Introduction

There are three parts in my presentation. First I will discuss briefly about the fiscal implications of the CSSA scheme. Then I will try to explain why CSSA payments increased sharply after 1993. Finally, I will present some tentative proposals to contain cost, and I emphasize the tentative nature of my proposals.

Fiscal Implications

Figure 1 shows that the CSSA total spending from 1975 (Public Assistance (PA) before 1993) up to now. In 1992/1993, when the spending started to increase exponentially, it was around 1.4 billion. In 1999 it was 13 billion, a 9-fold increase in 7 years.

The annual rate of increase of the CSSA spending starting to rise sharply in 92/93 and went up to 73.5 percent in 1993/94 (Figure 2). Even though it dropped to around 40 percent in the following year, it is still a huge rate of increase. Since then the rate has fluctuated around the same level. It was still 38 percent in the last fiscal year.

Figure 3 illustrates the financial implications of the CSSA scheme. The ratio of CSSA spending to salaries tax revenue increased rapidly over the years. By now well over 50 percent of Hong Kong's salaries tax is used to finance the CSSA scheme. If we use the total government revenue as a yardstick, 8 percent of the total government revenue is spent on CSSA.

If we add in the Social Security Allowance (SSA) which is the non means-tested part of the welfare program, then the percentage goes up to 75 percent (Figure 4). If we further include costs of the welfare system, like staff costs and the costs of all the subvented programmes, then the total welfare payment will exceed the total salaries tax. The salaries tax is not able to pay for the welfare expenditure in Hong Kong.

Sharp Increase After 1993

There are two components contributing to the sharp increase in CSSA spending after 1993, the number of recipients and the average amount of payment to each recipient. These two components are related. Let us first look at the number of recipients. Figure 5 shows that the number of CSSA cases had been relatively stable from 1975 up to the early 1990s. It increased sharply around 1993, from 80,000 cases to about 230,000 cases now. There has been a slight drop since the end of 1999. Even though the number of CSSA cases increased by threefold from 1992 to 1999, it cannot fully account for the 9-fold increase in CSSA spending. The other component that contributes to the large increase is the huge increase in the average CSSA payment.

Before turning to the average payment, let us look at the number of CSSA cases more closely. Figure 6 shows the four major components from 1990 to 2000, old age, single parent family, unemployed and low earnings cases. Except for a drop in the number of unemployed cases between 1999 and 2000, all four components show an increase over the years, but at different rates. Figure 7 shows the rates of increase of CSSA cases. The sharpest increase is in the number of unemployed cases which went up to 85 percent in the year 1995/1996 from 35 percent the year before. There was also a sharp rise in the number of single-parent cases. The number of elderly cases also increased over time but the rates of increase were more moderate. As a result of the differential rates of increase, the ratio of old age cases dropped to half of the total, while the other three components became relatively more important in proportion (Figure 8).

The rapid increase in the number of unemployed CSSA cases since 1993 has nothing to do with the unemployment rate. As a matter of fact when unemployed CSSA cases increased rapidly, in the mid 1990s, there was essentially full employment in Hong Kong. Figure 9 shows the unemployment rates in the past 25 years which show three peaks: 9 percent in 1975, 5 percent in 1983 and 6 percent in 1999. Matching the profile of unemployment rate (Figure 9) with the number of unemployed CSSA cases (Figure 10) shows that there is no relation between the two profiles. The number of unemployed CSSA cases simply went up after 1993 when the labour market was still rather tight, whereas in the previous two peak periods of unemployment (1975 and 1983), the number of CSSA cases was moderate.

How can we explain the increase in the number of unemployed CSSA cases if the underlying unemployment rate has an insignificant impact? Could the increase in the number of unemployed CSSA cases be accounted for by the relaxation of the asset limits of the eligibility requirement for receiving unemployed CSSA? Figure 11 shows that the asset limit for single able-bodied adults has been raised steadily and moderately over time. In 1995 the asset limit was increased somewhat faster than the trend. The increase is partly to offset the effect of tightening of the eligibility requirement in that year; henceforth CSSA recipients were no longer allowed to receive rental income from their property. There has been a reduction in the asset limits for single able-bodied adults after the review in June 1999. Over the years, the asset limits have basically been increased gradually to take into account inflation. Hence, the large increase in the number of unemployed cases since 1993 cannot be due to changes in asset limits which expand the number of eligible persons.

Figure 12 shows the take-up rate of unemployed CSSA cases. In the 80s and early 90s, the take-up rate is one to three percent, that is out of one hundred unemployed persons, only one to three persons would receive CSSA. The take-up rate shot up to 21 percent in 1997 which incidentally was a year of full employment. (The figure refers to the end of March in 1997 before the financial crisis.) In that year one out of five unemployed received CSSA, whereas historically only 3 out of 100 unemployed workers were recipients. Over time, there has not been significant changes in eligibility requirements that could account for this large increase in the take-up rate. Something drastic has obviously happened.

In fact the same pattern shows up in the elderly cases even though it is not as dramatic. The dotted line in Figure 13 shows the percentage of the population that is aged 60 or above from 1975 to 1999. It clearly shows that the population is aging. The solid line shows the take-up rate of old age CSSA cases. The take-up rate had been stable at around 6 percent in the 1980's. It started to increase rapidly after 1993 to reach around 13 percent in 1999.

It should be stressed that among the potentially eligible recipients -- the elderly and the unemployed, the take-up rate has increased sharply. What caused such a dramatic increase in the take-up rate? Figure 14 reveals the answer. There was a substantial increase in the CSSA standard rate in 1993, and again in 1996. (Separate standard rate of payment applies to different categories of recipients, such as single person able-bodied, elderly, disabled, etc.) It shows that the standard rates of 5 representative recipient families increased sharply in 1993. Figure 15 uses the median wage as a yardstick for comparison. The standard rate of a 4-person family went up steadily after 1993 and reached 70 percent of the median wage in 1996. Before 1993 it had been around 40 percent of the median wage. For a single-person family, the pattern is similar but less dramatic. The standard rate for a single-person family is around 25 percent of the median wage.

It should be noted that besides the standard rate, CSSA recipients are eligible for special allowances, like rental assistance, tuition fee and lunch subsidy for school children, dental reimbursement etc. The total amount received by a CSSA family includes the standard rate payment and the special allowances. Figure 16 shows the average monthly total payment for two categories of recipient families, as a percentage of the median wage. While the standard rate for a four-person family is around 70 percent of the median wage, the average actual amount which includes the special allowances for a four-person family is above 100 percent. In other words a 4-person-family on the average is receiving CSSA payment that is more than the income of half of the working population. The CSSA family members do not work and yet they are better off than a family of four whose head is working and receiving the median wage. This comparison should give an indication why there has been such a sharp increase in the number of CSSA cases starting from 1993. What happened in 1993 is that the Hong Kong Government increased the CSSA payment substantially, triggering a sharp increase in the take-up rate. Before 1993 a large number of people who were eligible for CSSA chose not to apply for the scheme because the payment was not that attractive, and they had other means to cope. Since 1993, due to the large increase in the payment rate, CSSA became very attractive. Eligible persons responded to the economic incentive and found it worthwhile to apply for the scheme.

The increase in the take-up rate of the elderly can be interpreted as a response to the increase in CSSA payment rate. Before 1993, many elderly people eligible for CSSA did not join the scheme and the take-up rate was low. They were either supported by their children or they had other means to cope. After the sharp increase in the CSSA payment in 1993, the children might tell their parents to apply for CSSA because the payment became more attractive, and they might withdraw their financial support for their elderly parents. As a result, there was a sharp increase in the take-up rate. The government was substituting CSSA support for family support. The same interpretation can be applied also to the unemployed. After the sharp increase in the standard rate, more of the unemployed found it attractive to apply for the CSSA rather than to look for a job.

The increase in the standard rate may cause an increase in the number of fraudulent cases, but the sharp increase in the number of CSSA cases cannot possibly be explained by fraud. Fraudulent cases could only account for a very small portion of the increase. The sharp increase in CSSA cases also cannot be explained by social workers working harder to help eligible persons to apply for CSSA. There is no reason to believe that social workers suddenly worked very much harder in 1993, the year when the number of CSSA cases took off. We can only conclude that the rapid increase in the number of CSSA cases is a response to the large increase in the CSSA payment after 1993. A large number of eligible persons, formerly foregoing CSSA and coping with other means, are now on CSSA. CSSA recipients respond to economic incentives just like everybody else.

Contaning costs

What can be done to contain costs? Several proposals are worth considering. Experience indicates that when the average CSSA payment was around 80 percent of the median wage in the late 1980s, the take-up rate was stable. However, when the payment was increased to above 100 percent of the median wage, there was a sharp increase in the take-up rate. Total payment and take-up rate are empirically related. To stabilize the take-up rate at a manageable level, CSSA payment should be linked to the median wage at, say, the 80 percent level.

Another possible measure to contain costs is to take into account a CSSA recipient's accrued benefits of the mandatory provident fund (MPF) in calculating CSSA payment. Consider a person earning $ 10,000 a month. The combined monthly MPF contribution from the employer and the employee will be $ 1,000, After 3 years, the MPF balance of this employed person will be $ 36,000, plus investment returns. This amount exceeds the current CSSA asset limit for a single person. If MPF accrued benefits are counted as part of asset, he will not be eligible for CSSA if he becomes unemployed subsequently. How can this unemployed person survive without any current income? We should consider amending the MPF Ordinance to allow unemployed workers to draw on their own MPF accrued benefits while they are unemployed. Basically it is asking the unemployed to internalize their welfare benefits as much as possible before resorting to public assistance. Unemployed persons should first make use of their own resources including their MPF accrued benefits before being supported by public money.

But if unemployed persons run down their MPF balance, how are they going to support themselves when they retire? There are two possibilities. One possibility is that their unemployment is short-term. After a short period of unemployment, they will find a job and start to accumulate their MPF balances again. Young workers have 20 to 30 years to accumulate for their retirement. Short spells of unemployment during their work career will not hurt their long-term retirement protection by much. The government will save a lot of money because it does not have to pay CSSA for these short unemployed spells. The second possibility is that their unemployment turns out to be long-term or they are approaching retirement and there is no time to accumulate the MPF benefits. If that is the case, their MPF balance will be run down and they will become eligible for CSSA. In any case they will have to be supported by CSSA after retirement. CSSA will continue to offer a safety net. To summarize, allowing unemployed workers to draw on their MPF balance before applying for CSSA will save public money, but this will require amending the MPF Ordinance.

As for the retirees, it is logical to count their accrued MPF benefits as part of their assets. The asset limit for old age CSSA should take into account MPF balances.

The third proposal is to provide greater incentive for work. The current CSSA scheme rules allow about $ 1,000 as disregarded income if the CSSA recipients work. The disregarded income can be increased so that the more they work, the more income they will get. CSSA recipients should be allowed to keep more of their earned income while receiving partial CSSA support up to a cutoff point when they will be on their own. More thought should be given to build in greater incentive for CSSA recipients to work.

The fourth point deals with single-parent families. The Government has to provide more child care and other support, and to help divorced women collect court mandated alimony from their ex-husbands. This will reduce their reliance on CSSA.

These proposals are tentative. The review of the CSSA scheme by the government in June 1999 is a step in the right direction. The standard rate for able-bodied recipients was reduced by about 10 percent to make it less attractive. A self-reliance requirement was introduced to make it more costly for the unemployed recipients to remain on CSSA. Since the review, there has been a decrease in the number of CSSA cases. It is still too early to say whether these factors can account for the reduction. We have to monitor the number of unemployed CSSA cases further before we can be sure of the effectiveness of the measures introduced in the review. The initial signs are, however, encouraging.

By Liu Pak Wai