(This essay was published in Hong Kong Economic Journal on 19 March 2014)
Last week I showed that half of government expenditure today is spent on four areas – housing, education, health and social welfare – compared to one-third 30 years ago. The growth in spending has been the result of economic growth, demographic change, and organized political pressure. Barring unforeseen circumstances, these underlying factors are expected to continue to drive public spending higher in all four areas.
I have singled out these four areas because how we manage public spending on them will determine to a large degree what kind of society Hong Kong will become in the future. A number of commentators and voices in society are wary of the Financial Secretary’s attempt to warn the public of runaway structural fiscal budget deficits. Some have lashed out at infrastructure spending they deem to be less essential then spending on social welfare, health care or other areas. Their voices are sometimes joined by the environmental lobbies, some of whom are totally opposed to development altogether. Then there are those with vested interests on both sides of the divide: some want more infrastructure spending and others oppose it.
How do we decide what should or should not be done? The issues at stake are more than just a matter of balancing the fiscal budget as an accounting exercise. The choices we make will impact whether Hong Kong will remain, in the long-term, a free and just society with a vibrant economy. Hong Kong’s ability to sustain economic prosperity and political stability depends on how we respond to the continuous growth in the demand for housing, education, health and social welfare spending.
Unlike episodic, one-time increases in infrastructure spending, these four areas require sustained increases in recurrent public spending in order to meet society’s rising demands. Meeting such demands through public spending alone will inevitably lead to a larger and more interventionist government, higher taxes, and the curtailment of some of the freedoms and liberties we are accustomed to. Life will not be the same anymore.
Some in Hong Kong may welcome this, but if society is to transform itself and reinvent a brave new world, we owe it to ourselves to have a serious political debate first. I am sure the central authorities in Beijing will be as interested in this debate as the people of Hong Kong. The basis for this debate must be the projected rising demand and expected costs in these four areas.
Before considering the possible choices and what is to be done, we need to first look at (1) the relationship between economic growth, demographic change, and public spending growth; and (2) why the growth in demand for public spending in each of these four areas will not be the same in the future. I shall start with health expenditure – the greatest challenge posed by an ageing population for containing fiscal budget deficits.
Salient Features of Health Spending
We know that spending on health care increases with (1) the number of persons in an economy, (2) the proportion of elderly persons in the population, (3) the average income per person, and (4) the cost of health care services. So a growing economy with an ageing population will spend more on health care. Health is a form of human capital. A healthier population can become economically more productive, for example, by having fewer hours lost due to sickness. For this reason, spending on health is not entirely consumption spending, but partly investment spending that enhances economic productivity.
Increasing public spending on health care also potentially enhances the capacity of the medical and health industry to export medical services. This contributes to economic growth and adds to public revenues in the long run. One notable form of public spending that is classified under education expenditures is the training of medical and health care personnel. These investments not only enhance the capacity to provide health care services, but also research and development activities that create demand for skilled jobs in the bio-medical sector and enhance long-term economic growth.
Increasing health-related public spending, therefore, affects the public budget on both sides of the equation. It raises both expenditures and revenues. The extent to which public revenues will rise is not clear. In principle it can be empirically estimated, although I am not aware that anyone has seriously done this kind of calculation for Hong Kong. It is highly likely that in the short run the effect of increasing public spending on health will be to enlarge the fiscal deficit, and that the offsetting effects will appear in the long run.
Health spending also varies enormously from person to person. Some people stay healthy most of their lives and die suddenly without incurring a great deal of medical and health expenses over their entire lives. Others have chronic diseases that require expensive treatments over long periods of time. The incidence of disease is not highly correlated with ability to pay. This implies health care spending has to be provided in part through some form of social insurance scheme, perhaps through public subsidies. Our current public health service provision is a form of heavily subsidized social insurance modeled largely after the British National Health System, but with a lot less competition among hospitals and clinics for containing costs and enhancing efficiency.
Public spending on health will increase rapidly in the future not because of rising population numbers but ageing. According to the Hospital Authority’s (HA’s) published health care services costs, treating elderly populations is a lot more expensive than other age groups (see Table 1). For example, it cost the HA on average $3.1 million per annum to treat a population of 1000 persons aged 15-64 in 2009-11, but $11.4 million to treat those aged 65-74 and $25.3 million to treat those aged 75 or above. This meant it was 8.15 times more expensive to treat those aged 75 or above than 15-64 year olds, and 3.68 times more expensive to treat those aged 65-74. Children age 0-14 were also slightly more expensive to treat by 1.05 times. These figures averaged out at $5.1 million per annum for treating 1000 persons.
Table 1: Health Care Costs per 1000 Population by Age Group 2009-2011 ($millions)
Demographic Drivers of Future Health Spending
These cost figures can be used to construct the underlying demographic demand for health care services. We can then construct a “health care cost standardized population” that would multiply the number of persons aged 65-74 by 3.68 times and the number of persons aged 75 or above by 8.15 times. This “standardized population” would embody the increase in demographic demand for health care due to changes in both population numbers and the age structure. In Figure 1 we use the projection figures of both the Census and Statistics Department and the United Nations to make population and “standardized population” projections up to years 2041 and 2100, respectively.
Figure 1: Actual and Projections of Population Numbers and Health Care Cost Standardized Population Numbers 1950-2100
Note: Projections based on figures provided by C&SD and UN.
Source : HKSAR Census and Statistics Department, Hospital Authority & UN Population Projection
Table 2 shows the population and “standardized population” numbers for various years in 1989-2100. The population in the years 2013, 2041, and 2100 is estimated or projected to be 7.18, 8.17, and 10.35 million. But the “standardized population” is found to be 12.29, 22.08, and 27.61 million. These latter numbers provide a much clearer idea of the demographic pressure on health expenditures.
Table 2: Population Numbers, Health Care Cost Standardized Population Numbers, and Growth Rates 1989-2100
Note: Projections based on figures provided by C&SD and UN.
Source : HKSAR Census and Statistics Department, Hospital Authority & UN Population Projection
In 2013, health care demand from the “standardized population” is 1.71 times higher than the population. In 2041 and 2100 it will be 2.61 and 2.67 times higher, respectively. The implied annual percentage growth rates of the population in 2013-2041 and 2041-2100 are 0.59% and 0.62%; those for the “standardized population” are 2.11% and 1.38%, respectively.
Clearly, the “standardized population” growth rates during 2013-2041 are on average significantly higher than the population growth rates, even when compared with the past, i.e., in the period 1989-2013. This clearly implies that the demographic demand for health expenditure will be quite overwhelming in the coming three decades. Fortunately projections for the second-half of the century will moderate with little difference in the projected population and “standardized population” growth rates.
Forecasting Future Income Growth
Health spending will increase in the future as per capita GDP rises, but by how much? Two factors are relevant. First, what is the income elasticity of demand for health? Second, by how much will GDP grow?
The income elasticity is a concept used by economists to describe the quantitative demand response of a good or service to changes in income. Estimates of the income elasticity of demand for health among the Organization of Economic Cooperation and Development (OECD) countries is around 0.8 when the US is excluded. We shall use this number to forecast our future “baseline scenario” of the demand for health because Hong Kong’s income level is about the same as the member countries in the OECD. The US is excluded because it has a uniquely higher income elasticity of demand due to its market insurance system, which drives demand for health services and produces higher levels of health spending.
We also produce a second “high growth scenario” forecast for health care by assuming an income elasticity of demand of 1 if the government is unable to resist the rising political pressure to increase health care spending and is unable to reform the present essentially-free public health care services delivery system.
Table 3 presents some figures on the relationship between real GDP and employment growth in the past and for the future. Real GDP per employed persons was 4% per year on average in 1961-2013. This is the labor productivity of the employed population for the past half-century. During 1961-97 it was 4.7% and in 2003-13 it fell to 3.1 %. Despite this drop, I have assumed that a 4% rate of productivity can be sustained into the future.
Table 3: Annual Percentage Growth of Real GDP per Employed Person, Real GDP per capita, and Real GDP
Note: Future employed population is defined as 73% of the population aged 20-64, which is the average rate in 2002-2012.
Source : HKSAR Census and Statistics Department
I believe there is a very reasonable chance that our productivity will not slow down as in other countries because our rapidly ageing population will trigger more investments in health-driven increases in the productivity of our working population. Similar investments in health around the world will extend life expectancies and by implication also lead to longer working lives in Hong Kong. Therefore the assumption of an employed population bounded between the ages 20-64 is likely to underestimate productivity growth that should be offsetted to produce more reliable forecasts.
Effects of Rising Health Costs
The final element determining rising health expenditures is the cost of providing medical and health services. A variety of price indices are available, but for the purpose of estimating the fiscal impact of rising health expenditure the most logical one to use is the medical services component in GDP. Figure 2 shows its price relative to the real GDP deflator. It has risen from 57.0 in 1989/90 to 100.0 in 2010/11. This is equivalent to an average annual rate of increase of 2.45%, which is roughly comparable to the 2.13% increase in the government consumption expenditure deflator relative to the GDP deflator. This is expected since the major component of the medical services deflator and the government consumption expenditure deflator is manpower costs.
Figure 2: Relative Price Index of Medical and Health Services and Share of Public Sector Health Expenditure
Note: *Relative Price of Health and Medical Services =Private Consumption Expenditure Deflator for Medical Services/GDP deflator (Index 2010=100)
Source : HKSAR Census and Statistics Department, Food & Health Bureau
How does the rising relative cost of medical and health services affect demand? One would expect it to decrease the quantity of health care services demanded by the population, so changes in the total expenditure would depend on the price elasticity of demand. If the price elasticity is equal to 1, then the total spending will be unchanged because a 1% decline in health care will be offset by a 1% increase in its price.
We assume, again relying on OECD calculations and excluding US estimates, that the price elasticity of demand is -0.4. In other words, the price elasticity is relatively inelastic; therefore, price increases in medical and health services will trigger increases in expenditure.
I shall also assume for the purpose of forecasting future health expenditures that the relative price of medical and health services will increase at 2.5% per year, which is the average rate in the period 1989-2011. How fast it will actually rise will depend critically on government policy to influence the salaries and incomes of medical and health care personnel. If we train more doctors and other health service providers and admit more foreign-trained ones, then the increase in costs will be more moderate. This is the single most important lever the government can pull to control future medical and health care costs, but its ability and willingness to do so have been quite limited to date.
Long-term Forecasts of Health Expenditure
Finally, for total health expenditure I use the domestic health accounts prepared by the government. Since these figures cover total health expenditure in both the private and public sectors, one has to determine the share of public expenditures in total future health spending to ascertain the impact on the fiscal budget. In Figure 2 we can see that the share of public health expenditures rose from 39.4% in 1989/90 to a peak of 57.7% in 2003/04 before falling to 48.7% in 2010/11. I shall assume that in the future government will be able to keep the public share of health expenditure roughly at 50% of total health spending.
The assumptions I have made up to this point for the “baseline scenario” are very simple: (1) population growth up to 2041 according to the Census and Statistics Department projections and to 2100 according to the United Nations projections, (2) growth in real GDP per working person at 4% per annum, (3) the income and price elasticities assumed to be 0.8 and -0.4, and (4) medical and health care prices to rise at 2.5% per year in real terms. For the “high growth” scenario, I assume the income elasticity will be 1.0 instead.
Although there is always room for some debate, I believe my assumptions are quite reasonable. They can be used to forecast future total health care expenditure in a straightforward manner using a spreadsheet model. Table 4 presents the estimated public expenditure forecasts as a percentage of GDP.
Table 4: Forecasts of Public Health Expenditures as a Percentage of GDP to 2041 and 2100
Note: Forecasts to 2041 based on CS&D population projections and those to 2100 based on UN ones; real GDP/working person is 4% each year; medical relative prices rise at 2.5% each year; and public health expenditure is 50% of total.
In 2011, 2.56% of GDP was spent on health in the public sector. This will increase to 5.65% in 2041, an increase of 121% under the “baseline scenario”. The corresponding figures for 2100 are 8.8% of GDP and an increase of 244% above the 2011 baseline public sector health expenditure figure. These are very large increases. Of course the high growth figures are even more alarming especially over the long-term. Such increases in public spending on health expenditures alone cannot be accommodated under the present fiscal arrangements without severely sacrificing the quality of health care. Politically no government can even contemplate such an outcome given the power of the elderly vote.
Health is not the only issue that will be seriously impacting our future fiscal budget. Next week I will begin discussions of how to manage public spending for an ageing population.
de la Maisonneuve, C. and J. Oliveira Martins(2013), “A Projection Method for Public Health and Long-Term Care Expenditures”, OECD Economics Department Working Papers, No. 1048, OECD Publishing
Thirteenth essay in the series on Rekindling Hong Kong’s Magic