The Effect of Housing Attributes on Occupant Health in a Less Developed Country: Evidence from Uganda

Presenter: Michelle Amaral, University of the Pacific

Abstract

This paper uses the third round of the Uganda National Household Survey (UNHS III) to study the relationship between housing quality and resident health. Count data models estimate the number of days ill in a 30-day time interval as a function of measurable housing characteristics, demographics, diet and location. Both Poisson and negative binomial models estimate the relationship. We show that some, but not all, of these characteristics adversely affect health. Most notably, we find that indoor burning necessary for cooking and lighting affects different measures of illness in expected ways. Consistent with other work, we find that household crowding also adversely affects health.

We stratified our sample by respiratory illness and gastrointestinal illness as a specification check and found estimated coefficients for control variables consistent with our base model. As a separate specification test we were able to stratify our sample by an illness that had no a priori expectations that housing quality would have an impact. The results are consistent with this expectation.

To our knowledge this is the first paper to demonstrate a relationship between objective housing measures and the health of housing occupants. Much of the housing-health relationship we show emanates from the lack of ubiquitous access to electricity in Uganda, which requires indoor burning of wood, kerosene, and paraffin for cooking and lighting purposes. Moreover, the impact of inferior cooking methods has the greater detrimental effect. Additionally, and consistent with other work, we show that crowding also has adverse health consequences.