130,000 applicants on the waiting list (Hong Kong Housing Authority, 2010b). The Housing Authority sets a target of maintaining the average waiting time at 3 years.
In the process of public housing allocation, applications are processed strictly in accordance with the application sequence numbers and the applicants’ choices of district.However, those who ful?ll the eligibility criteria of the Express Flat Allocation Scheme may have an earlier chance to be allocated public rental housing units. There are also separate criteria and allocation schemes for people affected by squatter clearance and public housing redevelopment.
According to existing public housing policies, publichousing tenants may apply for ?at transfer under various situations such as increased family size, special medical or social grounds, or major improvement or redevelopment programs implemented by the Housing Authority(Hong Kong Housing Authority, 2010c).Whenever possible, eligible tenants will be transferred to suitable ?ats within the same estate. Tenants may also apply for special transfer to another estate. Due to persistent shortage of public rental housing units, even if a tenant is eligible for ?at transfer, the waiting time can be quite long and there is no guarantee that the new ?at will match the tenant’s Preferences.
Besides applying for a public rental housing unit, eligible households can also apply to buy subsidized ?ats through the Home Ownership Scheme (HOS).In setting the price of HOS ?ats, the Housing Authority follows three guiding principles: (i) the mortgage-to-income ratio should not be more than 40%; (ii) 50% of the ?ats should be affordable to the target group; and (iii) ?ats should be offered at a 30% discount of market value.
Starting from the 3rd year of occupancy, HOS ?at owners may sell their ?ats in the open market after paying a premium to the Housing Authority. The premium is basically equal to the prevailing market value multiplied by the discount enjoyed by the owner at the time of purchase.For example, if the prevailing market value of a HOS ?at is HK$2000,000 and the ?at was offered at a discount of 30% at the time of purchase, then the premium is equal to HK$600,000. Owners of HOS ?ats can also sell their ?ats to existing public rental housing tenants. In this case, the original HOS owner will not be required to pay the premium to the Housing Authority before selling the ?at,but the purchaser will inherit the liability to pay the
20
premiumshould he or she seek to sell the ?at in the open market later on.With these restrictions in place, the secondary market for HOS ?ats is quite inactive.
Empirical framework
The data set we use in this study is a random sub-sample of the Hong Kong Population Census. Relying on information from the population census ?le, we compare the pattern of internal mobility between public housing and private housing occupants. A study of public housing in Hong Kong (Wong and Liu, 1988) adopts a similar approach, but the focus of that study is on the distortion in the level of housing consumption. Wong (1998) elaborates this and other distortions induced by public housing, and makes a case for privatization. By introducing another dimension to the misallocation problem (namely, mobility and location choice), our study adds to the debate concerning various policy alternatives such as privatization or private-market rent subsidy. Although housing tenure plays an important role in mobility, the cross-sectional data set used in this study does not contains adequate information for more in-depth analysis. Hence, this paper should be read with this limitation in mind.
Because of a persistent shortage of subsidized housing,it is much more dif?cult for a public housing tenant to obtain relocation to another public housing unit than for a private housing occupant to move to another ?at. Moreover, it is often an unrealistic option for public-housing tenants to move to a private ?at because doing so entails losing a substantial government subsidy. This is similar to the idea of the ‘‘job lock’’ effect in labor economics (Madrian, 1994; Gruber and Madrian, 1994). In the ‘‘job lock’’ effect, one may argue that the value of health insurance provided by the employer is not a pure subsidy because of possible compensating wage differences. In the case of government housing, the low rental is indeed a subsidy. We therefore expect the ‘‘housing lock’’ effect on internal migration to be even stronger than the ‘‘job lock’’effect on job mobility. We hypothesize that, other things equal, public housing occupants are more immobile than private housing occupants (Hypothesis I).
We argue that public housing occupants are less mobile than private housing occupants because the range of options available to them is more limited. By the same logic,if we
21
compare people who actually made a move, we expect to ?nd that a tenant in public housing is less likely to move to a unit that matches his preferences. Therefore,conditional on moving, we expect that public-housing tenants are more likely than their private housing counterparts to have to move farther away from their original district or region of residence (Hypothesis II).
Another distortion created by public housing is that public housing bene?ts tend to tie down tenants to their current location. When circumstances change, such as change to a new job or having a baby, one may consider moving to live in another location. Siegel (1975) examines the relationship between the home and job location of the household within a metropolitan area (see also Simpson,1987). However, tenants in public housing are less likely to move to more convenient locations and their location choices may not accurately re?ect their true location preferences. In an earlier study in the United Kingdom, Hughes and McCormick (1987) show that public housing inhibits local movement for job-related reasons. We therefore hypothesize that public-housing tenants are less likely to work in the same district or region as their place of residence (Hypothesis III).
Similarly, the market size for subsidized sale ?at is much smaller than that of private sector and these subsidized ?ats are often located far from centers of economic activities. When a household determines to buy a subsidized ?at, the location choice may not accurately re?ect their true location preference. Therefore, conditional on moving, we expect that owners of public housing are more likely than their private home owners to have to move farther away from their original district or region of residence(Hypothesis IV). When compare with the private sector,public home owners are less likely to move to more convenient locations. We therefore hypothesize that public home owners are less likely to work in the same district or region as their place of residence (Hypothesis V).
These hypotheses are tested using probit regressions on census data in 2001. The census data set contains information about whether a household has moved in the previous 5 years, and information about the past and present districts of residence. Additional variables such as income and education can be used to control for other differences between public and private housing occupants. However,there may be unobservable differences between public and
22
private housing occupants that affect our outcome variables. Because we do not have convincing instrumental variables, it is dif?cult to tease out the direct effect of public housing from the indirect effects induced by the correlation between public housing occupancy and unobservable variables. We will attempt to partially address this issue by introducing an extensive set of control variables. But to the extent that there are important unobserved differences across occupancy groups that affect the outcome variables, the estimates from our probit regressions cannot be given a causal interpretation.
Suppose a dummy variable M is equal to 0 if a tenant has not moved in the previous 1 years, and M = 1 if the tenant has moved. The probability that M = 1 can be represented by a probit model with M as dependent variable.The independent variable of interest is whether the tenant resides in public or private housing. We also include an extensive set of control variables such as age, gender, education, income, spouse, household size, marital status,employment status, dummy variables for young children,and industry and occupational dummy variables. Based on Hypothesis I, we predict the coef?cient estimate for the public housing tenant dummy variable to be negative and signi?cant.
Suppose the dummy variable S is equal to 0 if a tenant has moved to a new district in the previous 5 years, and S = 1 if the tenant stayed in the same district. Then the probability of S = 1 conditional on having moved in the past 5 years can be denoted by Prob(S =1|M = 1). This conditional probability can be estimated using the sub-sample of households that had moved in the previous 5 years.Hypothesis II predicts that the coef?cient estimate for the public housing tenant dummy variable of this probitmodel is negative and signi?cant.
Finally, let W= 1 if place of work and place of residence are in the same district (or in the same region), and W=0 otherwise. Hypothesis III predicts that in a probit regression with W as the dependent variable, the coef?cient estimate for the public housing tenant dummy variable is negative and signi?cant.
Empirical results
The data set we use in this study is the 5% random subsample of the 2001 Hong Kong Population Census. We include all heads of households aged 18 or above who were born in Hong Kong,Macau orMainland China. By de?nition,a household refers to a housing domestic
23
相关推荐：
loading