Demographics, Human Capital, and the Demand for Housing

1
Demographics, Human Capital, and the Demand for
Housing

• Piet Eichholtz
• Maastricht University
• Thies Lindenthal
• Maastricht University
• ICPM Netspar Conference
• Maastricht University, 30 October 2007

2
Expected change in total population, 2005-2050
Large differences across Europe
Source United Nations
3
Housing performance and demographic contraction
Limburg is lagging behind the national trend
4
Shrinking AmsterdamPopulation decline 1795-1814
drove down house prices and rents
5
Structure of presentation

• Introduction
• Method and Data
• Results
• Conclusion

6
Intention of the paper is to understand (future)
housing demand better

• How do demographic changes influence the demand
  for residential real estate?
• Will demand for housing decline when population
  growth slows down and societies become older?
• This paper contributes to the discussion in three
  ways
• Refined methodology
• Very detailed and high-quality data
• European evidence

7
Preview of results

• Demographics impact the demand for housing
• Human capital is one of the key drivers
• Education, income, health, employment status
• Housing demand does not decline with age, but
  increases
• Positive human capital effects get stronger with
  age
• Aging and slowdown in population growth do not
  necessarily imply a decline in overall housing
  demand
• Education effect may even offset shrinking
  population

8
Literature ReviewThe first wave of research

• Mankiw and Weil (1989) started the debate
  claiming that aging baby boomers will demand less
  housing in the future
• They predicted house price drop of 47
• Intense criticism by (inter alia) Peek and Wilcox
  (1991), Hendershott (1992), Engelhardt and
  Poterba (1991)
• Green and Hendershott (1996) find housing demand
  to stay constant with age
• Education is main driver of demand

9
International evidence Good empirical studies
are still very scarce

• England Ermisch (1996)
• Demand partly explained by demographics
• Japan Ohtake and Shintani (1996)
• Short run price effects of demographics, long run
  supply adjustment
• Sweden/OECD Lindh and Malmberg (1999)
• Demographics explain new construction in Sweden
  and OECD
• Austria Lee et al. (2001)
• Number of households is important
• The Netherlands Neuteboom and Brounen (2007)
• Demand for housing will not go down in aging
  society (due to cohort effects)

10
Agenda

• Introduction
• Method and Data
• Results
• Conclusion

11
First decompose, then predict demand for housing
Control for housing quality and the demographic
profile of household

• Decompose house into housing services
• Investigate willingness to pay for these services
• Investigate the role of households demographic
  situation and human capital
• Define a constant quality house
• Calculate the willingness to pay for this house
  as a household becomes older
• Predict housing demand with changing demographics

12
Refining the methodology Cohort variables versus
life-cycle variables

• Cohort variables do not change when households
  grow older
• Gender, ethnicity, education, birth-cohort
• Mankiw and Weil these variables change as
  households age
• Life-cycle variables depend on the household's
  position in life-cycle
• Household size, employment status, income, health
  of household members
• We take age as a proxy for the position in the
  life-cycle
• Explicitly model income differences over time
• Green and Hendershott these variables are
  constant as households age

13
English Housing Condition Survey (EHCS)Covers
both housing data and demographic information

• British government collected data on the current
  housing stock
• Study provides a representative cross-section of
  households and their houses
• We use the 2001 cross-section
• Excellent level of detail and quality of data
• More than 900 variables, 17,500 households
• Housing characteristics and values from
  professional inspections of dwellings
• Information on household based on interviews
• Subsidies distort picture exclude all subsidized
  housing, 10,000 left

14
Agenda

• Introduction
• Method and Data
• Results
• Conclusion

15
Hedonic regression dwelling related
variablesHow much are the components of a
dwelling worth on average?
16
Hedonic regression location related
variablesHow much are the components of a
dwelling on average worth?
17
Demographic regressionControlling for household
size and income
18
Demand increases with educationAdditional
educational achievement drives up reservation
prices
19
Impairments to human wealth drive down
demandNegative impact of disabilities, long-term
illnesses, and children
20
Additional results of the demographic regression

• Importance of education increases with age
• Older university graduates willing to pay more
  than younger ones
• No results for age and health
• Analysis of the interaction terms for age and
  chronic illness, and for age and disability does
  not yield any significant results
• Full time employment
• Drives down willingness to pay

21
Recap of the demographic regression results

• Willingness to pay for housing
• Increases with household size and income
• Decreases with children
• Human capital is key driver of demand
• Education drives up demand
• Chronic health problems and disabilities decrease
  demand
• Age has positive effect on demand
• Age-income effect is positive
• Age-education effect is positive
• When calculating future housing demand, the
  dynamics of these variables must be considered
• Cohort variables vs. life-cycle variables

22
Household's willingness to pay for constant
quality houseOverall, demand is upward sloping
as households become older
23
Demand for different dwelling typesUpward-sloping
with age for all types Detached houses and
bungalows steepest increase
24
Similar demand growth for the English population
scenariosBased on different assumptions
fertility, migration, and life expectancy
25
Will higher demand translate into higher prices?

• Malpezzi and Maclennan (2001) find supply
  elasticities between 0 and 1 for post-war UK
• The range depends on the assumptions for their
  models
• Office of the Deputy Prime Minister projects
  housing shortages if supply remains at current
  level

26
Agenda

• Introduction
• Method andData
• Results
• Conclusion

27
Conclusion

• Demographics influence the demand for housing
• Education, income, health, employment status, and
  household size are main drivers
• Housing demand does not decline with age, but
  increases
• A slowdown in population growth (or even a
  shrinkage) does not necessarily imply a decline
  in overall demand
• Human capital will keep on increasing
• Younger generations better educated
• Improving health
• Study provides analytical framework to apply to
  other (European) countries and regions
• Housing remains key asset in private retirement
  portfolio