Next Article in Journal
Economic Valuation of Improving Environmental Degradations in Korea Using Choice Experiment
Previous Article in Journal
Achieving Learning Outcomes of Emergency Remote Learning to Sustain Higher Education during Crises: An Empirical Study of Malaysian Undergraduates
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Sustainability
Volume 14
Issue 3
10.3390/su14031595
sustainability-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Factors That Attract the Population: Empirical Research by Multiple Regression Analysis Using Data by Prefecture in Japan

by
Keisuke Kokubun
1,2
1
Smart-Aging Research Center, Tohoku University, Sendai 980-8575, Japan
2
Economic Research Institute, Japan Society for the Promotion of Machine Industry, Tokyo 105-0011, Japan
Sustainability 2022, 14(3), 1595; https://doi.org/10.3390/su14031595
Submission received: 26 December 2021 / Revised: 23 January 2022 / Accepted: 26 January 2022 / Published: 29 January 2022
Browse Figures
Versions Notes

Abstract

:
The development of the local economy and correcting the concentration in the capital city have long been the target for many countries. Furthermore, in the wake of the recent COVID-19 pandemic, the momentum for rural migration has been increasing to prevent the risk of infection with the help of the rise of remote work. However, there is not enough debate about what kind of land will attract the population. Therefore, in this paper, we performed correlation and multiple regression analyses, with the inflow rate and the net inflow rate of the population as the dependent variables, using the average values of government statistics for each prefecture in 2010 and 2017. As a result of the analyses, in addition to economic factor variables, variables of climatic, amenity, and human factors correlated with the inflow rate, and it was shown that the model had the greatest explanatory power when multiple factors had been used in addition to specific factors. It indicated that local prefectures were required to take regional promotion measures focusing not only on economic factors but also on multifaceted factors to attract the outside population. In addition, when the dependent variable was replaced with the 2020 population inflow rate, the model in which human factors were used as the independent variable showed the largest improvement in explanatory power. Therefore, it was shown that human factors have become more important in attracting people during the COVID-19 pandemic.

1. Introduction

Correcting population concentration in large cities is an important issue in many countries. Japan has a heavily concentrated population in the Tokyo metropolitan area, the capital and coastal city of Japan, and its problems have been discussed concerning efficiency, urban traffic congestion, housing shortages, environmental pollution, disaster risk, and rural depopulation [1,2]. Therefore, the government has aimed to correct the overconcentration in Tokyo and promote the development of local areas through financial support, employment creation, etc., by implementing the national comprehensive development plan. However, the effect is limited, and there have been no signs of improvement in recent years. In contrast to this situation, the recent COVID-19 pandemic has forced people living in cities to reassess the value of local prefectures with a low risk of infection. Moreover, the rise of remote work, which is a work style that can be performed anywhere, has made it possible for people to move to rural areas. These changes are expected to boost population migration from urban to rural areas.
If the relative attractiveness of Tokyo is diminishing, the COVID-19 pandemic may be an opportunity to correct the overconcentration in Tokyo and promote regional development. However, if economic factors are the only factors that influence migration, the changes in migration may also be limited, as rural areas with scarce economic resources have a limited ability to attract the population. In this regard, the solution could be what some previous studies have shown, that in addition to economic factors, human, amenity, and climatic factors all influence population migration. For example, if people can be attracted to migrate by utilizing human networks, there are expectations for regional development even in prefectures where the jobs of industries that can pay high salaries are few. However, as far as the author knows, there are no studies that have analyzed how much better population migration can be explained by combining these different factors as compared to the case where they act alone. To consider migration and its factors during and after the pandemic, it is necessary to correctly understand not only economic factors, but also the effects of various combined factors. Therefore, this paper will tackle this theme through the use of hierarchical multiple regression analysis, with the population inflow rate and net inflow rate as dependent variables and various factors as independent variables using the data set of 47 prefectures in Japan obtained from government statistics.

2. Examination of Previous Research and Presentation of Hypothesis

Recent studies have attributed migration to a variety of factors [3,4,5,6]. For example, Yu et al. showed that in addition to employment opportunities and wages, good social services, including facilities for education, recreation and commuting, warm winters, less-humid summers, clean urban environments, and friendly and open relationships are factors that influence Chinese migration within the country [5]. Likewise, Liu and Shen added the factors of social and cultural amenities, including the number of doctors, teachers, and museums; the area of green land, climatic conditions such as temperature, precipitation, and humidity; and the openness to employment opportunities such as interregional wage differentials, when analyzing the patterns of skilled migration [6]. These can be roughly divided into four categories: economic, human, amenity, and climatic factors. Therefore, in this study, these four factors are also regarded as factors that determine population migration and are used as independent variables.

2.1. Economic Factors

Economic factors are the most representative and indispensable element in classical economics. In support of this, many studies have revealed that local employment, income/wage differences, and costs of living account for interregional migration [7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24]. There are also studies that show how the concentration of the tertiary industry in the Tokyo area has been a major cause of population migration from rural areas to the Tokyo area since the 1990s [25,26]. Another study shows that from 1970 to 2000, prefectures with higher financial strength had less population outflow to Tokyo [27]. Relatedly, there is also a study showing that from 2005 to 2010, the higher the financial strength of municipalities, the higher their population growth rate [28]. From the abovementioned studies, the following hypotheses have been derived.
Hypothesis 1 (H1).
Economic factors correlate with the population inflow rate.
Hypothesis 1a (H1a).
Income and wage levels correlate with the population inflow rate.
Hypothesis 1b (H1b).
Accumulation of tertiary industry correlates with the population inflow rate.
Hypothesis 1c (H1c).
Financial strengthcorrelates with the population inflow rate.

2.2. Climatic Factors

Few studies have examined the effects of climate on migration. It has been pointed out that in the United States, migration from the high-income and cold northern part to the low-income and warm southern part became predominant after a certain period [29]. A study of Japan also showed that after the period of a high economic miracle, there was a gradual change of net inflow to warmer regions [14]. Furthermore, although the study covered a more limited area, the study by Tomioka and Sasaki [30] showed that there is a positive correlation between the minimum temperature and the population inflow rate. However, temperature is not the only climatic factor. Domestic studies have shown that the population is declining rapidly in snowy areas [31,32]. In addition, humidity is greatly related to the degree to which humans find the climate unpleasant [33]. Indeed, studies in China have shown that high humidity has a negative correlation with population influx [5,6]. Therefore, the following hypotheses have been derived.
Hypothesis 2 (H2).
Climatic factors correlate with population influx.
Hypothesis 2a (H2a).
Temperature correlates with population influx.
Hypothesis 2b (H2b).
Snowy days correlate with population influx.
Hypothesis 2c (H2c).
Relative humidity correlates with population influx.

2.3. Amenity Factors

It has been pointed out that the uneven distribution of amenities (living environment), including road infrastructure development, has caused the overconcentration of the population [1]. Another study showed that the number of hospital beds and the infrastructure development correlate with the influx of the population [34]. Similarly, a study of young people under the age of 35 from 1970 to 2000 showed that various social and environmental amenities influenced migration [14]. In China, social services, including facilities for education, recreation, and commuting, play an important role in attracting migrants [5]. Another study in China showed that road maintenance was largely associated with the development of surrounding cities [35]. In addition, studies in Japan and Europe have shown that area-specific amenities such as parks are one of the determinants of the population [36,37]. There is also a study showing the positive correlation between the spread of modern sewage systems and population influx [30,38]. From the abovementioned studies, the following hypotheses have been derived.
Hypothesis 3 (H3).
Amenity factors correlate with the population inflow rate.
Hypothesis 3a (H3a).
Medical facilities, such as hospitals and beds, correlate with population influx.
Hypothesis 3b (H3b).
Infrastructure, such as roads, correlates with population influx.
Hypothesis 3c (H3c).
Entertainment facilities, such as parks, correlate with population influx.
Hypothesis 3d (H3d).
Sanitary facilities, such as sewage facilities, correlate with population influx.

2.4. Human Factors

Population and population density have often been used as proxy variables for urbanization (e.g., [39]). Therefore, studies show that population and population density have a positive correlation with the influx of population [40,41]. However, some previous studies have dealt with a wider variety of human factors. Isoda [42] suggested that not only economic factors, but higher education, employment, and human networks are important factors that determine migrations. In line with this, there is a study on how the size of the production-age population and the penetration rate of higher education positively correlate with the population influx [41]. Relatedly, there is evidence that the college-educated population itself becomes an amenity that attracts other well-educated migrants [43,44,45]. Another study shows that the presence of some types of social capital (migration counseling and agricultural support centers) influenced the influx of the population [46]. From these studies, the following hypotheses have been derived.
Hypothesis 4 (H4).
Human factors correlate with the population inflow rate.
Hypothesis 4a (H4a).
Population and population density correlate with the population inflow rate.
Hypothesis 4b (H4b).
Production-age population correlates with the population inflow rate.
Hypothesis 4c (H4c).
College-educated population correlates with the population inflow rate.
Hypothesis 4d (H4d).
Social capital correlates with the population inflow rate.

2.5. Examining a Model That Includes Multiple Factors

Not many studies have incorporated multiple factors into the model. Most of these studies also show that economic factors influence population influx more strongly than other factors [17,47,48,49]. On the other hand, Tomioka and Sasaki [30] targeted 208 cities in the Tohoku and Kanto regions and conducted a multiple regression analysis using the net inflow rate from 1991 to 1995 as the explained variables, and wage income per capita, land prices, minimum temperature, sewerage penetration rate, as well as the number of beds and the number of students at universities and junior colleges as explanatory variables. As a result, it was shown that there was a significant correlation between the net inflow rate and the minimum temperature, wage income, and land price at the 5% level, and the sewerage penetration rate, the number of beds, and the number of university/junior college students at the 10% level. Similarly, a questionnaire survey conducted in Sweden showed that demographic, socio-economic, and geographical factors influence housing preferences [50]. In the same vein, a study using questionnaire response data and socio-economic data of knowledge workers in Israel found that they tended to choose land with a high population density, more areas used for cultural and educational purposes, and a richer knowledge community as a place of residence [51]. However, as far as the author knows, there are no studies that have verified how much more explanatory the analytical model that incorporates various variables is than the model that focuses only on economic factors. On the other hand, in research in the field of human resource management, a model that includes intrinsic factors and social factors is better than a model that includes only extrinsic factors (economic factors) in explaining the organizational commitments by hierarchical multiple regression analysis [52,53,54,55]. Therefore, the following hypothesis has been established.
Hypothesis 5.
Variables composed of various factors explain population inflow better than variables composed of specific factors.

3. Analytical Model

In western countries, the use of gravity models has been prevalent in population migration research since the 1960s [56]. The gravity model formulates the rule of thumb that the number of migrants between regions is proportional to the product of the population of the regions and inversely proportional to the distance between regions [57,58]. However, there are criticisms of the gravity model, which directly links migration and utility [46]. Moreover, it is argued that population size plays a large role in the gravity model, so it may not be suitable for research on economies such as Japan where the degree of concentration in the metropolitan area is strong [59]. In the author’s opinion, the gravity model focuses on elucidating the mechanism by which population migration occurs, but it seems that there is a lack of perspective on regional development as to what kind of land attracts the population. Here, the inflow rate and the net inflow rate can be calculated by the following formulas [30,40,60].
Inflow rate = (the number of in-migrants)/population, unit: %
Net inflow rate = (the number of in-migrants − the number of out-migrants)/population, unit: %
Therefore, this paper performs a multiple regression analysis, with the inflow rate and net inflow rate of the population as the dependent variables and various factor variables as the independent variables instead of using the gravity model. In this regard, the author would like to clarify which factors in Japan have attracted a bigger population in recent years. To this end, we first look at the simple correlation between the inflow rate and the net inflow rate of the population and each variable, and then perform a hierarchical multiple regression analysis, with the inflow rate and net inflow rate as the dependent variables.

4. Data

The population inflow rate is the average of the values for 2010 and 2017. The value was calculated by using the following information in the formula shown above: the number of Japanese in-migrants and Japanese out-migrants from other prefectures obtained from the “Basic Resident Register Population Migration Report” [61] of the Statistics Bureau of the Ministry of Internal Affairs and Communications, and the total population (including foreigners) as of 1 October obtained from the “Population Estimate” [62] of the Statistics Bureau of the Ministry of Internal Affairs and Communications. The reason for limiting the inflow population in the numerator to Japanese is in consideration of the possibility that the incentives for migration differ between the Japanese and foreigners. On the other hand, the reason why foreigners are included in the denominator population is that the author thought we should see how much the total local population can accept migrants from the viewpoint of regional development. Similarly, the author chose 1 October instead of the beginning of the year because the author thought it would be reasonable to measure capacity based on the representative population of the year. The author chose the two years of 2010 and 2017 because they are relatively recent, many variables are easily available, and because the researcher wanted to avoid the year that had a big impact on the domestic economy such as the Lehman shock in 2008 and the Great East Japan Earthquake in 2011.
For Gross domestic product and income per capita, the average values for 2010 and 2017 obtained from the Cabinet Office’s “Prefectural Accounts” [63] were used. For the unemployment rate, the current research used the average values for 2010 and 2017 obtained from the “Labor Force Survey” [64] of the Statistics Bureau of the Ministry of Internal Affairs and Communications. For the average land price of the residential area, the current research used the average values for 2011 and 2017 obtained from the “Prefectural Land Price Survey” [65] by the Ministry of Land, Infrastructure, Transport, and Tourism. The distance from the three major metropolitan areas (Tokyo, Aichi, Osaka) is the shortest (geodesic length) in the spheroid (GRS80) recorded in “Distances between prefectural offices” [66] of the Geographical Survey conducted by the Institute of the Ministry of Land, Infrastructure, Transport, and Tourism. For the number of operating kilometers (km) of railway lines per 100 square kilometers, the average values for 2009 and 2013 obtained from the “Annual Report of Regional Transportation” [67] of the Japan Transport Research Institute were used. The population density per 1 km2 of the habitable land area is the average value for 2010 and 2017, which was obtained by dividing the population data recorded in the “Population Estimate” [68] of the Statistics Bureau of the Ministry of Internal Affairs and Communications, by the habitable area obtained from the “Social Life Statistics Index” [69] of the Statistics Bureau of the Ministry of Internal Affairs and Communications.
For the ratio of people having completed up to college and university levels, since data for multiple years was not available, only the 2010 values obtained from the “Social and Demographic System” [70] of the Statistics Bureau of the Ministry of Internal Affairs and Communications were used. “Social capital” [71], which means trust, norms, and networks, was calculated and standardized by the Cabinet Office (2003) [72]. This value is based on the answers of 3878 people to the 10-question survey on “friendship/exchange”, “trust”, and “social participation” as well as the “volunteer activity rate” and “community chest per capita” by prefecture. For the sex ratio, average age, population ratio under 15 years old, population ratio between 15 and 64 years old, and population ratio over 65 years old, the average values for 2010 and 2015 obtained from the “National Census” [62] of the Statistics Bureau of the Ministry of Internal Affairs and Communications were used. For others, the average values for 2010 and 2017 (partly, the values of the years before and after) obtained from the “Social and Demographic System” [70] of the Statistics Bureau of the Ministry of Internal Affairs and Communications were used. The financial strength index is the average value for the past three years obtained by dividing the standard financial income by the standard financial demand. It can be said that the higher the financial strength index, the larger the reserved financial resources for calculating the ordinary tax allocation, and the more financial resources there are [73].
For the dependent variables, inflow rate and net inflow rate, the ones from 2020 were used alternatively to the above material [61]. This was to see the changes that the COVID-19 pandemic had provoked and how it affected population migration. However, for all independent variables, the ones shown above for 2010–2017 were commonly used. This is because at the time of writing of this paper, data from 2018 onwards, excluding migration, were not available. For reference, the above materials show that the number of over-migrants to the Tokyo metropolitan area in 2020 decreased by 33.3% as compared to the previous year, but it is still higher than in the 1990s, and there are nearly 100,000 over-migrants to that area.

5. Result

Table 1 shows the correlation of the variables with the population inflow rate and net inflow rate, in addition to the average value and standard deviation of each variable. From the top, economic factors, climatic factors, amenity factors, and human factors are shown. Let us look at them in order.

5.1. Economic Factors

A significant correlation was seen in both the inflow rate and the net inflow rate regarding gross domestic product per capita (r = 0.323, p < 0.05), prefectural income per person (r = 0.475, p < 0.01), household income (r = 0.396, p < 0.01), and cash salary (r = 0.561, p < 0.01) (the numbers in parentheses are the correlation coefficient with the population inflow rate and the p-value). On the other hand, the unemployment rate (r = 0.299, p < 0.05) was significantly correlated with the inflow rate, but not with the net inflow rate. Regarding consumer prices (r = 0.474, p <0.01) and the average land price in residential areas (r = 0.714, p <0.01), there was a positive correlation with the inflow rate and the net inflow rate. Strangely, people gathered in places where commodity and land prices were high, but it also made sense that the causal relationship was reversed; prices rose where people gathered.
Regarding the industrial composition, the percentage of primary industry workers (r = −0.494, p <0.01) and the percentage of secondary industry workers (r = −0.365, p <0.05) were negative, while the percentage of tertiary industry workers (r = 0.678, p < 0.01) was positive, and a significant correlation was found with the inflow rate and net inflow rate. It seems that the population inflow was less significant in prefectures where the employment scale of the tertiary industry was smaller, centered on the primary and secondary industries. As for the financial strength index (r = 0.602, p < 0.01), a significant correlation was shown for both the inflow rate and the net inflow rate, and it seems that people tended to gather in prefectures with high financial strength.
Distance from Tokyo, distance from Aichi, and distance from Osaka were variables for investigating the effect of distance from the three major metropolitan areas on the inflow rate. Only the distance from Tokyo showed a significant negative correlation with the net inflow rate. This result indicates that it was difficult for people to gather as the distance from Tokyo increased. Regarding the result that the distance from Aichi and Osaka did not become significant, it may be an indication of the current situation where the proximity to these prefectures is less likely to be an incentive for population influx as the concentration in Tokyo progresses.
From the above, correlations with the population inflow rate were found in multiple economic factor variables. This is in favor of H1.

5.2. Climatic Factors

The yearly average of air temperature and the lowest temperature show a positive correlation with the inflow rate, while the yearly rainy days show a negative correlation. Moreover, the yearly average relative humidity and yearly snowy days show a negative correlation with the inflow rate and net inflow rate. It seems that the land with higher (minimum) temperature, less precipitation/snowfall days, and lower humidity tended to attract more people. From the above, correlations with the population inflow rate were found in multiple climatic factor variables. This is in favor of H2.

5.3. Amenity Factors

The number of general hospitals, general clinics, and dental clinics, the number of beds in general hospitals/clinics and public parks (per 100 km2 of the inhabitable area), the total length of roadbed, total real length of roads, and total real length of major roads (per 1 km2), the ratio of major roads paved, and the diffusion rate of sewerage showed positive correlations with the inflow rate and net inflow rate. In addition, the ratio of local roads paved and policemen (per 1000 persons) showed a positive correlation with the inflow rate. People seem to gather in areas with abundant amenities such as hospitals, sewers, transportation, parks, and areas with many police officers. Furthermore, for persons killed or injured by traffic accidents and persons killed by traffic accidents (both per 100,000 persons), the former did not show a significant correlation with the inflow rate and net inflow rate, while the latter showed a negative correlation. This shows that it is difficult for people to gather in an area where large traffic accidents occur frequently, which cause death instead of injury. From the above, correlations with the population inflow rate were found in multiple amenity factor variables. This is in favor of H3.

5.4. Human Factors

The ratio of people having completed up to college and university, population, population density of the inhabitable area, population sex ratio (male per 100 females) showed a positive correlation with the inflow rate and net inflow rate, and average age showed a negative correlation with the inflow rate and net inflow rate. Regarding the age group, the population ratio 15–64 years old showed a positive correlation with inflow rate and net inflow rate, while the population ratio over 65 years old showed a negative correlation with the inflow rate. It seems that the more densely populated the land is and the more college graduates, working-age population, male, and young people there are, the more immigrants are attracted. From the above, correlations with the population inflow rate were found in multiple human factor variables. This is in favor of H4.
Social capital has no significant correlation with the inflow rate but shows a negative correlation with the net inflow rate. Social capital is a trust-based relationship among rural people, which can create unity and at the same time eliminate clinging and strangers [74,75]. Therefore, the negative correlation with the net inflow rate shown here may be due to the negative nature of social capital that has kept immigrants away. Alternatively, the causal relationship can be reversed. The results of several previous studies have shown that ethnic diversity brought about by population influx lowers social capital [76,77,78]. The influx of strangers with different values, even if they are ethnically the same, may make it difficult for residents to unite and lower their social capital.

5.5. Multiple Regression Analysis

Table 2 presents the results of the multiple regression analysis. The left half is the inflow rate, and the right half is the analysis results with the net inflow rate as the dependent variable. From the left, the variables for each category were input in the following order: economic factors (Step 1), climatic factors (Step 2), amenities (Step 3), and human factors (Step 4), and finally, all variables that became statistically significant from Step 1 to Step 4 were input (Step 5). Since there are many variables, the variable selection was performed by the stepwise method. Looking at the analysis results with the inflow rate as the dependent variable, as for the economic factors, the percentage of tertiary industry workers (β = 0.543, p < 0.01), and the financial strength index (β = 0.434, p < 0.01) were selected as significant variables, showing a positive correlation with the inflow rate. Even if the dependent variable was changed to the net inflow rate, the selected significant variables were the same. However, wage and the economic level were not chosen as significant variables in both inflow and net inflow models. They support H1b and H1c, but not H1a.
Next, as for the climate variables, yearly snowy days (β = −0.461, p < 0.01) showed a negative and significant correlation in a model with the inflow rate as the dependent variable. Instead, in a model with the net inflow rate as the dependent variable, the yearly average relative humidity (β = −0.377, p < 0.01) showed a negative and significant correlation. However, temperature was not chosen as a significant variable in both inflow and net inflow models. They support H2b and H2c, but not H2a. As for amenity variables, the ratio of major roads paved (β = 0.332, p < 0.01) and public parks (per 100 km2 of inhabitable area) (β = 0.635, p < 0.01) showed positive and significant correlations in a model with the inflow rate as the dependent variable. However, in a model with the net inflow rate as the dependent variable, in addition to the ratio of major roads paved, the total real length of roads (β = 0.521, p < 0.01) and the diffusion rate of sewerage (β = 0.374, p < 0.01) were selected as significant variables. Variables of medical facilities were not selected in both inflow and net inflow models. They support H3b, H3c, and H3d, but not H3a.
On the other hand, with regard to human factors, in addition to the population density of the inhabitable area (β = 0.352, p < 0.05), the population ratio of 15–64 years old (β = 0.368, p < 0.05), the ratio of people having completed up to college and university (β = 0.346, p < 0.05), and social capital (β = 0.339, p < 0.01) showed positive correlation in a model with inflow rate as the dependent variable. Social capital did not show a significant correlation by simple correlation, but the correlation was instead shown by controlling the influence of other variables in a multiple regression analysis. In other words, it seems that prefectures with a large population per land area, number of university graduates, and size of the working-age population, as well as those with abundant social capital tended to attract more people. The population density of the inhabitable area and the population ratio of 15–64 years old were positive and significant even in a model with the net inflow rate as the dependent variable. They support H4a, H4b, H4c, and H4d. Finally, when all the statistically significant variables in Steps 1–4 were input, the percentage of tertiary industry workers (β = 0.543, p < 0.01) and the financial strength index (β = 0.414, p < 0.01) of the economic factors, and the social capital (β = 0.352, p < 0.01) of the human factors were selected as significant variables in a model with inflow rate as the dependent variable (the ratio of people having completed up to college and university of the human factor was not statistically significant at 5%, although it was selected). However, in a model with the net inflow rate as the dependent variable, the ratio of major roads paved (β = 0.185, p < 0.05) and the total real length of roads (β = 0.229, p < 0.05) of the amenity factors, and the population density of the inhabitable area (β = 0.438, p < 0.01) and population ratio of 15–64 years old (β = 0.277, p < 0.01) of the human factors were selected as significant variables.
Comparing the adjusted R-squared, the model in which all the significant variables in Steps 1–4 were input was at 0.705, whereas the one for the economic factor only had 0.613, the climatic factor only 0.195, the amenity factor only 0.555, and the human factor only 0.622 in a model with the inflow rate as the dependent variable. That is, the adjusted R-squared was higher, in the range of 0.083–0.510, in the model where all variables were input as compared with the case where only variables of individual categories were used. According to Cohen’s (1988) [79] criteria, 0.02 is small, 0.13 is medium, and 0.26 is large, which means that the model in Step 5 has improved, from a small to a large extent, as compared to Steps 1–4. This is in favor of H5. Even in the case with the net inflow rate as the dependent variable, the model in which all variables were input was better in the adjusted R-squared than the model in which each factor was input individually in the range of 0.065–0.671, also from a small to a large extent.
Table 3 replaced the dependent variables’ inflow rate and net inflow rate with those from 2020. There was no big change in the composition of variables. However, there was a difference in the adjusted R-squared, and the biggest difference was in Step 4 of the inflow rate (adjusted R2 = 0.728). This indicates that in 2020, human factors had a greater influence on the inflow rate than in 2010–2017 (ΔR2 = 0.106). On the other hand, in the model with the net inflow as the dependent variable, the adjusted R-squared was lowered in every step in 2020. However, the adjusted R-squared of the human factors also decreased less than the adjusted R-squared of the economic factors, and an increase in the relative influence of human factors can be observed (adjusted R2 = 0.619 for Step 1 < adjusted R2 = 0.644 for Step 4).
Figure 1 and Figure 2 are P-P plots for testing the normality of regression residuals. Since the line draws on the countermeasure line, it can be said that the residuals roughly followed the normal distribution [80]. Figure 3 and Figure 4 are the maps of immigration and net immigration rates by the prefectures, respectively. It is shown that as a large population gathers in the metropolitan areas centered in Tokyo, the population diminishes in rural areas such as the northern (Tohoku) and western (Chugoku, Shikoku, and Kyushu) areas. Figure 5 is a scatter plot of the population inflow rate, with the predicted value calculated from the regression equation in Table 2 on the horizontal axis and the measured value on the vertical axis. This means that the prefectures above the diagonal have attracted more influx than expected, and the prefectures below the diagonal have attracted less influx than expected. It seems that such a difference was caused by another factor that was not included in the model estimated in this paper.

6. Discussion

We analyzed the factors in a prefecture to clarify which ones collect the population influx, using the population inflow rate and net inflow rate as the dependent variables, and various variables related to economic factors, climatic factors, amenity factors, and human factors as the independent variables. Furthermore, we analyzed the explanatory power of the model using various elements to clarify how much it is improved compared to the model using individual elements. First, as the result of a simple correlation analysis, it was shown that many variables have correlations with the inflow rate and the net inflow rate. Next, as the result of the hierarchical multiple regression analysis by the stepwise method, the inflow rate was more correlated with independent variables when all the four-factor variables were used as compared with the case where only the variables of a single factor, for example, the economic factor, were used. In other words, in addition to excellent economic factors such as abundant tertiary industry workers and the municipality’s financial strength, the results show that comfortable climatic factors, including few snowy days and low humidity; sufficient amenity factors such as well-arranged roads, public parks, and sewage; abundant human factors, including dense population, a large volume of the working-age population, college graduates, and rich social capital are all important for attracting the population. These results did not make a big difference when the dependent variable was changed from inflow rate to net inflow rate.
Since Hicks [81], many previous studies have focused on economic factors in the analysis of population movement. However, as this study shows, economic factors alone are not sufficient as a determinant of migration. In addition, hierarchical multiple regression analysis did not select some economic factors such as wage levels, which have been used in previous studies, as a result of stepwise variable selection. One possible cause is the difference in position between economic agents. For example, high wage levels are an incentive for workers, but a disadvantage for employers. In this way, there is a large difference in how economic factor variables are perceived depending on the migrant’s position, so it is considered that the difference in reactivity between the two was offset in the analysis of this paper by the migration of diverse people, and the correlation was reduced accordingly.
Another possibility is the difference between generations. Previous studies have shown that young people tended to move to higher-income prefectures, and older people tended to move to lower-income prefectures. The authors of the paper speculated that many of the latter were retired people, and that the emphasis was on livability rather than income [82]. Interestingly, the research also showed that both young and old people tended to move to prefectures with many young people [82]. This is consistent with the results of this paper, which show that human factors such as the working-age population have high explanatory power. Alternatively, the national character may be influencing the outcome. A finding of an international study suggests that Chinese students were oriented toward employment opportunities and economic well-being, while Japanese students were more inclined to consider personal lifestyle and local amenities [83].
It is also worth noting that the variables chosen as economic factors were local financial strength, not wages or added value per capita. This shows the importance of the proper financial steering of local governments. For example, it has been reported that as a result of subsidies to attract factories, only poor quality factories were attracted and did not become an incentive to attract the population [84]. Considering these cases and the results of this paper, it is suggested that it is more effective to attract people by eliminating waste and improving public finances from a long-term perspective, rather than providing temporary employment opportunities. Temperature from the climatic factors and medical facilities from the amenity factors were also not selected as independent variables. Regarding the former, it is consistent with the argument that the discomfort felt by humans is not determined by temperature alone but occurs when temperature and humidity overlap under certain conditions [33]. Similarly, the latter, as with other facilities, can be understood since it cannot be fully utilized if the transportation infrastructure to the facilities is poor. These results also show the importance of developing from a comprehensive perspective and not just one factor in order to attract people through regional development.
Climatic factors were not selected as statistically significant variables in the model in which all variables were input for both inflow rate and net inflow rate. This indicates that the impact of climatic factors on population inflow is smaller than that of other factors. However, this result is consistent, for example, with a previous study held in China where climatic factors such as temperature, precipitation, and humidity, and amenity factors such as the number of doctors had a lower impact on the decision of where to move to as compared to economic factors [6]. However, in the model using the inflow rate in 2020 as the dependent variable, yearly snowy days were selected as the statistically significant variable. This can be understood as a result suggesting that snowy areas still tend to be avoided even though the value of a rural residence has been reassessed by the COVID-19 pandemic.
Social capital showed a simple negative correlation with the net inflow rate. However, the multiple regression analysis controlling other variables showed a positive correlation with the inflow rate. How can this result be interpreted? Social capital creates trust and unity among residents and sometimes leads to the exclusion of strangers [74,75,85]. However, some studies have shown that a positive aspect of social capital, high trust, leads to a positive attitude towards influxes by increasing tolerance to uncertainty [86,87,88]. Moreover, in Japan, in addition to measures such as subsidies, it has been shown that measures that utilize the social capital of residents play an important role in promoting migration and settlement [89,90]. Therefore, the results of the multiple regression analysis would mean that social capital alone is weak and could sometimes be harmful as a factor to attract people, but in a land where other factors are prepared, it helps to attract people by making good use of it. Recent studies have shown that social capital is effective in efforts to prevent COVID-19 infection [91,92], so it can be said that it is becoming a factor that cannot be ignored when considering population movement during and after the COVID-19 pandemic.
Migration has often been thought to be primarily due to economic factors. However, as this study shows, economic factors alone are not sufficient for determining migration, and that the combination of climate, amenity, and human factors can better explain migration. In particular, the influence of human factors was as great as that of economic factors, and it was shown that their influence increased more and more in the COVID-19 pandemic. This shows that the rural area has the potential to attract and develop more people by nurturing and disseminating the resources of the region, including human resources.

7. Implication

Today, in response to the COVID-19 pandemic, it is expected that migration to rural areas, the correction of overconcentration in large cities, and regional promotion will progress. The results of this paper show what kind of factors have attracted the population to a prefecture in recent years, and they can be used as a reference in promoting the flow of migration to rural areas. In particular, the result showing that the correlation between the economic factor variables only and population inflow is smaller than the correlation between the variables of diverse elements, including climatic, amenity, and human factors in addition to economic factors and population inflow will be good news for those considering promotion measures in lacking areas. In particular, it will be important to consider how to improve the human factors of retaining university graduates of working age in the region, which has a large association with the inflow rate. At the same time, social capital, which was correlated with the inflow rate, is one of the few resources that these regions have, and therefore strategies such as utilizing it to support immigrants physically and mentally will be realistic and successful.

8. Limitation

The analysis results in this paper are based on a cross-sectional analysis and do not show a causal relationship. Additionally, since data for a specific year are used, there remains anxiety in terms of robustness. Furthermore, since we used prefectural data, we must also consider the impact of sample size on the analysis results. In the future, it would be preferable to verify the results of this paper by analyzing what kind of difference has occurred in the population inflow between prefectures in order to identify those that have taken specific measures and those that have not, and what kind of correlation there is between variables depending on the time of analysis. Moreover, it seems possible to supplement the limitation of this paper by an analysis based on small units such as cities and towns.

9. Conclusions

Regional promotion and centralized correction in Tokyo have long been the goals of the Government of Japan. Furthermore, since the wake of the COVID-19 pandemic, the momentum for rural migration has been increasing in order to prevent the risk of infection, with the help of the rise of remote work. However, there is not enough debate about what kind of land will attract the population. Therefore, in this paper, we considered this problem by performing a correlation analysis and a multiple regression analysis, with the inflow rate and the net inflow rate of the population as the dependent variables, while using recent government statistics for each prefecture. As a result of the analysis, in addition to the economic factor variables, variables of the climatic, amenity, and human factors correlated with the inflow rate, and it was shown that the model has the greatest explanatory power when multiple factors were used in addition to specific factors. In particular, human factors had a great influence on population inflow along with economic factors, and it was shown that they were increasing under COVID-19. Therefore, local prefectures are required to take regional promotion measures that focus not only on economic factors but also on multifaceted factors to attract the outside population.

Funding

This research received no external funding.

Institutional Review Board Statement

Ethical review and approval were waived for this study because we used anonymous information that is open to the public.

Informed Consent Statement

Patient consent was waived because we used anonymous information that is open to the public.

Data Availability Statement

Publicly available datasets were analyzed in this study (available upon request).

Acknowledgments

An earlier version of this article is published on the following preprint server: https://arxiv.org/abs/2009.07144 (accessed on 26 December 2021).

Conflicts of Interest

The author declares no conflict of interest.

References

  1. Hirata, S.; Kawabata, Y.; Fujii, S. A Study on the effect of investment in road infrastructures on the population concentration into Tokyo. J. Jpn. Soc. Civil Eng. 2019, 75, 967–978. [Google Scholar] [CrossRef]
  2. Porter, M. The Competitive Advantage of Nations; Free Press: New York, NY, USA, 1990. [Google Scholar]
  3. Buch, T.; Hamann, S.; Niebuhr, N. What makes cities attractive? The determinants of urban labor migration in Germany. Urban Stud. 2014, 51, 1960–1978. [Google Scholar] [CrossRef]
  4. Clark, W.A. What matters for internal migration, jobs or amenities? Migr. Lett. 2014, 11, 377–386. [Google Scholar]
  5. Yu, Z.; Zhang, H.; Tao, Z.; Liang, J. Amenities, economic opportunities and patterns of migration at the city level in China. Asian Pac. Migr. J. 2019, 28, 3–27. [Google Scholar] [CrossRef]
  6. Liu, Y.; Shen, J. Spatial patterns and determinants of skilled internal migration in China, 2000–2005. Pap. Reg. Sci. 2014, 93, 749–771. [Google Scholar] [CrossRef]
  7. Arntz, M. What attracts human capital? Understanding the skill composition of interregional job matches in Germany. Reg. Stud. 2010, 44, 423–441. [Google Scholar] [CrossRef]
  8. Asada, Y. Interregional Migration in the Postwar Japan: Economic Analysis Using Regional Income Data; Departmental Bulletin Paper; Osaka Prefecture University: Osaka, Japan, 1996; Volume 41, pp. 91–125. [Google Scholar] [CrossRef]
  9. Attanasio, O.; Padoa-Schioppa, F. Regional Inequalities, Migration and Mismatch in Italy, 1960–1986. In Mismatch and Labour Mobility; Padoa-Schioppa, F., Ed.; Cambridge University Press: Cambridge, UK, 1991. [Google Scholar]
  10. Faggian, A.; McCann, P.; Sheppard, S. Human capital, higher education and graduate migration: An analysis of Scottish and Welsh students. Urban Stud. 2007, 44, 2511–2528. [Google Scholar] [CrossRef]
  11. Ferguson, M.; Ali, K.; Olfert, M.R.; Partridge, M. Voting with their feet: Jobs versus amenities. Growth Chang. 2007, 38, 77–110. [Google Scholar] [CrossRef]
  12. Herzog, H.W.; Schlottmann, A.M. The metro rating game: What can be learned from the recent migrants? Growth Chang. 1986, 17, 37–50. [Google Scholar] [CrossRef]
  13. Higuchi, Y. Nihon Keizai to Shugyo Kodo (Japanese Economy and Job Search Behavior); Toyo Keizai: Tokyo, Japan, 1991. (In Japanese) [Google Scholar]
  14. Ito, K. An analysis of income growth effect on the long-distance migration in postwar Japan. J. Popul. Stud. 2006, 38, 89–98. [Google Scholar] [CrossRef]
  15. Lee, Y. Economic gains youth receive from inter- regional migration. In Tokyoni Deru Wakamonotachi (The Brain Drain: Why Japanese Youth Move to Tokyo), Ishiguro, I.; Lee, Y.J., Sugiura, H., Yamaguti, K., Eds.; Minerva Shobo: Tokyo, Japan, 2012; pp. 47–90. (In Japanese) [Google Scholar]
  16. Ohta, S.; Ohkusa, Y. Regional Labor Mobility and Wage Curve in Japan. JCER Econ. J. 1996, 32, 111–132. [Google Scholar]
  17. Scott, A.J. Jobs or amenities? Destination choices of migrant engineers in the USA. Pap. Reg. Sci. 2010, 89, 43–63. [Google Scholar] [CrossRef]
  18. Storper, M.; Scott, A.J. Rethinking human capital, creativity and urban growth. J. Econ. Geogr. 2009, 9, 147–167. [Google Scholar] [CrossRef] [Green Version]
  19. Tachi, M. Shotoku no chiiki bunseki to kokunai jinkō idō: Demogurafi no kenchi kara, Guranto shohan hakkō san hyaku-nen o kinen shite (Regional analysis of income and domestic migration: From a demographic point of view, to commemorate the 300th anniversary of the publication of Grant’s first edition). Hitotsubashi Univ. Res. Series. Econ. 1963, 7, 179–246. (In Japanese) [Google Scholar] [CrossRef]
  20. Tanioka, K. Study on regional income disparity and population migration. J. Reg. Soc. 2001, 4, 58. Available online: https://core.ac.uk/download/pdf/233904962.pdf (accessed on 26 December 2021). (In Japanese).
  21. Toyoda, T. Changes in regional income inequality and migration in Japan: Using estimated household income adjusted for household size and age compositions. Ann. Assoc Econ Geogr. 2013, 59, 4–26. [Google Scholar] [CrossRef]
  22. Vakulenko, E.S. Econometric analysis of factors of internal migration in Russia. Reg. Res. Russ. 2016, 6, 344–356. [Google Scholar] [CrossRef]
  23. Watanabe, M. Chiiki Keizai to Jinkō (Regional Economy and Population); Nippon Hyoron Sha: Tokyo, Japan, 1994. [Google Scholar]
  24. Zhou, J.; Hui, E.C.M. Housing prices, migration, and self-selection of migrants in China. Habitat Int. 2022, 119, 102479. [Google Scholar] [CrossRef]
  25. Kondo, S. How has service economy and de-industrialization affected the overconcentration of Japan’s labor force and population in Tokyo and other large city areas? Reg. Anal. 2020, 58, 45–65. [Google Scholar]
  26. Watanabe, M. Internal migration and regional economic differentials in postwar Japan. J. Popul. Stud. 1989, 12, 11–24. [Google Scholar] [CrossRef]
  27. Mitsuta, N.; Goto, K.; Shishido, S. A study on demographic shift from regional areas to large metropolitan areas. Stud. Reg. Sci. 2011, 41, 705–719. [Google Scholar] [CrossRef]
  28. Morikawa, H. Seeking revitalization of declining small and medium-sized cities in non-metropolitan regions of Japan: A comparison with German cities. Geogr. Sci. 2016, 71, 1–18. [Google Scholar] [CrossRef]
  29. Graves, P.E. Migration and climate. J. Reg. Sci. 1980, 20, 227–237. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  30. Tomioka, T.; Sasaki, K. Jinkō idō o kōryo shita toshi it sui no it su-teki hyōka (Economic evaluation of urban amenities considering population migration). J. Appl. Reg. Sci. 2003, 8, 33–44. [Google Scholar]
  31. Goto, S.; Fukai, T. A population analysis of the specific district in Japan. Bull. Aichi Inst. Technol. 1997, 32, 67–76. [Google Scholar]
  32. Ye, K.R.; Kiuchi, N.; Kozuka, K. About the direction of the winter life support in a Hilly, Mountainous and heavy snowfall area, based on a characteristic of space and member of community. J. Constr. Manag. 2007, 14, 299–310. [Google Scholar] [CrossRef]
  33. Thom, E.C. The discomfort index. Weatherwise 1959, 12, 57–60. [Google Scholar] [CrossRef]
  34. Etzo, I. Determinants of interregional migration in Italy: A panel data analysis. SSRN 2008. [Google Scholar] [CrossRef] [Green Version]
  35. Baum-Snow, N.; Brandt, L.; Henderson, J.V.; Turner, M.A.; Zhang, Q. Roads, railroads, and decentralization of Chinese cities. Rev. Econ. Stat. 2017, 99, 435–448. [Google Scholar] [CrossRef]
  36. Zhang, J.; Seya, H.; Kaneshige, H.; Chikaraishi, M. Longitudinal analysis of factors affecting inter-prefecture population mobility based on a discrete choice model with spatial context dependency. Geogr. Sci. 2016, 71, 118–132. [Google Scholar] [CrossRef]
  37. Rodríguez-Pose, A.; Ketterer, T.D. Do local amenities affect the appeal of regions in Europe for migrants? J. Reg. Sci. 2012, 52, 535–561. [Google Scholar] [CrossRef]
  38. Hayashi, N.; Saito, S.; Takahashi, T. Migration for the young and improvement in the infrastructure in rural areas of Kyoto prefectures: Mainly from 1990 to 2000. J. Rural Plan. Assoc. 2005, 24, 115–122. [Google Scholar] [CrossRef]
  39. Rees, P.; Bell, M.; Kupiszewski, M.; Kupiszewska, D.; Ueffing, P.; Bernard, A.; Charles-Edwards, E.; Stillwell, J. The impact of internal migration on population redistribution: An international comparison. Popul. Space Place 2017, 23, e2036. [Google Scholar] [CrossRef]
  40. Aoyama, Y.; Kondo, A. A Migration Model Based on the Difference in Utility between Regions. Infrastruct. Plan. Rev. 1992, 10, 151–158. [Google Scholar] [CrossRef]
  41. Palkama, J. The Determinants of Internal Migration in Finland. 2018. Available online: https://aaltodoc.aalto.fi/handle/123456789/35572 (accessed on 26 December 2021).
  42. Isoda, N. Higher education and population concentration into Tokyo metropolitan area in Japan. Fukuoka Univ. Review Lit. Humanit. 2009, 41, 1029–1052. [Google Scholar]
  43. Betz, M.R.; Partridge, M.D.; Fallah, B. Smart cities and attracting knowledge workers: Which cities attract highly-educated workers in the 21st century? Pap. Reg. Sci. 2016, 95, 819–841. [Google Scholar] [CrossRef]
  44. Gottlieb, P.D.; Joseph, G. College-to-work migration of technology graduates and holders of doctorates within the United States. J. Reg. Sci. 2006, 46, 627–659. [Google Scholar] [CrossRef]
  45. Waldorf, B.S. Brain Drain in Rural America. In Proceedings of the American Agricultural Economics Association Annual Meeting, Portland, OR, USA, 28–30 July 2007. [Google Scholar]
  46. Abe, S.; Kondo, A.; Kondo, A. Factors analysis of “UIJ-turn” migration and policies of population inflow. Infrastruct. Plan. Rev. 2010, 27, 219–230. [Google Scholar] [CrossRef] [Green Version]
  47. Clark, D.E.; Hunter, W.J. The impact of economic opportunity, amenities and fiscal factors on age-specific migration rates. J. Reg. Sci. 1992, 32, 349–365. [Google Scholar] [CrossRef]
  48. Chen, Y.; Rosenthal, S.S. Local amenities and life-cycle migration: Do people move for jobs or fun? J. Urban Econ. 2008, 64, 519–537. [Google Scholar] [CrossRef]
  49. Niedomysl, T.; Hansen, H.K. What matters more for the decision to move: Jobs versus amenities. Environ. Plan A 2010, 42, 1636–1649. [Google Scholar] [CrossRef]
  50. Niedomysl, T. Residential preferences for interregional migration in Sweden: Demographic, socioeconomic, and geographical determinants. Environ. Plan A 2008, 40, 1109–1131. [Google Scholar] [CrossRef]
  51. Frenkel, A.; Bendit, E.; Kaplan, S. Residential location choice of knowledge-workers: The role of amenities, workplace and lifestyle. Cities 2013, 35, 33–41. [Google Scholar] [CrossRef]
  52. Kokubun, K. Education, organizational commitment, and rewards within Japanese manufacturing companies in China. Empl. Relat. 2018, 40, 458–485. [Google Scholar] [CrossRef]
  53. Kokubun, K. Organizational commitment, rewards and education in the Philippines. Int. J. Organ. Anal. 2019, 27, 1605–1630. [Google Scholar] [CrossRef]
  54. Kokubun, K.; Yasui, M. The difference and similarity of the organizational commitment–rewards relationship among ethnic groups within Japanese manufacturing companies in Malaysia. Int. J. Sociol. Soc. Policy 2020, 40, 1391–1421. [Google Scholar] [CrossRef]
  55. Kokubun, K.; Yasui, M. Gender differences in organizational commitment and rewards within Japanese manufacturing companies in China. Cross Cult. Strateg. Manag. 2020, 28, 501–529. [Google Scholar] [CrossRef]
  56. Greenwood, M.J.; Hunt, G.L. The early history of migration research. Int. Reg. Sci. Rev. 2003, 26, 3–37. [Google Scholar] [CrossRef]
  57. Haynes, K.; Fotheringhrum, A. Gravity and Spatial Interaction Models; Sage: Beverly Hills, CA, USA, 1984. [Google Scholar]
  58. Ravenstein, E.G. The laws of migration. J. R. Stat. Soc. 1885, 48, 167–235. [Google Scholar] [CrossRef]
  59. Ito, K. Kokunai chōkyori jinkō idō ni ataeru seikatsu suijun no eikyō it suite–shin kokumin seikatsu shihyō to 1990-nen kokusei chōsa shūkei kekka o riyō shite (Impact of living standards on domestic long-distance migration-using the New National Living Index and the results of the 1990 National Survey). Rev. Econ. Inf. Stud. 2004, 4, 662–692. (In Japanese) [Google Scholar]
  60. Ito, K. Sengonihon no jinkō idō ni taisuru shotokukakusa-setsu no setsumei-ryoku to kongo no kadai (Explanatory power of income disparity theory for postwar Japan’s migration and future issues). J. Reg. Soc. 2001, 4, 9–38. Available online: https://core.ac.uk/download/pdf/233904961.pdf (accessed on 26 December 2021). (In Japanese).
  61. Statistics Bureau, Ministry of Internal Affairs and Communications. Jūmin Kihon Daichō Jinkō Idō Hōkoku (Basic Resident Register Population Migration Report). Available online: https://www.stat.go.jp/data/idou/index2.html#kekka (accessed on 18 January 2022).
  62. Statistics Bureau, Ministry of Internal Affairs and Communications. Census. Available online: https://www.e-stat.go.jp/stat-search/files?page=1&layout=datalist&toukei=00200521&tstat=000001049104&cycle=0&tclass1=000001049105 (accessed on 6 September 2020).
  63. Cabinet Office. Prefectural Accounts. Available online: https://www.esri.cao.go.jp/jp/sna/data/data_list/kenmin/files/contents/main_h28.html (accessed on 18 January 2022).
  64. Statistics Bureau, Ministry of Internal Affairs and Communications. Labor Force Survey. Available online: https://www.e-stat.go.jp/stat-search/files?page=1&layout=datalist&toukei=00200531&tstat=000000110001&cycle=0&tclass1=000001011635&tclass2=000001011637 (accessed on 6 September 2020).
  65. Ministry of Land, Infrastructure, Transport, and Tourism. Todōfuken Chika Chōsa (Prefectural Land Price Survey). Available online: https://www.mlit.go.jp/totikensangyo/totikensangyo_fr4_000044.html (accessed on 6 September 2020). (In Japanese).
  66. Geographical Survey Institute, Ministry of Land, Infrastructure, Transport and Tourism. Todōfuken-Chō-Kan no Kyori (Distance between Prefectures). Available online: https://www.gsi.go.jp/KOKUJYOHO/kenchokan.html (accessed on 6 September 2020). (In Japanese).
  67. Japan Transport and Tourism Research Institute. Chiiki Kōtsū Nenpō (Regional Transportation Annual Report); Japan Transport and Tourism Research Institute: Tokyo, Japan, 2015. (In Japanese) [Google Scholar]
  68. Statistics Bureau, Ministry of Internal Affairs and Communications. Jinkō Suikei (Population Estimation). Available online: http://www.stat.go.jp/data/jinsui/index2.html (accessed on 6 September 2020).
  69. Statistics Bureau, Ministry of Internal Affairs and Communications. Social Life Statistical Index. Available online: https://www.stat.go.jp/data/shihyou/naiyou.html (accessed on 6 September 2020).
  70. Statistics Bureau, Ministry of Internal Affairs and Communications. Social/Demographic System. Available online: https://www.e-stat.go.jp/stat-search/files?page=1&layout=datalist&toukei=00200502&tstat=000001137306&cycle=0&tclass1=000001137307&result_page=1 (accessed on 6 September 2020).
  71. Putnam, R.D. Bowling Alone: The Collapse and Revival of American Community; Simon & Schuster: New York, NY, USA, 2000. [Google Scholar]
  72. Cabinet Office National Living Bureau. Sōsharu Kyapitaru: Yutakana Ningen Kankei to Shimin Katsudō no kō Junkan o Motomete (Social Capital: Seeking a Virtuous Cycle of Rich Relationships and Civic Activities). 2003. Available online: https://www.npo-homepage.go.jp/toukei/2009izen-chousa/2009izen-sonota/2002social-capital (accessed on 6 September 2020).
  73. Ministry of Internal Affairs and Communications. Explanation of Indicators. Available online: https://www.soumu.go.jp/main_content/000264701.pdf (accessed on 6 September 2020).
  74. Coleman, J.S. Social capital in the creation of human capital. Am. J. Sociol. 1988, 94, S95–S120. [Google Scholar] [CrossRef]
  75. Portes, A. Social capital: Its origins and applications in modern sociology. Annu. Rev. Sociol. 1998, 24, 1–24. [Google Scholar] [CrossRef] [Green Version]
  76. Alesina, A.; La Ferrara, E. Who trusts others? J. Public Econ. 2002, 85, 207–234. [Google Scholar] [CrossRef]
  77. Costa, D.L.; Kahn, M.E. Civic engagement and community heterogeneity: An economist’s perspective. Perspect Politics 2003, 1, 103–111. [Google Scholar] [CrossRef] [Green Version]
  78. Putnam, R.D. E pluribus unum: Diversity and community in the twenty-first century the 2006 Johan Skytte Prize Lecture. Scand. Polit. Stud. 2007, 30, 137–174. [Google Scholar] [CrossRef]
  79. Cohen, J. Statistical Power Analysis for the Behavioral Sciences, 2nd ed.; Erlbaum: Hillsdale, NJ, USA, 1988. [Google Scholar]
  80. Tranmer, M.; Murphy, J.; Elliot, M.; Pampaka, M. Multiple Linear Regression, 2nd ed.; Cathie Marsh Institute Working Paper 2020–01; Cathie Marsh Institute for Social Research: Manchester, UK, 2020; Available online: https://hummedia.manchester.ac.uk/institutes/cmist/archive-publications/working-papers/2020/2020-1-multiple-linear-regression.pdf (accessed on 18 January 2022).
  81. Hicks, J. The Theory of Wages; Macmillan: London, UK, 1932. [Google Scholar]
  82. Tamura, K.; Sakamoto, H. Nihon no todōfuken-kan jinkō idō no sedai-kan hikaku (Intergenerational comparison of Japan’s inter-prefectural migration). AGI Working Paper Ser. 2016, 2016–2017, 1–11. Available online: http://id.nii.ac.jp/1270/00000114/ (accessed on 26 December 2021). (In Japanese).
  83. He, Z.; Zhai, G.; Asami, Y.; Tsuchida, S. Migration intentions and their determinants: Comparison of college students in China and Japan. Asian Pac. Migr. J. 2016, 25, 62–84. [Google Scholar] [CrossRef]
  84. Okubo, T.; Tomiura, E. Industrial relocation policy, productivity and heterogeneous plants: Evidence from Japan. Reg. Sci. Urban Econ. 2012, 42, 230–239. [Google Scholar] [CrossRef] [Green Version]
  85. Takagishi, M.; Kiminami, L. I-Turn Promotion in Rural Areas: Case Study from Chichibu City, Saitama Prefecture. Bull. Fac. Agric. Niigata Univ. 2012, 65, 1–14. Available online: https://agriknowledge.affrc.go.jp/RN/2010834121.pdf (accessed on 26 December 2021).
  86. Economidou, C.; Karamanis, D.; Kechrinioti, A.; Xesfingi, S. The Role of Social Capital in Shaping Europeans’ Immigration Sentiments. IZA J. Dev. Migr. 2020, 11, 20200003. [Google Scholar] [CrossRef] [Green Version]
  87. Herreros, F.; Criado, H. Social trust, social capital and perceptions of immigration. Polit. Stud. 2009, 57, 337–355. [Google Scholar] [CrossRef]
  88. Rustenbach, E. Sources of negative attitudes toward immigrants in Europe: A multi-level analysis. Int. Migr. Rev. 2010, 44, 53–77. [Google Scholar] [CrossRef]
  89. Sakuno, H. The Increase of Migrants into Local Areas and Regional Correspondence: What does “Return to the Country” Mean for Local Areas? Ann. Jpn. Assoc. Geogr. 2016, 62, 324–345. Available online: https://www.jstage.jst.go.jp/article/jaeg/62/4/62_324/_pdf (accessed on 26 December 2021).
  90. Takeda, Y.; Kaga, A. A study of the policy at migration and settlement in regional hub cities and dweller’s characteristics. J. City Plann. Inst. Jpn. 2018, 53, 1153–1160. [Google Scholar] [CrossRef]
  91. Kokubun, K.; Yamakawa, Y. Social capital mediates the relationship between social distancing and COVID-19 prevalence in Japan. Inquiry. 2021, 58, 00469580211005189. [Google Scholar] [CrossRef]
  92. Kokubun, K.; Ino, Y.; Ishimura, K. Social and psychological resources moderate the relation between anxiety, fatigue, compliance, and turnover intention during the COVID-19 pandemic. Int. J. Workplace Health Manag 2022. ahead of print. [Google Scholar] [CrossRef]
Sustainability 14 01595 g001 550
Figure 1. Normal P-P plot of the regression of standardized residuals. Note(s): Dependent variable is the annual average population inflow rate, 2010–2017.
Figure 1. Normal P-P plot of the regression of standardized residuals. Note(s): Dependent variable is the annual average population inflow rate, 2010–2017.
Sustainability 14 01595 g001
Sustainability 14 01595 g002 550
Figure 2. Normal P-P plot of the regression of standardized residuals. Note(s): Dependent variable is the annual average population net inflow rate, 2010–2017.
Figure 2. Normal P-P plot of the regression of standardized residuals. Note(s): Dependent variable is the annual average population net inflow rate, 2010–2017.
Sustainability 14 01595 g002
Sustainability 14 01595 g003 550
Figure 3. The map of immigration flow rates.
Figure 3. The map of immigration flow rates.
Sustainability 14 01595 g003
Sustainability 14 01595 g004 550
Figure 4. The map of net migration flow rates.
Figure 4. The map of net migration flow rates.
Sustainability 14 01595 g004
Sustainability 14 01595 g005 550
Figure 5. Scatter plot of predicted and actual population inflow rates. Note(s): The red circle (●) is the Tokyo metropolitan area.
Figure 5. Scatter plot of predicted and actual population inflow rates. Note(s): The red circle (●) is the Tokyo metropolitan area.
Sustainability 14 01595 g005
Table
Table 1. Mean, standard deviation, and correlation with an inflow rate.
Table 1. Mean, standard deviation, and correlation with an inflow rate.
ItemUnitMeanSDAnnual Average Population Inflow Rate, 2010–2017Annual Average Population Net Inflow Rate, 2010–2017Annual Average Population Inflow Rate, 2020Annual Average Population Net Inflow Rate, 2020
Annual average population inflow rate, 2010–2017%1.5680.39010.732 **0.963 **0.650 **
Annual average population net inflow rate, 2010–2017%−0.1210.1740.732 **10.802 **0.902 **
Annual average population inflow rate, 2020%1.6740.4140.963 **0.802 **10.760 **
Annual average population net inflow rate, 2020%−0.1370.1790.650 **0.902 **0.760 **1
Economic factors
  Gross domestic product per capita1000 yen3803.125747.8330.323 *0.522 **0.370 *0.263
  Prefectural income per person1000 yen2810.404472.8410.475 **0.663 **0.526 **0.427 **
  Monthly wages and salaries of household head per household1000 yen405.68743.0860.396 **0.436 **0.419 **0.351 *
  Cash salary (1 month per person)1000 yen299.46831.7960.561 **0.746 **0.649 **0.607 **
  Unemployment rate%3.6050.6560.299 *0.2080.2710.349 *
  Regional Difference Index of Consumer Prices (All items, less imputed rent)Total = 10098.8761.6570.474 **0.493 **0.484 **0.423 **
  Average land price of residential area (per 1 m2)yen51,340.42653,438.7800.714 **0.751 **0.733 **0.634 **
  Percentage of primary industry workers%6.0173.329−0.494 **−0.703 **−0.594 **−0.636 **
  Percentage of secondary industry workers%25.6184.952−0.364 *−0.094−0.235−0.165
  Percentage of tertiary industry workers%68.3665.0910.677 **0.551 **0.617 **0.576 **
  Financial strength index-0.5030.1980.602 **0.799 **0.687 **0.702 **
Climatic factors
  The yearly average of air temperature°C15.6192.3070.373 **0.2460.360 *0.214
  Highest temperature among monthly averages of the highest daily temperatures°C32.8261.3740.2470.1030.2500.050
  Lowest temperature among monthly averages of the lowest daily temperatures°C0.8723.1620.353 *0.2520.327 *0.239
  The yearly average of relative humidity%69.1814.390−0.316 *−0.377 **−0.385 **−0.355 *
  Yearly sunshine hourshours1938.755212.7430.2850.330 *0.335 *0.259
  Yearly precipitationmm1748.000469.135−0.140−0.098−0.170−0.165
  Yearly clear daysdays24.22311.7980.354 *0.320 *0.406 **0.252
  Yearly rainy daysdays120.64930.122−0.426 **−0.264−0.421 **−0.227
  Yearly snowy daysdays31.69232.978−0.461 **−0.344 *−0.473 **−0.259
Amenity factors
  General hospitals (per 100 km2 of inhabitable area)hospitals7.9497.6930.633 **0.615 **0.628 **0.536 **
  General clinics (per 100 km2 of inhabitable area)clinics112.688160.2980.654 **0.654 **0.662 **0.564 **
  Dental clinics (per 100 km2 of inhabitable area)clinics75.317125.1620.674 **0.686 **0.679 **0.582 **
  Number of beds in general hospitals/clinics (per 100 km2 of inhabitable area)beds17.33617.1910.660 **0.647 **0.665 **0.584 **
  The diffusion rate of sewerage%65.51118.0010.367 *0.577 **0.437 **0.584 **
  Total length of roadbed (per 1 km2)km10.3439.2230.677 **0.706 **0.695 **0.658 **
  Total real length of roads (per 1 km2)km4.3732.3040.624 **0.735 **0.681 **0.704 **
  Total real length of major roads (per 1 km2)km0.6320.1800.525 **0.535 **0.533 **0.485 **
  The ratio of major roads paved%97.7451.9610.424 **0.391 **0.414 **0.393 **
  The ratio of local roads paved%81.6609.9020.332 *0.2240.2830.115
  Public parks (per 100 km2 of inhabitable area)parks105.617122.7400.683 **0.687 **0.700 **0.661 **
  Persons killed or injured by traffic accidents (per 100,000 persons)persons581.783237.6430.0840.0210.0840.022
  Persons killed by traffic accidents (per 100,000 persons)persons4.0941.168−0.545 **−0.514 **−0.543 **−0.576 **
  Police men (per 1000 persons)persons1.9090.3090.365 *0.2490.316 *0.113
  Distance from Tokyokm456.651322.065−0.132−0.331 *−0.240−0.264
  Distance from Aichikm368.130261.304−0.049−0.208−0.129−0.068
  Distance from Osakakm367.287262.428−0.068−0.109−0.0930.038
Human factors
  Population1000 persons2710.4892718.2120.608 **0.768 **0.650 **0.739 **
  Population density of the inhabitable areaPersons/km21364.2451750.7990.687 **0.709 **0.711 **0.648 **
  Average ageyears old46.5561.643−0.578 **−0.726 **−0.684 **−0.716 **
  Population ratio under 15 years old%13.0720.9910.0980.1460.1720.199
  Population ratio 15–64 years old %60.5112.2840.668 **0.815 **0.755 **0.787 **
  Population ratio over 65 years old%26.4172.687−0.604 **−0.747 **−0.705 **−0.742 **
  The ratio of people having completed up to college and university%14.7473.9080.691 **0.773 **0.735 **0.670 **
  Social capital-0.0000.621−0.211−0.425 **−0.248−0.415 **
  Population sex ratio (male per 100 females)persons93.1083.7500.354 *0.634 **0.517 **0.618 **
Note(s): n = 47; * Significance at the 5% level; ** Significance at the 1% level.
Table
Table 2. Results of hierarchical multiple regression analysis, 2010–2017.
Table 2. Results of hierarchical multiple regression analysis, 2010–2017.
VariableAnnual Average Population Inflow Rate, 2010–2017Annual Average Population Net Inflow Rate, 2010–2017
Step 1Step 2Step 3Step 4Step 5Step 1Step 2Step 3Step 4Step 5
Economic factors
  Percentage of tertiary industry workers0.543 ** 0.543 **0.337 **
  Financial strength index0.434 ** 0.414 **0.695 **
Climatic factors
  Yearly snowy days −0.461 **
  The yearly average of relative humidity −0.377 **
Amenity factors
  The ratio of major roads paved 0.332 ** 0.247 ** 0.185 *
  The total real length of roads 0.521 ** 0.229 *
  Public parks (per 100 km2 of inhabitable area) 0.635 **
  The diffusion rate of sewerage 0.374 **
Human factors
  Population density of the inhabitable area 0.352 * 0.537 **0.438 **
  Population ratio 15–64 years old 0.368 * 0.392 **0.277 **
  The ratio of people having completed up to college and university 0.346 *0.272
  Social capital 0.339 **0.352 **
  R20.6290.2130.5740.6550.7300.7410.1420.6950.7410.812
  Adjusted R20.6130.1950.5550.6220.7050.7290.1230.6740.7290.794
  F37.366 **12.168 *29.691 **19.940 **28.450 **62.790 **7.445 **32.640 **62.970 **45.264 **
Note(s): n = 47; * Significance at the 5% level; ** Significance at the 1% level.
Table
Table 3. Results of hierarchical multiple regression analysis, 2020.
Table 3. Results of hierarchical multiple regression analysis, 2020.
VariableAnnual Average Population Inflow Rate, 2020Annual Average Population Net Inflow Rate, 2020
Step 1Step 2Step 3Step 4Step 5Step 1Step 2Step 3Step 4Step 5
Economic factors
  Percentage of tertiary industry workers0.448 ** 0.1680.397 **
  Financial strength index0.549 ** 0.579 **
Climatic factors
  Yearly snowy days −0.473 ** −0.247 **
  The yearly average of relative humidity −0.355 *
Amenity factors
  The ratio of major roads paved 0.319 ** 0.262 ** 0.220 *
  The total real length of roads 0.477 ** 0.271 *
  Public parks (per 100 km2 of inhabitable area) 0.654 ** 0.324 **
  The diffusion rate of sewerage 0.397 **
Human factors
  Population 0.320 *
  Population density of the inhabitable area 0.288 *
  Population ratio 15–64 years old 0.504 **0.592 ** 0.538 **0.573 **
  The ratio of people having completed up to college and university 0.342 **
  Social capital 0.343 **0.360 **
  R20.6540.2240.5900.7510.8050.6360.1260.6710.6600.728
  Adjusted R20.6380.2070.5720.7280.7820.6190.1070.6480.6440.709
  F41.527 **12.999 **31.696 **31.724 **33.937 **38.392 **6.507 *29.258 **42.617 **38.428 **
Note(s): n = 47; * Significance at the 5% level; ** Significance at the 1% level.
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Kokubun, K. Factors That Attract the Population: Empirical Research by Multiple Regression Analysis Using Data by Prefecture in Japan. Sustainability 2022, 14, 1595. https://doi.org/10.3390/su14031595

AMA Style

Kokubun K. Factors That Attract the Population: Empirical Research by Multiple Regression Analysis Using Data by Prefecture in Japan. Sustainability. 2022; 14(3):1595. https://doi.org/10.3390/su14031595

Chicago/Turabian Style

Kokubun, Keisuke. 2022. "Factors That Attract the Population: Empirical Research by Multiple Regression Analysis Using Data by Prefecture in Japan" Sustainability 14, no. 3: 1595. https://doi.org/10.3390/su14031595

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop