## wpiea2020200-print-pdf

## Source details

**Canonical URL:** [wpiea2020200-print-pdf](https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020200-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2020/english/wpiea2020200-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2020/english/wpiea2020200-print-pdf.pdf.json)

---

### I. Introduction — overview and key findings
- Japan is undergoing demographic transition with slowing population growth due to low fertility and increased life expectancy.
- Main empirical questions:
  - Pattern of prefecture-level population dynamics since 1975.
  - Evolution of residential house prices (residential land price as proxy) across Japan and by prefecture.
  - Importance and time variation of demographic drivers of house price dynamics.
  - Policy implications.
- Summary of principal findings:
  - Stark urban–rural demographic divergence: by 1998 thirty-four prefectures (out of forty-seven) had negative population growth; by 2017 forty prefectures had negative growth. Seven prefectures recorded positive population growth in 2017: Tokyo-to, Saitama, Chiba, Kanagawa, Osaka, Fukuoka, and Nagoya.
  - Residential house prices declined steadily after the early 1990s bubble burst; since 2013 (Abenomics) the downward trend reversed: prices stabilized with mild increases in some prefectures and sharp condominium price appreciations in some large cities and tourist destinations.
  - Positive correlation between population growth and house prices: "a 1 percent increase in population growth is associated with a 5 percentage points increase in house prices."
  - Nonlinear/asymmetric population–price relationship: population declines are associated with larger magnitudes of house price decreases than symmetric population increases produce in price gains (consistent with Glaeser and Gyourko (2005) durable housing model).
  - Time variation: in the last ten years the positive correlation between population increases and house price increases has weakened, while the correlation between negative population growth and house price declines remains.

### II. Demographic trends and migration (prefecture-level)
- Distributional shift and timing:
  - Kernel density of prefecture-level population change shifted left from 1975–1990 to 1990–2015.
  - 1975–1990: every prefecture except Akita and Nagasaki experienced positive population growth; average population growth around 9 percent; median 6.7 percent.
  - Since 1990: thirty prefectures out of forty-seven experienced population declines in 1990–2015; largest declines in Tohoku, Shikoku and Kyushu.
  - 1998: thirteen prefectures with net positive population growth.
- Role of internal migration:
  - Natural population change (births and deaths) does not fully explain prefecture-level dynamics; cross-prefecture net migration is important.
  - Net migration into four metropolitan areas since 1971: until 2007 all four (Tokyo, Osaka, Nagoya, Fukuoka) showed positive net migration; since 2007 only Tokyo-to and the Greater Tokyo Metropolitan area continue to show positive growth, with rates picking up since 2012.
  - Example: internal migration accounted for half of population changes due to natural decrease in Tohoku in 2017; in Tokyo, Kanagawa, Chiba, Saitama, Aichi, and Fukuoka natural increase was negative but net migration was positive and exceeded the negative natural change.
- Population concentration changes (1996–2015):
  - Tokyo concentration increased by 1.3 percent.
  - Hokkaido experienced the largest decline: -0.3 percent.
  - Only six prefectures recorded increased concentration: Tokyo, Kanagawa, Saitama, Hyogo, Shiga, and Fukuoka.
- Persistence/asymmetry regression results (following Glaeser and Gyourko (2005)):
  - POPLOSS = 0 if prefecture grew in period t; equals actual percentage decline if prefecture lost population. POPGAIN analogously defined for gains.
  - 1975–1990: a 1 percent population loss in the previous year associated with a 0.78 percent decline in the current year; population gains impact 0.71 percent (column (2) referenced).
  - Recent period (column (5) referenced): population gain persistence 0.77 versus population loss persistence 0.17 — concentration into gain prefectures is accelerating while losses in others are more gradual.

### III. Households and family composition
- Total households and projections:
  - Total households increased until 2016 and stabilized at around 53 million households.
  - Official projection: peak at "54.2 million, and start a gradual decline n to around 50.8 million households by 2040, which is slightly below the total household numbers in 2010."
- Vital statistics:
  - Number of births peaked at 2,091,983 in 1973 and fell thereafter.
  - FY2017 births: 946,065; deaths: 1,340,397.
- Structural household shifts:
  - Movement from three-generation households to nuclear families.
  - Among households with persons aged 65 years and over, three-generation share declined dramatically; one-person and couple-only households combined exceed 50 percent.
  - Average family size declined; decline larger in rural areas (-31.9 percent) versus large cities (-28.2 percent) between 1970 and 2015.
  - Household growth by area: major cities experienced an increase in household numbers of 115.6 percent compared to other areas at 18.3 percent (1970–2015).

### IV. Housing price developments — aggregate, regional, and by type
- Long-run and recent aggregate dynamics:
  - Residential land price (¥/㎡, public assessment) used as proxy for house prices.
  - Bubble period: since 1988 annual average increase of 20 percent; bubble burst in 1991 with sharp fall.
  - Post-bubble: steady downward trend with a "Mini bubble (2002-2008)."
  - Since Abenomics: prices increased by 2.7 percent since 2014 and 0.7 percent in 2018; current level recovered to the level in 2013.
  - Recent nationwide trend: average 1.9 percent nationwide increase since 2013 (with large cross-prefecture variation).
- Prefecture-level heterogeneity:
  - Cumulative price changes (2002–2018): all prefectures except Tokyo experienced declines (in the -5 to -15 percent range).
  - Since 2014: numerous prefectures experienced appreciations; magnitude varies.
  - Notable changes since 2013: Miyagi, Fukushima: +18 percent; Tokyo: +16 percent; Akita: -6.5 percent; Shimane: -5.5 percent; Yamanashi: -4.8 percent.
  - Hokkaido: commercial land price up by 60% in 2018; residential land prices also rose.
- By housing type (since 2013):
  - Condominium prices: increased by 23 percent at the national level since 2013.
  - Condominium prices in Tokyo: increased by 45 percent since 2013.
  - Detached houses: increase of 4.2 percent since 2013.
- Drivers of condominium appreciation:
  - Tax incentives (inheritance tax assessment treats land and dwellings separately).
  - Better-developed secondary market, ease of rental, investment expectations.
  - Foreign investment: total acquisition of commercial real estate by foreign investors ~¥680 billion in H2 2017.
  - Institutional investment: J-REIT transaction volume steadily increasing since 2014.

### V. Tokyo specifics — supply, demand, and market structure
- Tokyo experienced continued population inflows and is the only region where land house price increased despite the bubble crash.
- Supply-demand indicators: highest number of housing starts and low vacancy rates in Tokyo.
- Supply-side concentration:
  - Market share of major seven condominium developers in Tokyo rose from about 20 percent until 2007 to close to 50 percent in 2017, increasing price sensitivity to major developers’ supply decisions.
  - Reverse relationship found between condominium prices and supply of newly-built condominiums in Tokyo metropolitan area.
- Other contributors: tourism-related construction, Olympic expectations, low interest rates by Bank of Japan, inheritance tax incentives favoring condominiums.
- Note: "In March 2020, the postponement of the Tokyo Olympics by one year was officially announced, due to the COVID-19 pandemic."

### VI. Theory and empirical modeling of population and prices
- Theoretical framework:
  - Durable housing model (Glaeser and Gyourko, 2005): asymmetric sensitivity of house prices to population change; housing durability and kinked supply (elastic when building, inelastic when demolishing) cause losses to have larger price effects than gains.
- Empirical strategy:
  - Baseline regression uses publicly assessed residential land price (¥/m2, inflation-adjusted) as P and decomposes population change into POPLOSS (declines; zero if growth) and POPGAIN (growth; zero if loss), typically using t-1 lags.
  - Expected empirical sign: coefficients positive with β > γ (price decline larger for given population loss than price increase for equal population gain).
- Empirical findings:
  - Sample periods analyzed: (i) 1975–2015, (ii) 1996–2015, (iii) 2010–2015.
  - 1975–2015: estimated coefficient for population loss higher than for population gain, confirming asymmetry.
  - Recent periods: correlation between house price decline and population decline stays robust; correlation between price increase and population increase disappears.
  - Implication: further housing price decline is a medium- and long-term trend in many shrinking-prefecture areas; potential vicious cycle where expectations of declines spur sales, increase oversupply, and depress ownership incentives.

### VII. Policy implications, risks, and measures
- Macro and distributional risks:
  - House price declines reduce household financial and non-financial asset values and collateral values for financial institutions, worsening profitability.
  - Local government revenues from housing transactions fall, potentially reducing social spending and accelerating outflows.
  - Dwelling-related liabilities constitute about 75 to 90 percent of total household liabilities (National Survey of Family Income and Expenditure of 2014), amplifying balance-sheet exposure to house price declines.
  - Regional inequality: greater balance-sheet deterioration expected in rural prefectures than in metropolitan areas, with potential intergenerational transmission.
- Vacant houses (akiya) and oversupply:
  - Vacant/abandoned houses nationwide: approximately 8.5 million houses and vacancy rate of 13.6 percent in 2018.
  - Highest-prefecture vacancy rates: Yamanashi 21.3 percent; Wakayama 20.3 percent; Nagano 19.5 percent; Tokushima 19.4 percent; Kouchi 18.9 percent; Kagoshima 18.9 percent.
- Causes of continued new supply despite vacancies:
  - Weak secondary market for detached houses; preference for newly built houses; tax treatments favor new houses; difficulty valuing used houses; lender reluctance to finance used properties.
  - Property tax distortion: taxable base on land reduced to one sixth of the appraised value for small residential properties when a residential structure remains, creating incentives to retain poor housing stock.
- Policy responses enacted and recommended:
  - Legislative/administrative:
    - Vacant Houses Special Measures Act ("Akiya Taisaku Tokubetsu Sochi Hou") introduced in May 2015 (official announcement November 2014) — municipalities can order demolition of “specially designated vacant houses” under strict procedures.
    - MLIT web-based system to provide information on public assets "for sale" and "for rent."
  - Needed policy directions:
    - Holistic strategy combining urban planning, precise projections of housing supply and demand, mortgage and real-estate regulation, tax policy adjustments, and public-private partnerships to better utilize unused assets.
    - Policies to promote more even regional growth and regional revitalization to retain and attract population to rural areas.
    - Policies to align housing supply with future demographic trends to avoid over-investment, and policies to spur transactions in secondary markets to reduce vacant houses.
- Macro-critical concern:
  - Asymmetric population–price relationship implies persistent downward pressure across most regions, risk of vicious cycle of outflows and falling prices, and potential negative spillovers to banks and debtors from sharp declines in condominium and commercial real estate prices.

### VIII. Selected econometric coefficients and annex highlights
- Annex A — households and housing market:
  - Longer sample (1970–2015): increase in number of households significantly positively correlated with increasing housing prices; no such relationship for decreases in household numbers.
  - Recent sample (2000–2015): decline in number of households significantly related to decline in housing prices.
- Selected regression estimates (preserve reported values):
  - Table 2 (Number of Households and Population Growth)
    - Population: 0.750*** (0.0402) for 1970-2015
    - Population: 0.728*** (0.0423) for 1970-2000
    - Population: 0.419*** (0.132) for 2000-2015
    - Constant: 0.0123*** (0.000768) for 1970-2015; 0.0125*** (0.000696) for 1970-2000; 0.00846*** (0.000223) for 2000-2015
    - Observations: 2,067; 1,315; 658
    - R-squared: 0.860; 0.756; 0.649
  - Table 3 (Number of Households and Population Change)
    - POPLOSS: 0.541*** (0.0947) for 1970-2015; 0.477* (0.268) for 1970-2000; 0.158 (0.109) for 2000-2015
    - POPGAIN: 0.672*** (0.0338) for 1970-2015; 0.627*** (0.0382) for 1970-2000; 0.822*** (0.105) for 2000-2015
    - Constant: 0.0125*** (0.000714); 0.0130*** (0.000667); 0.00749*** (0.000281)
    - Observations: 1,936; 1,205; 596
    - R-squared: 0.829; 0.677; 0.662

*Source — wpiea2020200-print-pdf (IMF Working Paper content as provided).*

### References ________________________________________________________________27

### wpiea2020200-print-pdf - References ________________________________________________________________27

### I. INTRODUCTION
- Advanced economies are undergoing a demographic transition marked by slowing population growth due to low fertility and increased life expectancy (OECD, 2019; United Nations, 2017).
- A shrinking population poses economic challenges: potential lower productivity growth, lower potential growth, and rising age-related fiscal costs such as pensions and medical care (IMF 2018, 2020; McGrattan et al. 2018; Westelius and Liu, 2016; Colacelli and Fernandez-Corugedo, 2018).
- Japan:
  - Total population growth has turned negative since 2011 (Figure 1).
  - Old age dependency ratio exceeded 40 percent in 2014 and is expected to accelerate, reaching above 70 percent in the next 50 years (IMF 2017; Shirakawa 2012).
- Regional heterogeneity:
  - Population is declining in most prefectures due to low fertility and outbound migration to large cities, while population grows in key major cities with inbound migration, increasing population concentration in large cities.
  - Empty houses (akiya) are rising in rural areas; by some estimates, akiya may comprise as much as one third of total houses by 2033.
- Research questions addressed using prefecture-level data (1975 onward) and residential land price as a proxy for house prices:
  - Pattern of prefecture-level population dynamics in the last forty years.
  - Evolution of residential house prices over time in Japan and by prefecture.
  - Importance of demographic factors as drivers of house price dynamics; whether the relationship has evolved over time.
  - Potential policy implications.
- Summary of main findings:
  - Stark urban–rural demographic divergence: by 1998 thirty-four prefectures (out of forty-seven) had negative population growth; by 2017 forty prefectures had negative growth. Seven prefectures recorded positive population growth in 2017: Tokyo-to, Saitama, Chiba, Kanagawa, Osaka, Fukuoka, and Nagoya.
  - Residential house prices declined steadily after the early 1990s bubble burst. Since 2013 (Abenomics), the downward trend reversed: prices stabilized with mild increases in some prefectures and sharp condominium price appreciations in some large cities and tourist destinations.
  - Positive correlation between population growth and house prices: a 1 percent increase in population growth is associated with a 5 percentage points increase in house prices.
  - The population–house price relationship is nonlinear: declines in population are associated with larger magnitudes of house price decreases than the magnitudes of house price increases associated with symmetric population increases (consistent with the durable housing model of Glaeser and Gyourko (2005)).
  - Time variation: in the last ten years the positive correlation between population increases and house price increases has weakened, while the correlation between negative population growth and house price declines remains.
- Policy implications emphasized:
  - Promote more even growth across regions.
  - For rural areas: attract and maintain workers; prevent over-supply of houses as populations decline.
  - Risks for rural prefectures failing to attract people: intensified brain drain, negative effects on productivity and regional growth, downward pressure on house prices, potential financial stability concerns from deterioration of household and regional bank balance sheets.
  - For prefectures facing higher population concentration: implement policies to address within-prefecture distributional issues, especially urban poverty.

### II. DEMOGRAPHIC TRENDS IN JAPAN
- A. Population Change at Prefecture Level
  - Distributional shift:
    - Kernel density of prefecture-level population change shifted left from 1975–1990 to 1990–2015, indicating many prefectures that had growth before 1990 experienced decline since 1990 (Figure 2).
  - 1975–1990:
    - Every prefecture except Akita and Nagasaki experienced positive population growth.
    - Average population growth across regions around 9 percent; median 6.7 percent.
  - Since 1990:
    - Thirty prefectures out of forty-seven experienced population declines in the 1990–2015 period.
    - Largest declines concentrated in Tohoku, Shikoku and Kyushu regions.
    - 1998: thirteen prefectures with net positive population growth.
  - Role of migration:
    - Natural population change (births and deaths) does not fully explain prefecture-level dynamics.
    - Cross-prefecture net migration (inflows minus outflows) is an important driver.
    - Figure 3: rate of net migration into four metropolitan areas (Tokyo, Osaka, Nagoya, Fukuoka) since 1971:
      - Until 2007 all four showed positive change in net migration rate.
      - Since 2007 only Tokyo-to and the Greater Tokyo Metropolitan area continue to show positive growth; rates picked up since 2012.
    - Internal migration example: internal migration accounted for half of population changes due to natural decrease in Tohoku in 2017; in Tokyo, Kanagawa, Chiba, Saitama, Aichi, and Fukuoka natural increase was negative but net migration was positive and exceeded the negative natural change (Population Statistics Year 2019).
  - Population concentration changes (1996–2015):
    - Tokyo concentration increased by 1.3 percent.
    - Hokkaido experienced the largest decline: -0.3 percent.
    - Only six prefectures recorded increased concentration: Tokyo, Kanagawa, Saitama, Hyogo, Shiga, and Fukuoka.
  - Regression analysis (following Glaeser and Gyourko (2005)):
    - POPLOSS variable = 0 if prefecture grew in period t; equals actual percentage decline if prefecture lost population.
    - POPGAIN variable = 0 if prefecture lost population; equals actual population growth rate if it gained population.
    - Results (Table 1, multiple samples and periods) show:
      - Positive coefficients for both population gains and losses, confirming acceleration of inflows into already-concentrated prefectures and outflows from prefectures with negative growth.
      - Time variation and increasing asymmetry: 1975–1990 a 1 percent population loss in the previous year associated with a 0.78 percent decline in the current year; population gains impact 0.71 percent (column (2)).
      - For recent periods (column (5)): population gain persistence 0.77 versus population loss persistence 0.17 — concentration into gain prefectures is accelerating while losses in others are more gradual.

- B. Japanese Households: Total Number and Composition
  - Total number of households:
    - Despite declining total population since 2011, total households continued to increase until 2016 and have since stabilized at around 53 million households.
    - Official projection by the National Institute of Population and Social Security Research: number of households will reach its peak in 2023, at around (projection value follows in the source beyond supplied excerpt).
  - Population vital statistics (context):
    - Number of births peaked at 2,091,983 in 1973 and continued declining thereafter.
    - FY2017 births: 946,065; deaths: 1,340,397 (Population Statistics Year 2019).

*Italic: Source — wpiea2020200-print-pdf - References ________________________________________________________________27*

### 54.2 million, and start a gradual decline n to around 50.8 million households by 2040, which

### wpiea2020200-print-pdf - 54.2 million, and start a gradual decline n to around 50.8 million households by 2040, which

### Households and Family Composition
- Projected household count peak and decline:
  - "54.2 million, and start a gradual decline n to around 50.8 million households by 2040, which is slightly below the total household numbers in 2010."
- Structural shifts in household types:
  - Transition from three-generation households (grandparents, parents, children) to nuclear families (parents and children) as the standard.
  - Among households with persons aged 65 years and over, the share of three-generation families declined dramatically, while the sum of one-person and couple-only households increased to exceed 50 percent.
- Family size and regional differences:
  - Average family size declined steadily, reflecting increases in nuclear and one-person families.
  - The decline in the number of family members per household was larger for rural areas (-31.9 percent) compared to large cities (-28.2 percent), reflecting out-migration of younger generations from rural prefectures.
- Household growth by area:
  - Major cities experienced an increase in household numbers of 115.6 percent compared to other areas at 18.3 percent (observed between 1970 and 2015).

### Housing Price Developments — Aggregate and Historical Context
- Long-run dynamics:
  - Residential land price used as proxy for house prices (public assessment values, yen/㎡).
  - Bubble Period: Since 1988, house price appreciation intensified with an annual average increase of 20 percent; the bubble burst in 1991 with a sharp fall thereafter.
  - Post-bubble: Steady downward trend in house prices except for a "Mini bubble (2002-2008)."
  - Since the beginning of Abenomics, prices increased by 2.7 percent since 2014 and 0.7 percent in 2018; current level recovered to the level in 2013.
- Recent nationwide trend:
  - Average 1.9 percent nationwide increase since 2013 (noting large cross-prefecture variation).

### House Price Changes Across Prefectures (regional heterogeneity)
- Cumulative price changes (2002–2018) and (since 2014):
  - Between 2002 and 2018, all prefectures except Tokyo experienced declines (from -5 to -15 percent range).
  - Since 2014, numerous prefectures experienced appreciations; magnitude varies greatly.
- Notable prefecture changes since 2013:
  - Miyagi, Fukushima: gains by 18 percent (low-base effects related to the earthquake).
  - Tokyo: +16 percent since 2013.
  - Akita: -6.5 percent; Shimane: -5.5 percent; Yamanashi: -4.8 percent.
- Sectoral and local drivers:
  - Large cities and tourist destinations exhibited appreciations, especially via condominium construction booms and investment-seeking behavior in a low-interest-rate environment.
  - Hokkaido: commercial land price up by 60% in 2018 and residential land prices also rose.

### House Price Changes by Housing Type
- Condominium versus detached houses (since 2013):
  - Condominium prices: increased by 23 percent at the national level since 2013.
  - Condominium prices in Tokyo: increased by 45 percent since 2013.
  - Detached houses: increase of 4.2 percent since 2013.
- Drivers of condominium price appreciation:
  - Tax incentives (inheritance tax assessment treats land and dwellings separately, favoring higher-dwelling buildings and lowering assessed values for condominium owners).
  - High investment returns: better-developed secondary market for condominiums, ease of rental, and expectations of rising prices.
  - Foreign investment: aggressive since 2016; total acquisition of commercial real estate by foreign investors ~¥680 billion in H2 2017 (second highest after slightly above ¥800 billion in H1 2007).
- Institutional investment:
  - J-REIT transaction volume has been steadily increasing since 2014; market attracted international investors.

### Tokyo Story — Demand, Supply, and Market Structure
- Population and price divergence:
  - Tokyo (and surrounding prefectures) experienced continued population inflows and is the only region where land house price increased despite the bubble crash.
- Supply-demand indicators:
  - Tokyo has the highest number of housing starts and low vacancy rates, suggesting high demand relative to supply.
- Supply-side concentration:
  - Market share of major seven condominium developers in Tokyo rose from about 20 percent until 2007 to close to 50 percent in 2017, increasing price sensitivity to major developers’ supply decisions.
  - Reverse relationship observed between condominium prices and supply of newly-built condominiums in the Tokyo metropolitan area.
- Other contributors:
  - Tourism-related construction activity (e.g., inflow of tourists, Olympic-related expectations).
  - Low interest rate environment by Bank of Japan boosting real estate investment demand.
  - Inheritance tax incentives increasing demand for condominiums.
- Note on Tokyo Olympics:
  - "In March 2020, the postponement of the Tokyo Olympics by one year was officially announced, due to the COVID-19 pandemic."

### Modeling Population Growth and Housing Prices — Theory and Empirics
- Theoretical framework:
  - Durable housing model (Glaeser and Gyourko, 2005): sensitivity of house prices to population change is asymmetric — population loss causes larger housing price declines than population gain causes price increases, due to housing durability and asymmetric supply elasticity.
  - Kinked supply curve: supply elastic when building, inelastic when demolishing.
- Empirical strategy:
  - Baseline regression specification uses publicly assessed residential land price (¥/m2, inflation-adjusted) as P and decomposes population change into POPLOSS (population decline; zero if growth) and POPGAIN (population growth; zero if loss), using t-1 lags.
  - Expected empirical result: coefficients positive with β > γ (price decline larger for given population loss than price increase for equal population gain).
- Empirical findings:
  - Sample periods: (i) 1975–2015, (ii) 1996–2015, (iii) 2010–2015.
  - 1975–2015: estimated coefficient for population loss higher than for population gain, confirming asymmetry.
  - More recent periods: correlation between house price decline and population decline stays robust; relationship between price increase and population increase disappears — consistent with an overall regional decline in housing prices since the bubble burst.
  - Implication: further housing price decline is a medium- and long-term trend in many prefectures with shrinking populations; potential for a vicious cycle where expectations of price declines spur sales and reduce ownership incentives, increasing oversupply and downward pressure.

### Policy Implications and Risks
- Broad macro and distributional risks:
  - House price declines negatively affect household financial and non-financial assets and collateral values for financial institutions, worsening profitability.
  - Local government revenues from housing transactions fall, leading to potential reductions in social spending and possible acceleration of population outflows.
  - Household liabilities: dwelling-related liabilities (purchase of house and/or land) constitute about 75 to 90 percent of total household liabilities (National Survey of Family Income and Expenditure of 2014), amplifying balance-sheet exposure to house price declines.
  - Regional inequality: greater balance-sheet deterioration expected in rural prefectures than in metropolitan areas, potentially transmitting intergenerationally and widening inequality.
- Specific problems: overinvestment and vacant houses (akiya)
  - Vacant/abandoned houses nationwide: approximately 8.5 million houses and vacancy rate of 13.6 percent (Japan average) in 2018.
  - Prefectures with high vacancy rates: Yamanashi (21.3 percent), Wakayama (20.3 percent), Nagano (19.5 percent), Tokushima (19.4 percent), Kouchi (18.9 percent), Kagoshima (18.9 percent).
- Causes of continued new supply despite vacancies:
  - Weak secondary market for detached houses: preference for newly built houses; tax treatments favor newly built houses; difficulty assessing used house values; lender reluctance to finance used properties.
  - Property tax distortions: taxable base reduction for land when a residential structure remains (taxable base on land reduced to one sixth of the appraised value for small residential properties), creating incentives to keep poor housing stock rather than demolish it.
- Policy responses and measures:
  - Legislative and administrative actions:
    - Vacant Houses Special Measures Act ("Akiya Taisaku Tokubetsu Sochi Hou") introduced in May 2015 to provide municipal powers to address abandoned houses (official announcement in November 2014); authorities can order demolition of “specially designated vacant houses” with strict procedures.
    - MLIT launched a web-based system to provide information on public assets "for sale" and "for rent" to facilitate use of public assets.
  - Needed policy directions:
    - Holistic strategy combining urban planning, precise projections of housing supply and demand, mortgage and real-estate regulations, tax policy adjustments, and public-private partnerships to better utilize unused investment assets.
    - Policies to promote more even regional growth and regional revitalization to retain and attract population to rural areas.
    - Policies to align housing supply with future demographic trends to avoid over-investment, and policies to spur transactions in secondary markets to reduce vacant houses.
- Potential macro-critical concern:
  - The asymmetric relationship between population change and housing prices implies persistent downward pressure in most regions, risk of vicious cycle of population outflows and falling prices, and potential negative spillovers to banks and debtors from sharp declines in condominium and commercial real estate prices.

*Italic: Source — wpiea2020200-print-pdf (IMF Working Paper content as provided).*

### REFERENCES

### REFERENCES

### Bibliographic references
- Colacelli, Mariana and Emilio Fernandez Corugedo, 2018, “Macroeconomic Effects of Japan’s Demographics,” IMF Working Paper 18/248, International Monetary Fund, Washington DC.
- Feiveson, Laura and John Sabelhaus, 2018, “How Does Intergenerational Wealth Transmission Affect Wealth Concentration?” FEDS Notes, June 01, 2018. https://www.federalreserve.gov/econres/notes/feds-notes/how-does-intergenerational-wealth-transmission-affect-wealth-concentration-20180601.htm
- Glaeser, Edward L. and Joseph Gyourko, 2005, “Urban Decline and Durable Housing,” Journal of Political Economy, Vol. 113 (2), pp. 345–75.
- Glaeser, Edward, and Joseph Gyourko, 2017, “The Economic Implications of Housing Supply,” No. w23833. National Bureau of Economic Research.
- Glaeser, Edward, and Albert Saiz, 2004, “The Rise of the Skilled City,” FRB Philadelphia Working Paper No.04-2.
- Hutton, Mercedes, 2018, “Japan’s Hokkaido, the ‘Aspen of Asia,’ Offers Strong Investment Opportunities in a Growing Market,” Mansion Global. https://www.mansionglobal.com/articles/japan-s-hokkaido-the-aspen-of-asia-offers-strong-investment-opportunities-in-a-growing-market-112545
- International Monetary Fund, 2017, “Asia: At Risk of Growing Old before Becoming Rich?” Chapter 2 in May 2017 Regional Economic Outlook for Asia and Pacific, International Monetary Fund, Washington DC.
- International Monetary Fund, 2018, “Japan: 2018 Article IV Consultation – Staff Report,” IMF Country Report 18/333, International Monetary Fund, Washington DC.
- International Monetary Fund, 2020, “Japan: 2019 Article IV Consultation – Staff Report,” IMF Country Report 20/39, International Monetary Fund, Washington DC.
- Kobayashi, Masahiro, 2016, “The Housing Market and Housing Policies in Japan,” ADBI Working Paper 558. Tokyo: Asian Development Bank Institute. Available: http://www.adb.org/publications/housing-market-and-housing-policies-japan/
- McGrattan, R. Ellen, Kazuaki Miyachi and Adrian Peralta-Alva, 2018, “On Financing Retirement, Health and Long-Term Care in Japan,” IMF Working Paper 18/249, International Monetary Fund, Washington DC.
- Nguyen, Anh Thi Ngoc, 2020, “Japan’s Inbound Tourism Boom: Lessons for its Post-COVID-19 Revival,” Working Paper 20/169, International Monetary Fund, Washington DC.
- Nishioka, Hachiro, Masato Shimizu, Yoshimi Chitose, Shiro Koike, and Hiroshi Kojima, 2010, “Recent Migration Trends in Japan Overview of the Results of the Sixth National Survey on Migration (2006)”, The Japanese Journal of Population, Vol. 8, No.1.
- OECD, 2019, “Adapting to Demographic Change,” OECD publishing, Paris. http://www.oecd.org/g20/summits/osaka/OECD-Ageing-and-Demographic-change-G20-JPN.pdf
- Saita, Y, C. Shimizu, T. Watanabe, 2013, “Aging and Real Estate Prices: Evidence from Japanese and U.S. Regional Data,” Tokyo Center for Economic Research (TCER), Paper No. E-68.
- Shapiro, Jesse, 2006, “Smart Cities: Quality of Life, Productivity, and the Growth Effects of Human Capital,” Review of Economics and Statistics, Vol. 88 (2), pp. 324-35.
- Shirakawa, Masaaki, 2012, “Demographic Changes and Macroeconomic Performance: Japanese Experiences,” Speech at the BOJ-IMES Conference, hosted by the Institute for Monetary and Economic Studies, Bank of Japan, Tokyo, May 30.
- United Nations, 2017, “World Population Prospects,” Department of Economic and Social Affairs, Population Division.
- Westelius, Niklas and Yihan Liu, 2016, “The Impact of Demographics on Productivity and Inflation in Japan,” IMF Working Paper 16/237, International Monetary Fund, Washington DC.

### Annex A. Household Numbers and the Housing Market — key findings
- Longer sample (1970 to 2015): an increase in the number of households has a significantly positive correlation with increasing housing prices; no such relationship exists with a decrease in household numbers (Table 1).
- Recent sample (2000 to 2015): a decline in the number of households is most significantly related to the change in housing prices — a positive coefficient implying that a decline in household numbers is related to a decline in housing prices.
- Regression results: number of households increased in proportion to population growth, but the correlation weakens after 2000 when population decline begins (Table 2).
- Population gain/loss separate regressions (Table 3) show:
  - Population losses have weakened impact on household numbers in recent years; population losses do not affect household numbers and only population gains strongly increased household numbers.
  - Household number increases are strongly skewed by population and household dynamics in large cities.

- Selected reported regression coefficients and statistics (Table 2 and Table 3):
  - Table 2 (Number of Households and Population Growth), variables and estimates:
    - Population: 0.750*** (0.0402) for 1970-2015
    - Population: 0.728*** (0.0423) for 1970-2000
    - Population: 0.419*** (0.132) for 2000-2015
    - Constant: 0.0123*** (0.000768) for 1970-2015; 0.0125*** (0.000696) for 1970-2000; 0.00846*** (0.000223) for 2000-2015
    - Observations: 2,067; 1,315; 658
    - R-squared: 0.860; 0.756; 0.649
    - Number of code: 47; 47; 47
    - Fixed effects: Y Y Y
    - Time effects: Y Y Y
    - Time-averaging: 1970-2015; 1970-2000; 2000-2015
    - Robust standard errors in parentheses; significance: *** p<0.01, ** p<0.05, * p<0.1
  - Table 3 (Number of Households and Population Change), variables and estimates:
    - POPLOSS: 0.541*** (0.0947) for 1970-2015; 0.477* (0.268) for 1970-2000; 0.158 (0.109) for 2000-2015
    - POPGAIN: 0.672*** (0.0338) for 1970-2015; 0.627*** (0.0382) for 1970-2000; 0.822*** (0.105) for 2000-2015
    - Constant: 0.0125*** (0.000714); 0.0130*** (0.000667); 0.00749*** (0.000281)
    - Observations: 1,936; 1,205; 596
    - R-squared: 0.829; 0.677; 0.662
    - Number of code: 47; 47; 47
    - Fixed effects: Y Y Y
    - Time effects: Y Y Y
    - Time-averaging: 1970-2015; 1970-2000; 2000-2015
    - Robust standard errors in parentheses; significance: *** p<0.01, ** p<0.05, * p<0.1

### Annex B. Development of Housing Market in Japan from 1982 to 2012 — narrative timeline and key numeric facts
- 1. Bubble period (1982 to 1990):
  - Real estate prices rose by as much as six to seven times during the 1980s asset bubble period.
  - Bank of Japan lowered interest rates from 5.5 percent to 2.5 percent in 1987.
  - Corporate tax rates reduced from 43.3 percent to 42 percent; top marginal income tax rates cut from 70 percent to 40 percent.
  - Yen appreciation after the Plaza Accord: about 50 percent from around ¥240 /USD to about ¥120/USD in less than a year.
  - Land in Ginza 4 Chome reported traded at ¥90,000,000 (US$750,000 at the time) per square meter.

- 2. Bubble burst:
  - In 1990, Ministry of Finance restrictions on total loan volume of real estate lending (Soryo-kisei) and other transaction restrictions curtailed credit and transactions.
  - Bank of Japan tightened interest rates to peak of 6 percent in 1990.
  - Nikkei average reached all-time high of 38,915 in December 1989, then crashed below 20,000 within nine months.
  - By 1995, Bank of Japan slashed rates to 0.50 percent.

- 3. Mini Bubble (2002 to 2008):
  - Bank of Japan cut rates further to 0.1 percent and introduced QE (Quantitative Easing).
  - J-REIT market emergence; property prices in some parts of central Tokyo rose as much as 70 to 100 percent compared to lows seen in 2002.
  - Overall market continued to languish despite central Tokyo gains.

- 4. Mini Bubble burst:
  - By 2006 Ministry of Finance moved to restrict investment in real estate loans.
  - The Great Financial Crisis (2007–2008) led to disappearance of foreign investors and collapse of securitized non-recourse lending market.
  - Real estate prices in Tokyo slumped but by 2010 stabilized around 50 percent above previous lows.

### Annex C. Japan Tax Policies Related to the Housing Market — instruments and effects
- Stamp tax: charged to register and transfer the title of the property.
- City planning tax: charged on the land and structures at 0.3% of the taxable base.
- Real estate acquisition tax: charged on land and structures at 4% of the taxable base.
- Consumption tax: payable on sale of a building (when purchased from a taxable person) but does not apply to sale of land.
- Income Tax Deduction for Mortgages: 1% of the outstanding balance of a mortgage is deducted from the amount of income tax.
- Property tax: charged on the land and structures at 1.4 percent of the taxable base.
  - Exceptional treatment for small residential properties: taxable base on the land is reduced to one sixth of the appraised value if there remains a residential structure on it; cited as contributing to high vacancy rates because owners retain structures to benefit from lower property tax.
  - New legislation enacted in 2014 to accelerate removal of vacant houses.
- Assistance for the elderly: subsidies for construction cost, accelerated depreciation for income tax, reduction of property tax, and JHF mortgages for construction or purchase of properties accommodating elderly accessibility needs.
- Inheritance/Bequest tax: inheritance tax burden in Japan is much heavier than in the United States and most developed economies.
  - Historical valuation practices (until the early 1990s) evaluated real estate below market value for bequest tax assessment, creating tax incentives to hold real estate and take out housing loans (mortgage interest tax deductibility).
  - Strong bequest motives combined with tax system help explain preference for owner-occupied houses and elderly retaining houses until death.

*Source: wpiea2020200-print-pdf - REFERENCES*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020200-print-pdf.pdf_