their old age, which may be the reason for their higher fertility rate compared to that of cities.
  • People who live in cities may delay marriage and parenthood because they are seeking higher education and employment, which may be why we observe declining fertility.

Thus, it is possible that the preferences and characteristics of households and the way households are geographically distributed may make population agglomeration appear to be a cause of declining fertility. If that is the case, we should be careful because it indicates that agglomeration itself is not a direct cause of low fertility rates.

Additionally, the endogenous choices made by households about the place of residence exaggerate the appearance of declining fertility in areas of high population agglomeration. For example, the following are also possibilities:

  • Households with many children change their place of residence from cities to rural areas, while families with few children remain in the cities.
  • A trend of households with few children moving from rural areas to cities because it is easier for them to change their place of residence.

Thus, we may observe lower fertility rates in areas of agglomeration because of actions taken by people after they have children. In that case, it is more accurate to say that agglomeration affects people's choice of where to live rather than directly curbing fertility behavior. Accordingly, it is possible that households that want children or already have them move out of areas of agglomeration, exaggerating the direct effect that agglomeration has in curbing fertility behavior (Note 8).

What I want to emphasize here is that we need to ask two questions: is there really a causality such that population agglomeration itself reduces the number of children, or do other factors make it appear as if population agglomeration causes declining fertility in urban areas? Understanding the cause correctly is important for appropriate policymaking. Thus, we need to confirm the robustness of our results, and, if we find that agglomeration really is a problem in terms of declining fertility, we have to be able to measure the extent of its impact. This is a pressing issue with which we researchers must deal for evidence-based policymaking (Note 9).

Regional policies based on an agglomeration economy could cause even lower fertility rates

In an aging and depopulating society with a low birth rate, there are constraints preventing it from increasing its labor force. Accordingly, how to increase the quality of the labor force and total factor productivity is critical to the revitalization of regional economies and their economic growth. So, how can productivity be increased? One approach that is attracting attention is agglomeration economies. Many books and discussion papers concerning agglomeration economies have been published by RIETI as well (Note 10). However, we should be keep in mind that a growth strategy based on an agglomeration economy could cause fertility to decline further in areas as noted above.

It is necessary to take a growth strategy based on an agglomeration economy, but it is also necessary to have countermeasures to mitigate the decline in the resulting fertility rates at the same time. As I emphasized earlier, any discussion of what specific policies to pursue requires that we find the correct cause of declining fertility in areas of agglomeration. Here I would like to discuss two factors and associated measures that might be considered as examples.

The first is regional differences in fertility rates as caused by different child-rearing costs and costs of living in different regions. Assuming identical incomes, areas with higher child-rearing costs and costs of living will have a lower potential demand for goods and services. In other words, even if people in different regions have equal nominal incomes, they will have unequal real incomes. Instead of paying the same benefits for children throughout the country, paying benefits that reflect regional differences in child-rearing costs and costs of living in line with the age of the child might be a better way to correct the disparity in fertility behavior. The second is the extent of the problem of children on childcare waiting lists in different regions. This is not just a simple problem of too few childcare centers. There is also the problem that in areas of high population agglomeration, the search and matching costs to find the right childcare center are rather high. If we assume that the result of this is a lower fertility rate, we should be focusing all the more on policies that enable households to start using childcare services quickly (Note 11). Such an approach would be characterized by regionally-based policymaking and seemingly would be effective in an agglomeration economy.

Naturally, any such policy should cover both households and enterprises and offer support for work-life balance and child-rearing (Note 12). It seems particularly necessary to take declining fertility countermeasures that respect a wide variety of life cycles. One big difference between urban and rural areas is the timing of childbirth. Figure 6 shows differences in childbirth timing by prefecture. It is particularly noteworthy that in households in the Tokyo metropolitan area, more women give birth between the ages of 35 and 44 than is the case in rural areas. In other words, the timing of childbirth—whether to have children at a younger or older age—is related to the size of the city. In that case, we cannot eliminate the possibility that marriage and childbirth later in life will happen even if urban cores are established in rural areas. It is important to offer assistance that respects the life cycle of each household so that the number of births will be closer to their ideal number of children (Note 13).

Figure 6: Fertility Rates by Prefecture and Age Class (in five-year increments) (per 1,000 women in population)
Figure 6: Fertility rates by prefecture and age class (in five-year increments) (per 1,000 women in population)
Note: Prefectures are categorized into six quantile levels.
Source: Created by author based on 2010 Specified Report of Vital Statistics, Table 27.
*All of the ranges are half-open intervals that are closed on the right.

Achieving both a growth strategy and countermeasures against declining fertility going forward

In this report, I discussed that a growth strategy based on an agglomeration economy can further lower the fertility rate in a region if it is not supplemented with policies to counter declining fertility. In policymaking, it is important to respect the diverse life cycles of households and consider how to prevent fertility rates from declining while simultaneously revitalizing local economies and promoting economic growth. When the population is falling, a debate that focuses simply on shifting population to rural areas may lead to an unproductive contest between municipalities to attract residents. We should avoid policy that leads to the prosperity of one region by stealing the riches of another.

Finally, I would like to give my personal opinion on the subject of municipalities that are "in danger of vanishing," a topic that has garnered much attention recently. What we must be careful to remember is that households and businesses do not exist for the sake of the survival of the municipalities in which they are located. Now is the time when we must consider how the municipal governments should be at the same time as these other issues. It is important to design appropriate institutions with a perspective considering the households and businesses that support local economies as policy targets.

The original text in Japanese was posted on August 15, 2014.

September 11, 2014

Acknowledgement:

I would like to express my gratitude to Masayuki Morikawa for his helpful comments. All remaining errors are solely mine.

Footnote(s)
  1. ^ The word "vanishing" as used here is somewhat misleading. In this case, it means that the population of women aged 20-39 is supposed to fall by 50% or more between 2010 and 2040. It is estimated that such an event will take place in 896 municipalities. These are called "municipalities in danger of vanishing."
  2. ^ The word "agglomeration" as used here means an agglomeration of population. It emphasizes the geographical distribution of population and the population size of individual cities. "Industrial agglomeration," on the other hand, is concerned with the distribution of industries across regions as well as the structure of industry in a region. Note that the purpose of this report is not to explain the "time-series" trend toward a declining fertility rate, but rather to focus on the "spatial" discrepancies in fertility rates. Part I, Chapter 2 of the Health, Labour and Welfare White Paper (Ministry of Health, Labour and Welfare, 2005) likewise emphasized the importance of having perspective of the community. However, it should also be noted that a separate discussion is necessary on the decline in the fertility rate observed in time-series data. Considering the issue of declining fertility from two different perspectives—time-series and spatial—is important for discussing the cause of such decline.
  3. ^ The total fertility rate (TFR) is a synthetic rate of the age-specific fertility rates for women aged 15-49. It is classified as either the total period fertility rate (TPFR) or total cohort fertility rate (TCFR). For details, see the following explanation by the Ministry of Health, Labour and Welfare.
    http://www.mhlw.go.jp/toukei/saikin/hw/jinkou/geppo/nengai11/sankou01.html (Retrieved on August 5, 2014, in Japanese)
  4. ^ Because population density is highly dependent on the relative size of the administrative district, the population density calculated here was spatially smoothed over a zone with a 30 km radius from the center of the municipality.
  5. ^ Naturally, the relative quality and selection of childcare services also has an effect. In areas of high population agglomeration, the cost of searching for a childcare center is higher, and, if one place is not a good match, the parents have to search again.
  6. ^ See Hotz et al. (1997) for an economic survey on household fertility behavior.
  7. ^ In fact, the relationship such that wages are higher in areas with higher population density was analyzed in Chapter 5 of Morikawa (2014) using worker individual data.
  8. ^ In reality, the workplace and home are not necessarily close to each other, commuting time should also have an effect on fertility behavior. Thus, it seems likely that the congestion cost caused by agglomeration would indirectly trigger a low fertility rate.
  9. ^ I am currently conducting an empirical study on this. A rough draft is scheduled to be published this fiscal year.
  10. ^ For example, refer to Morikawa (2014) and project results of the RIETI's Regional Economies program.
  11. ^ For households with the assumption more in terms of upper elementary school-aged children, benefits that are uniform nationwide may have a particularly big impact on their decision regarding number of children they will have. On the other hand, for households with the assumption more in terms of infants, improving the quality and quantity of child care services may have a greater impact on such decision. The above points require research going forward.
  12. ^ This point is also important in raising the average fertility rate in Japan as a whole over time.
  13. ^ For details on completed fertility, see Annual Population and Social Security Surveys conducted by the National Institute of Population and Social Security Research.
Reference(s)
  • Becker, Gary S. (1960) "An Economic Analysis of Fertility," in Universities-National Bureau ed. Demographic and Economic Change in Developed Countries, New York: Columbia University Press, pp. 209-240.
  • Becker, Gary S. (1992) "Fertility and the Economy," Journal of Population Economics 5(3), pp. 185-201.
  • Hotz, V. Joseph, Jacob Alex Klerman, and Robert J. Willis (1997) "The Economics of Fertility in Developed Countries," in Rosenzweig, Mark R. and Oded Stark eds. Handbook of Population and Family Economics Vol. 1A, Amsterdam: Elsevier, Chapter 7, pp. 275-347.
  • Sato, Yasuhiro (2007) "Economic Geography, Fertility and Migration," Journal of Urban Economics 61(2), pp. 372-387.
  • Willis, Robert J. (1973) "A New Approach to the Economic Theory of Fertility Behavior," Journal of Political Economy 81(2), pp. S14-S64.
  • Ministry of Health, Labour and Welfare (2005) Annual Health, Labour and Welfare Report 2004-2005 (in Japanese), Tokyo: Ministry of Health, Labour and Welfare.
  • Morikawa, Masayuki (2014) Productivity in Service Industries: Empirical analyses using microdata (in Japanese), Nippon Hyoronsha Co., Ltd.

September 11, 2014

Article(s) by this author