The Relationship Between Homeownership and Life Satisfaction in Germany

Abstract

The article investigates the relationship between homeownership and life satisfaction in Germany, using German Socio-Economic Panel Study data from 1992 to 2009. The data not only allow controlling for a wide range of variables, but also tackle various measurement problems of previous studies. Ordered logit models support a positive relationship between homeownership and life satisfaction. In addition, regression models with fixed effects also reveal unobserved differences between homeowners and renters. Further results show that homeownership is particularly important for low-income households and that there is a significant interaction between homeownership and the condition of the dwelling as well as homeownership and the financial burden of the household. Overall, the findings indicate that policies aim at a higher degree if life satisfaction should not focus on the promotion of homeownership alone, but also support home maintenance costs.

Keywords::

• Homeownership
• housing
• subjective well-being
• life satisfaction
• maintenance
• finance

Keywords::

• D1
• R20
• R21

Notes

¹ The terms life satisfaction, subjective well-being, and happiness are used interchangeably.

² A major contribution of Rossi and Weber's study is the extensive range of research questions they identified and which paved the way for various follow-up studies. These include the following questions: Is homeownership in reality a preference for single-family detached housing? What are the benefits and costs of homeownership and are they in or out of balance? Is there a difference between homeowners and renters in their participation behavior in local community activities?

³ For the determinants on life satisfaction in Germany, see for example, Frijters et al. (2004).

⁴ Observations for health status in 1993 are completely missing. I imputed the missing values by the mean of health status for each individual in 1992 and 1994.

⁵ The correlation coefficient between children and persons per household reveals a high degree of association. Accordingly, the variable also serves as a control for children.

⁶ For a more detailed description of the ordered logit model, see Wooldridge (2001) or Cameron & Trivedi (2005).

⁷ The Hausman test revealed that the fixed effects model is the appropriate model compared with the random effects model.

⁸ For a more detailed description of the fixed effect model, see Wooldridge (2001) or Cameron & Trivedi (2005).

⁹ In practice, differentiating between ordinality and cardinality does not make much difference (Frijters et al., 2004).

¹⁰ A critical comment on the ordered fixed effects model can also be found in Berger (2009).

¹¹ The results of the ordered logit models are comparable with those derived from ordinary least squares regressions.

¹² In addition, ordered logit regressions for different age groups show that homeownership is more important for an individual's life satisfaction, if the individual is approaching retirement age. The odds ratio of higher life satisfaction between homeowners and renters is greater for people between 45 and 65 years, than for those up to 30 years, and even greater for people above 65, i.e. those who have already reached retirement age. (The detailed results are not listed here.)

¹³ The exact phrasing changed slightly over time. Unfortunately, data for 2003 and 2004 are not available.