(This essay was published in Hong Kong Economic Journal on 27 February 2019.)
The Tenant Purchase Scheme (TPS) was introduced in 1998 as one of two core policy measures for attaining the Government’s goal of achieving 70% homeownership in Hong Kong. The other measure was to supply 85,000 units each year for the following ten years. Both policies were halted in the midst of the protracted deflation and economic stagnation brought on by the 1997 Asian Financial Crisis.
Many today are convinced that it would have been better to temporarily halt these two measures rather than abandon them.
The need to supply more housing units has not gone away and the government today is trying desperately to increase housing supply to make up for the accumulated shortage. There is, however, no will to revive the TPS.
The TPS was implemented 20 years ago from 1998 to 2005. As at September 2018, total of 140,298 units in total have been sold to sitting tenants, representing a very substantial stock of public assets worth at least $300 billion – an amount much larger than the construction cost of the MTR Shatin to Central Link.
These numbers are straightforward to crunch. But what has been less discussed are the labor market effects of the TPS.
The 39 public housing estates selected for the TPS were all built in the years 1982-94, and they were offered for purchase under the scheme between 1998 and 2006. Another 60 estates were also built in the same period, offering an interesting point of comparison.
According to the 1996 By-Census, there were 186,000 households living in the 39 future TPS estates and 209,000 households living in the 60 that would remain as solely public rental housing estates. The labor force participation rates and unemployment rates of typical working age adults (defined as 18-64 years old) can thus be compared before and after TPS was implemented.
In 1996, the estimated labor force participation rate for the typical adult working age man living in one of the 39 future TPS estates was 0.973, which was marginally lower than the estimated 0.975 for those living in the 60 public rental housing estates. Among adult working age women the rates were 0.578 for those living in future TPS estates and 0.601 for those living in the 60 public rental housing estates, representing a significantly higher rate.
The unemployment rate for men in the TPS estates was 0.043, marginally higher than the 0.041 rate for men in the 60 public rental housing estates. For women the rates were identical in both estates, at 0.039.
In 1996, therefore, the labor picture was relatively comparable between both kinds of estates. But after the TPS started to be rolled out and beginning with the 2001 Census data, that began to change.
TPS estates had significantly higher labor force participation rates and lower unemployment rates than public rental housing estates over time among both adult working age men and women (see Table 1).
Among adult working men, the labor force participation rates of TPS estate occupants were 0.950 in 2001, 0.952 in 2006, 0.934 in 2011, and 0.938 in 2016. These are all higher than the corresponding figures of 0.945 in 2001, 0.941 in 2006, 0.924 in 2011, and 0.931 in 2016 for those living in the public rental housing estates. In 2016, the labor force participation rate of men was 0.7% higher among TPS estate occupants, which translates into an additional 1,230 men in the workforce. Over the period 2001-2016 the average increase in men’s labor force participation rate was 0.8 per cent.
Among adult working women, the labor force participation rates of TPS estate occupants were 0.675 in 2001, 0.742 in 2006, 0.771 in 2011, and 0.758 in 2016. These are all higher than the corresponding figures of 0.673 in 2001, 0.719 in 2006, 0.742 in 2011, and 0.746 in 2016 for those living in the public rental housing estates. In 2016, the labor force participation rate of women was 1.2 per cent higher among TPS estate occupants, which translates into an additional 2,000 women in the workforce. Over the period 2001-2016 the average increase in women’s labor force participation rate was 1.6 per cent.
Similar results can be seen in the unemployment rate. Among men living in TPS estates, the rates were 0.083 in 2001, 0.073 in 2006, 0.053 in 2011, and 0.049 in 2016. These are almost all lower than the corresponding figures for men living in the public rental housing estates of 0.080 in 2001, 0.083 in 2006, 0.062 in 2011, and 0.056 in 2016. In 2016, men’s unemployment rate was 0.7% lower among TPS estate occupants, which translates into an additional 860 men in the workforce. Over the period 2001-2016 the average decrease in men’s unemployment rate was 0.5%.
Among women, the unemployment rates of TPS estate occupants were 0.047 in 2001, 0.077 in 2006, 0.057 in 2011, and 0.048 in 2016. These are all higher than the corresponding figures of 0.044 in 2001, 0.077 in 2006, 0.060 in 2011, and 0.053 in 2016 for those living in the public rental housing estates. In 2016, the unemployment rate of women was 0.6% lower among TPS estate occupants, which translates into an additional 680 women in the workforce. Over the period 2001-2016 the average decrease in women’s unemployment rate was 0.1%.
Why are there these differences in labor force participation and unemployment rates between occupants from TPS and public rental housing estates?
Normally one would expect the labor market behavior of homeowners to differ from renters. Economic theory tells us that households and individuals who become wealthier will increase their demand for leisure and work less.
Although the typical TPS unit homeowner has to pay the outstanding unpaid premium of their unit equal to half the market value at the time of purchase, for the vast majority this ultimately is a gain that has been further enriched by the capital appreciation of their TPS units over time.
On this basis of this increasing wealth, TPS homeowners should have lower labor force participation rates and probably also higher unemployment rates than public rental housing estate occupants. But this is the exact opposite of what has happened.
The reason for this anomaly lies in the public rental housing program.
Tenants in public rental housing estates are means tested after 10 years of occupancy. If they are found to have become well-off, then they have to pay up to double rent. This requirement provides an incentive for some members of the household to withdraw from the workforce either entirely or from time to time to avoid becoming a well-off tenant. Such opportunistic behavior lowers their observed labor force participation rates and raises their observed unemployment rates.
TPS homeowners do not have to engage in such opportunistic behavior because means tests are only applicable to renters and not homeowners. As a consequence, TPS estate occupants after 1998 are observed to have higher labor force participation rates and lower unemployment rates.
Opportunistic behavior must be quite prevalent and substantial if it is able to offset the wealth effect of rising property prices such that it discourages labor market participation among tenants. This is borne out by the Census statistics of observed labor market behavior between homeowners and renters in TPS estates (see Tables 3 and 4). Note that some flats in these estates remain public rental housing because the sitting tenant chose not to purchase.
In 2016, the difference in labor force participation rates between male TPS homeowners and renters was 0.018 or 1.8%, and for females 0.028 or 2.8% for women. For the period 2001-2016, the average difference among men was 0.022 or 2.2%, and 0.041 or 4.1% for women.
In 2016, the difference in the unemployment rates for men was 0.018 or 1.8%, and for women 0.010 or 1.0%. For the period 2001-2016, the average difference among men was 0.027 or 2.7%, and among women 0.021 or 2.1%.
These figures imply that implementing the TPS in 1998 encouraged former tenants who became homeowners to become more attached to the workforce by removing the disincentive to work due to means testing and the double rent threat. The result was an increase in the workforce of 1,270 men and 2,070 women in 2016 due to higher labor force participation rates and lower unemployment rates.
If the TPS were introduced in all 220 public rental housing estates in Hong Kong in 2016, it is probable that the total increase in workers could be as many as 6,300 men and 10,400 women. This would increase the male labor force by 1.0% and the female labor force by 1.8%, producing a meaningful relief for our labor supply tightness.
It would also discourage opportunistic behavior, foster greater attachment to the labor market, and encourage more human capital investment through on-the-job training – to add to the other housing market benefits that I have written about in previous columns.