NUS Department of Economics
Page 1 of 1
myEMAIL
IVLE
LIBRARY
MAPS
CALENDAR
SITEMAP
CONTACT
HOME
DEPARTMENT
PEOPLE
PUBLICATIONS
EVENTS
GRADUATE
UNDERGRADUATE
CENTRE/UNIT
D epart m ent of Econom i cs
Faculty of Arts and Social Sciences
Si ngapore Cent re f or Appl i ed and Pol i cy Econom i cs
Home About SCAPE Steering Committee Staff Directory Working Papers Events Newspaper Articles Econometric Studies Unit Housing affordability in Singapore
Methodological notes on lifetime income and housing affordability
Main references: Abeysinghe, Tilak and Jiaying, Gu (2011) "Lifetime income and housing affordability in Singapore" Urban Studies, 48, 1875-1891 (published online Nov 18, 2010). An earlier version is available at the SCAPE Working Paper site: [Link] Abeysinghe, Tilak and Jiaying, Gu (2013) "Estimating bubbles and affordable housing price trends: A study based on Singapore", SCAPE Working Paper 2013/01. [Link] A common measure of housing affordability is to look at what percentage of the house buyers monthly income needs to be allocated to finance the housing cost, mainly the monthly mortgage payment. There are several problems with this measure: (i) given the hump shaped nature of the age-income profile, housing may appear less affordable for young and more affordable for middle aged, (ii) monthly mortgage payment can be reduced simply by increasing the amortization period that can create an elusive sense of improved affordability, and (iii) it is essentially a measure of short run affordability though a house purchase needs long-term considerations. To address these problems, the housing affordability measure used in this exercise is the ratio of house price to lifetime income (or financial wealth), which under some conditions is equivalent to the ratio mortgage payments to permanent income. Therefore, it has the same interpretation as the common measure, what percent of lifetime income needs to be spent to buy a house or what percent of monthly permanent income is used up for mortgage payments. Applying a regression technique to age-income data over different years an age-income profile can be worked out for a representative household in a given birth cohort. Note that the regression technique separates income growth over years into an age effect and a cohort effect. In a growing economy the age-income profile for a given birth cohort typically takes a hump shape peaking at a certain age; this reflects the life-cycle effect. As the economy grows the cohort effect at a given age will result in an upward shift of the age-income profile. Typical cross sectional income data represent a mixture of these two effects and an apparent upward move of an individual in the income ladder may not be nothing more than the life-cycle effect, moving along the same age-income profile instead of shifting to higher income groups. From the constructed age-income profile, the lifetime income at a given age can be computed as the discounted present value of the future income stream. With this information at hand, individual house buyers can easily work out whether the house they are aiming at is affordable in the long run. Housing affordability entails two aspects: (i) housing accessibility and (ii) long-run affordability. Housing accessibility is simply the ratio of mandatory upfront payments to available savings. A ratio less than or equal to unity indicates that the buyer has enough savings for the upfront payments. Long-run affordability is the ratio of house price including transaction cost to lifetime income or financial wealth. The latter is the sum of available savings and the discounted present value of the projected future income stream. A ratio less than or equal to 0.3 or 30% of income is usually taken to indicate an affordable house. Since the focus of this exercise has been largely on first time house buyers, the computations are done for those who are 30 years old. (The average age of marriage for males in Singapore is 30 and that for women is 28.) Income data (including employers CPF contributions) corresponding to the upper end of the percentile groups, 1st-10th , 11th-20th,,81st-90th, which we denote by P10, P20,,P90 respectively, are provided by the Department of Statistics. Savings over age 20-29 are purely savings out of household income and do not include capital gains. For HDB resale flats we computed approximate upfront payments as 10% of the valuation limit (=transaction price COV) plus transaction cost which is about 3% of transaction price plus COV (COV refers to cash over valuation price). Chart 1 below plots the average age-income profile for two income groups (P10 and P80) corresponding to two birth cohorts (1976 and 1981). The chart shows that for the higher income group income peaks around age 58 and then drops slowly. However, for the lowest income group the peaks occur around age 45 and then drops sharply towards age 64. This is likely to result from the fact that low income groups are low-skilled and engage in jobs that require physical labor whereas high income groups earn more out of their human capital. Although human capital tends to appreciate with education and experience, age effect works against physical labor. Although the age-income profile shifts upward for the 1981 cohort, for the lower income group the shift sharply disappears as the person ages. This again seems to reflect the age effect on physical labor and therefore on incomes.
Chart 1: Age-income profile for two income groups (P10, P80) and two birth cohorts (1976, 1981)
[Link]
14/11/2013