(Reprinted from HKCER Letters, Vol. 42, January, 1997)


Poverty and Income Disparity in Hong Kong

Lui Hon-Kwong



In the last few months, various organizations have released studies on poverty and income inequality in Hong Kong. These studies have aroused public concern. Among the reports, one published jointly by the Hong Kong Council of Social Service (HKCSS) and Oxfam Hong Kong in December 1996 has become the center of attention. This report states that there were 640,000 people living in abject poverty in Hong Kong in 1994/95. Another study suggests that there were 850,000 people with household incomes below the poverty level in 1996.

These alarming figures caused widespread concern and received a great deal of media coverage. As a result, policy recommendations intended to address the poverty problem have been put forth by scholars, politicians, and pressure groups. However, there are reasons to be skeptical about the validity and reliability of the estimated numbers of people living in abject poverty in Hong Kong. A slight change in the assumptions made in the measurement method can easily lead to very different numbers.

In December 1996 the Hong Kong government released summary results of the 1996 Population By-Census, which added extra fuel to the debate about the poor. According to these results, the official Gini coefficient of household income distribution rose sharply from 0.476 in 1991 to 0.518 in 1996. The Gini coefficient, which has a value between zero and one, is frequently used as an income inequality index. A larger coefficient represents a higher degree of income disparity. The results of the Population By-Census thus suggest that income disparity has increased considerably in the past five years. Many people take this to mean that poverty is on the rise in Hong Kong, as well.

Poverty and income inequality are two distinct concepts, however. Poverty may become less severe in an economy in which income inequality is increasing. This is often difficult for the general public to grasp. This article attempts to clearly explain the difference between poverty and income inequality. It also discusses the definition of abject poverty and issues related to income distribution in Hong Kong.

Poverty, Abject Poverty, and Relative Poverty

There are no universally accepted definitions of poverty and abject poverty. Broadly speaking, poverty can be defined in absolute and relative terms. If we define a person in poverty as one whose income allows him or her to buy the minimum amount of basic necessities for living, then a person in absolute poverty has an income which does not allow him or her to secure these minimum basic necessities. The definitions of poverty and abject poverty are thus built upon the concept of a subsistence level of income. However, it is not clear what commodities or how much of them should be regarded as basic necessities.

On the other hand, relative poverty is defined in relation to a predetermined amount of money. For example, an interest group has proposed that anyone with personal income less than half that of the median income of the working population should be classified as poor. Again, there is a question of why the median income should be used as the predetermined amount, and why 50 percent of that amount should be used as an indication of poverty.

HKCSS/Oxfam Study and the Income Elasticity Technique

The study of HKCSS and Oxfam classifies households with income insufficient to pay for basic and necessary food expenditures as those in abject poverty. It estimates that a minimum of around $600 per person per month is needed to purchase basic and necessary food . I have serious reservations about the way in which the study arrives at this magic number. The analysis is based on data collected from Household Expenditure Survey conducted by the Census & Statistics Department in 1994/95. Due to confidentiality problem, researchers carrying out the study were not allowed to access the raw data. Instead they were provided with special tabulations by the Census & Statistics Department based on their ad hoc requests.

The researchers involved in the report study the expenditure patterns of eight expenditure groups in five household-size groups. The expenditure groups are arranged in ascending order of total spending. For each household-size group, they plot the expenditure share on food against the expenditure groups. They then join the eight data points as a line chart and define the subsistence level as the turning point of the plotted line.

The theoretical framework of the estimation method described above is adapted from the quantity income elasticity technique. The aim of this analysis is to try to determine the inflection point at which the proportion of income spent on basic necessities begins to decline. The income level at the inflection point can be looked upon as the poverty threshold. People whose income level places them below the poverty threshold are living in poverty. The idea is that when people live in poverty, if they get additional income, they will spend all of it on necessities. When income increases beyond a certain point (the poverty threshold), the proportion of the budget devoted to necessities decreases, and people begin to switch to higher-quality commodities rather than consuming more lower-quality basic necessities. In short, we observe a clear quality-quantity trade-off when household income reaches the poverty threshold. However, as the Household Expenditure Survey did not collect information on quantities purchased, HKCSS and Oxfam's report uses the dollar value of expenditures as a substitute for quantity of physical units.

Reliability of the HKCSS/Oxfam Study

As was discussed earlier, there is no universally accepted definition of poverty. The income elasticity technique is only one of many possible ways to determine the poverty threshold. Assuming we agree that this approach is appropriate, the HKCSS/ Oxfam report is still worthy of serious doubts.

First, it is incorrect to try to locate a point of inflection from a graph consisting of no more than eight discrete data points. Second, even if we assume that the line chart in the report is continuous, the researchers’ method is still questionable. It is inappropriate to subjectively choose a turning point from a graph and take it as a point of inflection. The essence of the income elasticity approach is to study the change in the marginal propensity to consume basic food necessities. An inflection point indicates that the marginal propensity to consume slows in relation to income, but a turning point represents a fundamental change in marginal propensity caused by factors other than income, such as changes in tastes and in goods available. As such, a turning point cannot be used as a substitute for a point of inflection.

Furthermore, each expenditure group represents not less than 5 percent of households in Hong Kong. It is likely that a point of inflection is embedded inside a discrete (aggregated) data point. If we really want to study the change in marginal propensity to consume, we should use the raw data (raw household data is highly preferable), to plot an Engel expenditure curve (with quantity purchased as a function of income) instead of analyzing a few highly aggregated data points.

Finally, the report (though not by choice) uses expenditure data as a substitute for quantity. Unfortunately, expenditure data alone does not indicate whether people spend more for better quality or for a larger quantity. For those living in abject poverty, the quantity of a good consumed is more important than its quality. As income rises, quality becomes increasingly important. An inflection point shows where such a trade-off takes place. However, the quantity consumed is not used in the study. As a result of these problems, I am reluctant to believe that the estimated number of people living in abject poverty reflects the real picture.

Rising Income Inequality

In 1996 the Gini coefficient of household income distribution in Hong Kong stood at 0.518, a record high level. Twenty years ago the estimated coefficient was 0.43, and it rose only gradually to 0.476 in 1991. Some people and pressure groups argue that higher income inequality also means a higher level of poverty. This misconception is rather deep-rooted. To clarify the misconception, consider the following three cases.

Case 1: Suppose we adopt the absolute poverty concept and assume that the poverty level is $1,000 per person. In an economy, 99 percent of the population all receive a fixed income of $1,001, and 1 percent earn an average of $100,000. In such an economy, since none of the residents earn an income below the poverty line, no one is classified as poor. However, as 1 percent of the population account for half the total income, the distribution of income is highly uneven.

Case 2: Suppose the poverty line is still $1,000, but all residents in the economy earn a fixed income of $999. Since everyone takes an equal share of the total income, the economy is in a state of perfect equality. However, as the income of all residents is below the poverty line, the whole population is in poverty.

Case 3: Suppose we adopt the relative poverty concept and use 50 percent of the median income as the poverty line. In the economy of Case 1 above, in which 99 percent of the population receive a fixed income of $1,001 and 1 percent earn an average of $100,000, the median income is $1,001. Clearly no one in this economy is regarded as poor, as no one receives an income below $1,001. However, income is still very uneven.

These examples clearly show that there is no direct relationship between poverty and income disparity. Rising income inequality does not necessarily indicate increasing poverty, and vice versa.

Factors for Change in Income Dispersion

In analysing income distribution, many commentators and researchers use the Gini coefficient as the main measure of income disparity. When comparing the Gini coefficients for 1991 and 1996, they jump to the conclusion that Hong Kong has experienced a significant increase in income inequality. However, there have been other significant changes in the composition of our population in recent years. For example, the impact on income distribution of the constant and heavy influx of immigrants from Mainland China into Hong Kong should not be underestimated. At present, Hong Kong accepts 150 Chinese immigrants daily, or 54,750 per annum. The average economic conditions of these new immigrants do not compare favorably with those of the average local resident. Even if these immigrants have the same productivity level as do local residents, their incomes tend to be lower when they first enter the local labor market. It takes time for them to adapt to the new working environment and to acquire knowledge pertaining to their local jobs. A study completed in 1995 indicates that if we remove new Mainland immigrants from the population, there will be less income dispersion.

Upward social mobility is another major factor to be considered in analyzing changes in income inequality over time. Households can move up and down the social ladder within a society. The increased accessibility of education at all levels in the last two decades has helped young people move up to higher-income groups. Tertiary education is now available to children from all income groups. Human capital theory suggests that educated workers receive higher wages than do uneducated workers, other factors being equal.

In addition, the income level of an individual varies across the different stages in his or her life. Suppose that in a given economy, all people have equal earning ability, and income differential is purely due to differences in working experience. If we then study lifetime income distribution, this economy shows perfect equality. However, if we measure income distribution at any time point, we observe a certain degree of income disparity due to the fact that people are in different stages of their life.

Obviously, all the factors mentioned above are not reflected in changes in the Gini coefficient. The recent sharp increase in this coefficient might be due to a number of causes. Widening income dispersion and increasing poverty are only symptoms, and we need a careful diagnosis of the actual situation. Policy recommendations have been put forth by various interest groups before much is known about the real causes. Implementing these policies would be like prescribing drugstore medicine to an uninformed and undiagnosed patient, with possibly disastrous results. Further and more detailed study is necessary before effective policies can be put into practice.


The perception that the poverty condition in Hong Kong has become severe is ill grounded based on the results of HKCSS and Oxfam's joint study. The accuracy and reliability of this study remains open as it has committed errors in methodology. Besides, the reported increase in Gini coefficients is not necessarily reflecting the whole picture of the current situation of income dispersion because of its deficiency. In order to find out the truth, we need better, more up-to-date, and more comprehensive data. Data collected by the Hong Kong government in general, and the Census & Statistics Department in particular, should be made more accessible to the general public. Academics and other interest groups are always willing to provide free and voluntary data analysis. I understand that due to legal restrictions, only census officers can access raw data collected by the Census & Statistics Department. To bypass this technical problem, interested researchers can be appointed as temporary census officers by the government and have to observe all legal restrictions in relation to data confidentiality. Statistical authorities elsewhere often employ full-time researchers to analyse data these authorities collect.

Dr. Lui Hon-kwong is assistant professor in the Department of Marketing and International Business, Lingnan College.


| Index | Research Projects | HKCER Letters |
| Speaker Program / Conference | Index of Economic Freedom |
The Hong Kong Centre for Economic Research
School of Economics and Finance
The University of Hong Kong
Phone: (852) 2547-8313 Fax: (852) 2548-6319
email: hkcer@econ.hku.hk