Where you live may be an important factor in the sorts of jobs you find, where your children go to school, and how safe you feel walking home after working late. A tight rental market clearly limits housing choices for the less well-off. But are there other factors that might affect this – such as ethnicity?
How we tested for discrimination
To investigate this, we designed an experiment to compare the experience of Anglo, Indian and Muslim Middle Eastern “renters” in Sydney, to see if they were treated differently. We chose to look at these two ethnic minorities because other studies have reported they experience housing discrimination.
Each week between August and November 2013, we sampled properties advertised for rent across Sydney. We used paired testers (people of the same gender and age, but different ethnicities) as “mystery shoppers”. They phoned the listed agent and then, if possible, attended an inspection in person.
The testers completed a detailed questionnaire about what the agents told them. We completed a total of 537 phone-call tests and 369 in-person inspection tests.
The study covers only the enquiry and inspection stages of the housing search. We couldn’t test what happens once an application is actually submitted, because that would have involved lying. However, even at these early stages, we found some clear differences in how Anglo testers were treated compared to Muslim Middle Eastern or Indian testers.
What happened in the tests?
The chart below summarises the proportion of “tests” (properties that both testers visited) falling into one of three categories: the minority tester was favoured; both testers were treated identically; and the Anglo tester was favoured.
We tested which of these differences were statistically significant (not likely to be the result of random variation). The net difference between the numbers of Anglos and minorities favoured is shown in the chart labels, along with the statistical significance of the difference.
On some dimensions, minority testers were more likely to be favoured. Agents were more likely to tell the minority tester when the property would be available and that all adult occupants must sign a lease.
Agents were also more likely to take the minority testers' details – although they were less likely to contact them after the inspection.
But on most dimensions, particularly those that might meaningfully affect someone’s search for housing, agents favoured the Anglo tester. Agents were three to four times more likely to offer Anglo (but not minority) testers an individual appointment to inspect the property, to ask Anglo testers about their housing needs and to tell them of other available housing.
Agents also gave Anglo testers information about the application process that the minority partner was not given (such as the deadline for applications). And, although agents did this in only a small number of cases, they were more likely to contact Anglo compared to minority testers after the inspection.
All of these differences might make it far easier for Anglo renters to find the right property, at the right price and in the right location.
What can we reasonably conclude?
Paired testing has its limitations. It is difficult to match individuals perfectly. Underlying characteristics (such as how confident someone is) might affect how they are treated.
We screened testers carefully and spent time training them. We also monitored the weekly survey responses to control quality.
One obvious limitation is that we looked at the experiences of only two ethnic minority groups, because we had limited funds. It’s likely that other ethnic and racial groups might have quite different experiences in the rental housing market. Our results also don’t say anything about other Australian cities.
Despite these limitations, we believe paired tests offer unique insight into how people of different ethnicities are treated in the market. Surveying agents or renters about their perceptions would have resulted in just that – a study of perceptions.
No-one can accurately compare their own experience of the world to someone else’s. So, relying only on complaints of discrimination is likely to underestimate substantially the actual incidence of differential treatment. We think our study offers a much more robust glimpse at housing inequity and has the potential to supplement anti-discrimination enforcement.
Our paper on the study, Rental discrimination in the multi-ethnic metropolis: evidence from Sydney, will soon be accessible when published in Urban Policy and Research.