(This essay was published in Hong Kong Economic Journal on 27 March 2019.)
The Tenant Purchase Scheme (TPS) introduced in 1998 sold 140,298 public rental housing (PRH) units to sitting tenants. Initially about 57% of the units available for sale were taken up and in the past decade this increased to 72% of all units offered. Through promoting homeownership, the TPS has increased the labor force participation rate and decreased the unemployment rate of those families who bought these units.
There were two other important effects. The TPS also lowered inequality in housing property wealth and promoted family stability.
Family stability was promoted because the TPS provided an opportunity for the tenants to become homeowners. It created an economic stake for the family to keep together because they now possess a jointly owned asset, whose value appreciates over time that can then be shared if they stay together. If the family divorces, sharing in the full present and future value of the housing asset becomes less certain.
Unlike TPS homeowners, PRH estate tenants suffer very little economic loss when a married couple divorces because they do not jointly own an asset to begin with. Normally, one of the divorced tenants keeps the PRH unit, often the mother who has custody of the children. The other spouse, often the husband, leaves the present PRH unit but is free to apply for another one and join the applicants’ waiting list. The waiting time is faster if the applicant remarries and has children in the meantime.
This implies that TPS homeowners have a strong incentive to maintain family stability to avert the economic loss from a divorce. Perversely, PRH tenants face no deterrence against divorce, since both parties are still eligible to apply for PRH units again. In a sense, this may even expand their total accommodation. Given this, the current government policy to shift public housing policy towards homeownership and away from rental housing is to be lauded.
The extent of family stability contributed by homeownership can be seen by looking at the TPS twenty years after it was introduced in 39 housing estates, which were selected from among 99 PRH estates built between 1982 and 1994.
The first piece of supporting evidence is that the average divorce rates for ever married men and women aged 18 to 54 were uniformly lower in TPS estates relative to PRH estates by 1.3% for men and 2.2% for women (see Table 1).
The average divorce rates for men in TPS and PRH estates were, respectively, 3.5% and 4.7%in 2001, 5.2% and 6.8% in 2006, 7.4% and 8.6% in 2011, and 8.4% and 9.4% in 2016. For women, the divorce rates in TPS and PRH estates were, respectively, 6.4% and 7.7% in 2001, 11.0% and 13.6% in 2006, 13.0% and 15.2% in 2011, and 12.8% and 15.3% in 2016. The reason why divorce rates were uniformly higher for women was presumably because remarriage opportunities are better for men.
The link between homeownership and divorce behavior is clearly evident when comparing TPS homeowners and tenants (the latter chose not to purchase the offered flat and remained as tenants).
The average divorce rates for owners in TPS estates were uniformly lower than those for tenants, by 3.7% for men and 9.2% for women (see Table 2). The divorce rates of men who were owners versus tenants were, respectively, 2.3% and 4.7% in 2001, 3.7% and 7.7% in 2006, 6.0% and 10.3% in 2011, and 7.4% and 11.4% in 2016. For women, they were 3.6% and 8.9% in 2001, 7.1% and 17.0% in 2006, 9.3% and 20.1% in 2011, and 9.9% and 20.7% in 2016. Again the divorce rates were uniformly higher for women than men.
A second piece of evidence of homeownership fostering greater family stability is the incidence of single parent households. The average single parenthood rates among ever married men and women aged 18 to 54 were uniformly lower in TPS estates than PRH estates by 0.7% for men and 2.4% for women (see Table 3).
The average divorce rates for men in TPS versus PRH estates were, respectively, 1.9% and 2.3% in 2001, 2.1% and 3.3% in 2006, 3.1% and 4.2% in 2011, and 3.5% and 3.6% in 2016. For women, the rates were 5.8% and 6.1% in 2001, 6.4% and 9.0% in 2006, 6.2% and 10.0% in 2011, and 7.0% and 10.0% in 2016. As with divorce rates, the single parenthood rate was uniformly higher among women than men.
Comparing tenants and owners in TPS estates also yielded similar results to the divorce rates. The average single parenthood rate was lower among owners than tenants, by an overall 2.1% for men and 8.8% for women (see Table 4). For men, the single parenthood rates among owners and tenants broke down to, respectively, 1.0% and 2.5%in 2001, 1.4% and 3.3% in 2006, 2.3% and 4.6% in 2011, and 2.9% and 5.6% in 2016. For women, the single parenthood rates for owners and tenants were, respectively, 2.6% and 9.2% in 2001, 2.7% and 12.0% in 2006, 2.5% and 12.7% in 2011, and 4.4% and 13.2% in 2016. Again, in all years single parenthood rate were higher among women than men.
A third piece of evidence of homeownership fostering greater family stability can be seen in the proportion of young single men and women, aged 18-39, who live with their parents. This figure was mostly higher in TPS estates relative to PRH estates, by 3.0% and 4.3%, respectively, for both men and women (see Table 5). A stable family environment fosters familial bonding and encourages young single adults to stay with their parents.
The breakdown of these figures shows that the proportion of single men living with their parents in TPS and PRH estates was, respectively, 74.1% and 74.5%in 2001, 85.8% and 79.8% in 2006, 85.1% and 79.8% in 2011, and 78.7% and 77.6% in 2016. For single women, the proportion was 60.5% living with parents in TPS estates and 62.5% in PRS estates in 2001, 76.4% and 67.7% in 2006, 77.5% and 69.9% in 2011, and 69.6% and 66.8% in 2016. Overall, there was a smaller proportion of young single women living with their parents than men, presumably more women married at a younger age and moved out of the parents’ household.
When comparing owners and tenants in TPS estates, the proportion living with their parents was mostly higher, by 2.4% for men and 7.3%, for women (see Table 6). Among young men, the proportion was 75.2% in owner flats and 68.3% rental flats in 2001, 87.4% and 82.7% in 2006, 85.6% and 83.6% in 2011, and 77.8% and 81.9% in 2016. For women, the figures were 63.7% and 49.9% in 2001, 80.4% and 69.0% in 2006, 79.6% and 72.3% in 2011, and 68.8% and 72.3% in 2016. Again there were proportionately more young single men than women.
The evidence from Census figures from 2001 to 2016 shows that the TPS scheme introduced in 1998 to bring subsidized homeownership to 39 PRH estates produced three effects that reduced family instability. Comparing TPS estates with PRH estates, first, average divorce rates were significantly lower; second, average single parenthood rates were also significantly lower; third, the proportion of young single adults living with their parents was significantly higher in TPS estates than PRH estates.
These results are further corroborated by evidence that similar outcomes were seen when comparing owner TPS households and tenant TPS households.
An important consequence of these outcomes is that they have increased the average household size in TPS estates relative to PRH estates. For the period 2001-2016, the average household size in TPS estates was 3.2 while in PRH estates it was 3.0 (see Table 7).
Why is this important? Because it suggests the TPS has reduced the demand for housing. TPS owner families were living together voluntarily in the same housing unit and having fewer members move out to set up a separate household. This indirectly reduces the demand pressure in the private rental market, especially the demand for sub-divided flats.
The additional number of individuals accommodated in the larger households of the 39 TPS estates amounted to 23,900 in 2001, 19,300 in 2006, 15,600 in 2011, and 3,500 in 2016. This averages out over the four census years to 15,600. If the TPS had been introduced in all 220 PRH estates, one could by projection have increased the total accommodation of persons within the existing stock by an average 50,900 persons per year. The housing shortage that we are now facing would have been less severe.
The divorce and single parenthood rates would also have been impacted. Over the four census years, the TPS estates had 3,400 fewer divorced individuals and 1,700 fewer single parents between the ages of 18 to 54 when compared to PRH estates.. If all 220 PRH estates had adopted the TPS, the projected decreases would have been 11,800 and 8,000, respectively. This would have been a significant contribution to family stability, which could well help alleviate the concern about intergenerational immobility. And all this is in addition to the benefits of higher labor force participation rates and lower unemployment rates from the TPS that I discussed in my last article.