Share This
June 2006 Cover Art

Search Finance & Development

Advanced Search
About F&D


Back Issues

Write Us

Copyright Information

Free Email Notification

Receive emails when we post new items of interest to you.

Subscribe or Modify your profile

F and D logo
A quarterly magazine of the IMF
September 2006, Volume 43, Number 3

Gauging the Cost of Aging
John Bryant and Audrey Sonerson

Why population aging is not the main cause of rising government health expenditure in New Zealand

The average person aged 65 or older costs New Zealand's public health system five times as much as the average person under 65. Over the next 50 years, the proportion of the population aged 65 and over is expected to double. Naturally, numbers like this suggest that population aging will put pressure on government health spending in New Zealand and, based on global statistics, in many other countries as well. However, it is reasonable to be skeptical about the close link between population aging and spending pressures.

Econometric studies have produced mixed findings on the relationship between changes in age structure and changes in health expenditure. In fact, the focus on age structure may be misplaced. The first reason is that underlying health status, rather than age itself, explains why old people require more health care—in other words, ill health, rather than age, is what matters for health expenditure. Indeed, the relationship between health status and age is changing because illness and disability rates among old people are falling. The second reason is that "non-demographic" expenditure drivers—such as higher wages for health workers, rising administrative costs, new treatments, and better coverage of the population—outweigh demographic expenditure drivers. Population aging accounts for only a small proportion of growth in health expenditure.

To explore the implications of these issues for New Zealand, the New Zealand Treasury collaborated with the Ministry of Health in 2002–04 to build a health expenditure model. The model shows how the demographic and health profile of the New Zealand population is changing and how these changes create pressures for increased government health expenditure. The results suggest that future demographic changes may be less threatening than is often assumed. But maintaining the current growth rates of government health expenditure would mean substantial increases in the ratio between expenditure and GDP.

The New Zealand model

Our model has two components. The first, the population and health component, was built on a conventional population projection showing the predicted population, by age and sex, up to the year 2051. For a person of a given age and sex, health care spending tends to be higher if that person is disabled or in the last year of life (the "distance to death effect"). Older people require more spending than younger people because older people are more likely to be disabled or close to death. The second component of the model, expenditure, translated the trends in age structure and health status into trends in health expenditure. We focused exclusively on government health care expenditure, which, including long-term care, accounts for about 80 percent of all health expenditure in New Zealand. A key variable was "non-demographic expenditure growth," which captured expenditure growth not attributable to changes in age structure and health. We also made assumptions about labor productivity growth, which allowed us to project trends in expenditure as a proportion of GDP.

Our model followed an approach developed by other researchers (Cutler and Sheiner, 1998; and, Jacobzone, Cambois, and Robine, 2000), which we modified to suit New Zealand data and policy settings. The model differs from standard fiscal models because it allows for the improving health status of the population over time and for the large proportion of lifetime costs associated with the last year of life. Both of these factors should work to reduce demographic pressure on health spending. There were gaps in the available data, which forced us to indirectly estimate the differences in expenditure on the disabled and nondisabled. But experiments confirmed that taking this approach should not have significantly affected our principal results, meaning that the model could prove useful for other countries with similar data gaps.

Most health expenditure models, including ours, include a link from health status to expenditure, but not from expenditure to health status. These models imply that variation in the growth rate for health expenditure does not show up in health trends. Model builders have little choice but to take this approach, since there is no research consensus on the strength of the relationship between health expenditure and health. However, the loss of accuracy is probably small, provided the variation in growth rates is not too large.

Gazing into the crystal ball

We started out by asking what the main drivers of health expenditure were between 1951 and 2000. Our analysis revealed that non-demographic factors—rather than demographic factors—have dominated expenditure growth (see Chart 1). In the 1950s and 1960s, government health outlays grew extremely rapidly, though demographic conditions were, if anything, reducing the need for spending. However, real per capita expenditure fell several times during the 1980s and early 1990s, just as demographic change started to absorb extra expenditure. Spending has increased quickly since the early 1990s, because growth in non-demographic factors has risen to 3–4 percent a year. The growth rates of future health outlays are also likely to hinge predominantly on non-demographic factors. Our model suggests that average demographic expenditure growth in the period 2001–51 will be about 0.5–0.75 percent a year, with the results depending greatly on assumptions about disability trends. The largest increases in growth rates occur around 2015–20, when the baby boomers start to turn 70. Unless growth rates for non-demographic factors become much lower and more stable in the future than they have been in the past, they will overwhelm the effects of aging.

Chart 1. What is the driver?

So what drives the all-important non-demographic expenditure growth? This is a difficult question that we did not try to answer in depth. However, Chart 2 provides a hint. Expenditure growth tends to be high when the economy is doing well, and low when the economy is doing badly. This suggests two possible mechanisms: governments may respond to increases in tax revenues by putting more money into the health sector, and pressure to raise health workers' wages may be highest during periods of rapid economic growth. The projection model used in the New Zealand Treasury's first Statement on the Long-Term Fiscal Position includes an explicit link between health expenditure growth and GDP growth.

Chart 2. Moving together

Second, we asked to what extent improvements in health status could offset the extra cost pressures produced by population aging. To answer this, we constructed two expenditure projections. One projection incorporated plausible improvements in health, while the second did not. Expenditure growth over the period 2002–51 was about one-third lower in the first projection than in the second. This means that plausible improvements in health could offset about one-third of the extra expenditure pressures resulting from population aging. This conclusion is not affected by our assumptions about non-demographic expenditure growth. However, it does depend heavily on our assumptions about future health trends. We assumed that disability rates within each age group would decline at about 0.5 percent a year. Faster declines would produce a larger offset; slower declines would produce a smaller offset.

Third, we asked how expenditure would be distributed among different age groups. We found that even if the health of older people improves, the share of total health outlays going toward older people is projected to grow substantially over the coming decades. In 1951, people aged 65 and over comprised 9 percent of New Zealand's population but, by 2051, will comprise almost 25 percent. It is not surprising that older people's share of total health spending is rising. Nevertheless, the results shown are striking. Our model suggests that by 2051, older people will account for over 60 percent of total health spending (see Chart 3). Their share is predicted to be twice that of the working-age population and 10 times that of children.

Chart 3. More of the pie going to the elderly

Fourth, we asked how quickly expenditure could grow in the future. That is, how fast can expenditure grow, net of what is required to offset population aging? The answer depends on what proportion of national resources New Zealanders are willing to devote to health care. We looked at what would happen if expenditure in 2051 were kept at its 2002 level of 6.2 percent of GDP. We then tried two alternative hypothetical targets of 9 percent of GDP and 12 percent of GDP. Even to keep expenditure at the upper target of 12 percent of GDP would mean holding growth in health expenditure to 2.1 percent a year (see table). In the five years up to 2002, the growth rate was almost 4 percent a year. Clearly, growth rates in health outlays have to fall in the long run, or New Zealanders will have to accept very large proportions of national income going toward health expenditure.

Future expenditure growth
How fast we can go depends on where we are willing to get to.

Expenditure as percent of GDP in 2051
Associated annual growth rate, 2002–51

Source: New Zealand Treasury.
Note: This is the growth rate of health expenditures, net of that required to offset demographic change.

Don't blame it on the elderly

In sum, the number of older New Zealanders will grow rapidly over the coming decades, but future generations of old people may well be healthier than current generations. Even if improvements in health are relatively modest, they could offset about one-third of the extra health care costs imposed by population aging. Once health effects are taken into account, future demographic change is likely to add 0.5 to 0.75 percentages points to annual growth in government health expenditures. This is not trivial, but it is far short of a crisis. The main causes of expenditure growth have been, and will continue to be, non-demographic factors. These have dominated New Zealand's health expenditure growth in the past and are likely to do so in future.

New Zealand's demographic outlook—a likely doubling in the proportion of elderly people by 2050—is similar to that of many countries. Moreover, health care studies in other countries have reached similar conclusions about past and future expenditure growth (Newhouse, 1993; Cutler and Sheiner, 1998; Jacobzone, Cambois, and Robine, 2000). Internationally, aging is likely to exert the same consistent, but manageable, pressure on health expenditure. Thus, if health expenditures reach unsustainable levels in coming decades, it will not be because of population aging.


Bryant, John, Audrey Teasdale, Martin Tobias, Jit Cheung, and Mhairi McHugh, 2004, "Population Aging and Government Health Expenditures in New Zealand," Working Paper No. 04/14 (Wellington: New Zealand Treasury).

Cutler, David M., and Louise Sheiner, 1998, "Demographics and medical care spending: Standard and non-standard effects," NBER Working Paper No. 6866 (Cambridge, Massachusetts: National Bureau of Economic Research).

Jacobzone, Stéphane, Emmanuelle Cambois, and Jean-Marie Robine, 2000, "Is the health of older persons in OECD countries improving fast enough to compensate for population ageing?" OECD Economic Studies No. 30, pp. 149–90 (Paris: OECD).

Newhouse, Joseph. P., 1993, "An iconoclastic view of health cost containment," Health Affairs 12 (Supplement 1) (Bethesda, Maryland: Project Hope).

John Bryant is at the Institute for Population and Social Research, Mahidol University, Thailand, and was formerly a Senior Analyst at the New Zealand Treasury. Audrey Sonerson is a Senior Analyst at the New Zealand Treasury.