According to the 1996 by-census, 186,000 households lived in the 39 future for-sale estates and 209,000 households lived in the 60 rental estates. Among typical working-age adults, defined as between 18 and 64 years old, the employment picture was similar for both types of housing. But with the roll-out of the Tenants Purchase Scheme and beginning with the 2001 census, the numbers began to change.

Over time, both men and women in the for-sale estates had significantly higher labour force participation rates and lower unemployment rates than those in the rental estates.

By 2016, the labour force participation rate for men living in the for-sale estates was 0.938, against 0.931 for those living in the rental estates – a 0.7 per cent difference, which translates into an additional 1,230 men in the workforce.

For for-sale housing, the proportion of women working or looking for work was 0.758 in 2016 against 0.746 for rental housing – a 1.2 per cent difference, which translates into an additional 2,000 women in the workforce.

The unemployment rates show a similar trend. In 2016, the jobless rate for men in the for-sale estates was 0.049, against 0.056 for those in the rental estates. Among women, the respective rates were 0.048 and 0.053.

Now, why the difference between people in the for-sale and the rental estates?

One would usually expect labour-market behaviour to be different between homeowners and renters. Economic theory tells us that when households and individuals become wealthier, they have stronger demand for leisure and work less.

Although the typical homeowner under the Tenants Purchase Scheme has to pay an outstanding premium for their unit, equal to half the market value at the time of purchase, this is ultimately a gain for the vast majority. Also, the homeowners’ units will appreciate over time.

On the basis of this increasing wealth, the homeowners should have lower labour force participation rates, and probably higher jobless rates, than the renters. However, in Hong Kong’s public housing, the opposite has happened.

The reason for this anomaly lies in the public rental housing programme.

This provides an incentive for some members of a household in a rental estate to withdraw from the workforce either entirely or from time to time. This sort of opportunistic behaviour lowers their observed labour force participation and raises their observed unemployment.

The homeowners do not have to go out of their way to avoid the penalty, because means tests are only applicable to the renters. This is why the employment picture has been rosier for residents of the for-sale estates after 1998.

But that opportunistic behaviour must be quite prevalent and substantial if it is able to offset the wealth effect from rising property prices and discourage participation in the labour market. This is borne out by census statistics comparing labour-market behaviour between homeowners and renters within the for-sale estates (some flats in these estates remain rental units because the sitting tenants chose not to purchase them).

In 2016, homeowning men were 1.8 per cent more likely to be participating in the labour force than renting men and 1.8 per cent less likely to be unemployed. The corresponding figures for women were 2.8 per cent and 1 per cent. If the Tenants Purchase Scheme was introduced in all 220 public rental housing estates in Hong Kong, the total number of workers could increase by as many as 6,300 men and 10,400 women, producing meaningful relief for the tight labour supply.

In one fell swoop, expanding the scheme would remove the disincentive to work, strengthen attachment to the labour market, and encourage investment in human capital through on-the-job training. This would add to the benefits to the housing market from homeownership, which i have written about previously.

Richard Wong is professor of economics and Philip Wong Kennedy Wong Professor in political economy at the School of Economics and Finance at the University of Hong Kong