 |
|
|
|
|
Global Demographic Change and Demand for Food in Australia by Ron Duncan, Qun Shi and Rod Tyers
January 2005
RIRDC Publication No 05/014 RIRDC Project No ANU-51A
These demographic changes have implications for many facets of economic life, including work force structure, savings and investment, retirement incomes, health expenditures, and consumption of goods and services. The changes are even likely to have implications for the quantity of food consumed, the types of food consumed, and the ways in which food is consumed.
How is the slowing of population growth and the aging of the global population likely to affect the demand for food in Australia (both domestic and export demand)? Market expectations for Australian food exporters have been based on expectations of continuing large increases in populations in most of the importing countries. These expectations about increases in the world’s population and the implications for food consumption need reappraisal. In this paper we focus on the projections of the size and global distribution of the world’s population and the impact of population forecast uncertainties on projections of economic behaviour, and thereby on projections of the composition of demand facing Australian commodity exporters.
There are numerous sources of demographic uncertainty, particularly fertility, but also mortality, migration and the sex ratio at birth, some of which have not been well understood by researchers. In most countries, infant mortality fell throughout the course of the last century and adult life expectancy increased. There have also been sharp declines in fertility, first in developed countries and recently in many developing countries. Before this present century is half over, populations in Japan and some European countries are likely to be smaller than they were in 1990, with these declines in total population being preceded by declines in the number of people of working age.
The extent to which demographic change has been imperfectly forecast is indicated by the size and persistence of errors in the UN Population Division’s global population projections since the 1950s.
Errors in forecasts of country and regional populations have been as large as 17 per cent over 30 years.
The latest global population projections by the UN Population Division may be seriously flawed because of the assumptions that are made about the future paths of fertility rates. Given their population size, what happens to fertility rates in China and India will have a large impact on the size of the global population. In the case of China the UN has assumed an increase in the total fertility rate (TFR) from the present 1.8 to 1.9 by 2010-15 and thereafter. The UN assumes that the TFR in Bangladesh and India will continue to decline until 2025 and then hold at the replacement rate of 2.1.
There is no basis in the experience of other countries to support such assumptions.
There is no sign of a turnaround in fertility rates even in those countries that have already reached very low levels. It is obvious that there is limited understanding of the reasons for the fertility rate decline and especially for the persistence of rates below the replacement level. Thus, the assumptions by the UN of a turn-around in fertility rates in higher-income countries and a slowing of the decline in fertility rates in lower-income countries are highly doubtful. Maintenance of current fertility rates or a continuation of the existing downward trends would appear to be more appropriate assumptions.
Demographers have been working to define the extent of uncertainty that should be associated with the UN’s population projections and in the process to provide probabilistic forecasts based on this uncertainty. The range of uncertainty of the resulting forecasts is quite large. For the global population there is a 60 per cent probability range of around one billion (from 7.6 to 8.6 billion). This indicates a 40 per cent chance that the global population in 2030 will be outside the 7.6 to 8.6 billion—an error range of almost 50 per cent of the projected increment to the global population. The 95 per cent probability range of the forecast is 2.4 billion. Hence, there is a reasonable likelihood that the projection for 2030 could be 500 million too low or too high (ranging from 7.8 billion to 8.8 billion) and a small likelihood that the projection error could be more than one billion too high or too low.
In the simulations of the impacts of different population growth outcomes, we use the median forecasts as our baseline scenario of population size and take the error bands (the measures of uncertainty of the forecasts) as our ‘policy shocks’ to analyse the economic implications of higher or lower population growth than forecast by the UN.
As fertility and death rates continue to decline, a key consequence will be the aging of populations—a process already well under way in the mature industrial economies. Although most global population projections have over-estimated the fertility rate, they nevertheless predict significant aging in many regions in the world during this century. As well as affecting many other aspects of an economy, population aging also affects food demand—both directly and indirectly.
The most rapidly aging economy is that of Japan—a country that is very important for Australian commodity exports. Japan’s total fertility rate has been below 2.1 since the early 1970s and was estimated at 1.3 in 2000. In 2002, 24 per cent of Japan’s population was 60 years and older and, according to the UN baseline, by 2050 about 42 per cent of its population will be in this age group.
Japan’s population is projected to begin to decline in 2007, but the decline may well begin before then.
Total fertility rates have also reached very low levels in Western Europe, especially in the southern European countries.
This aging pattern is not restricted to the mature industrial economies. For example, China’s population is also aging very quickly. In 2002, people aged 60 years and older in China accounted for 10 per cent of the total population. The UN base line has this proportion tripling to 30 per cent by 2050. Given the UN’s conservative assumption about fertility decline in China, this projection of the aging of the population may well be an under-estimate.
What impact will this population aging have on food consumption patterns in developing countries, alongside the income growth-related changes in diets that are still under way? What will be the impact of aging on food consumption in the high-income countries? Good answers to these questions are not possible at present because the detailed information that is necessary for analysis of the impact of population aging on consumption patterns is generally not available. So we restricted the analysis to an examination of the impact of changes in population growth projections and changes in incomes on food consumption and exports.
Answering these questions requires a multi-sector, general equilibrium model of the world economy.
The model chosen for the analysis allows for international capital mobility, capital accumulation, and investment. These features allow analysis of the long-run implications of changes in populations, labour forces and incomes for the composition of global food demand. Moreover, the model allows a high level of country and commodity disaggregation—up to 65 countries and regions and 54 commodity and product groups. This commodity disaggregation also allows the assessment of changes in the international distribution of populations and incomes on the demand for products over the full range from inferior through to income-elastic commodities.
The construction of the baseline scenario is a substantial task. Not only does it require assumptions about the exogenous growth rates of primary factor supplies like labour and skills, it also rests on assumptions about the pace of technical change and the extent of international capital mobility. The latter point is of particular importance because investment and capital accumulation drive the dynamics of this model. In the simulations we chose a level of international capital mobility that is at the high end of those available, reflecting the increasing tendency for savings in some regions to finance investment in others.
Eight population growth scenarios that deviate from the median of population forecasts—the baseline scenario—are simulated for comparison. They are divided into four groups: (1) high and low population growth in the mature industrial (mainly OECD) countries, excluding Australia; (2) high and low population growth in the developing countries; (3) high and low population growth in both the mature industrial (mainly OECD) and the developing countries, again excluding Australia—the ‘world’ case; and (4) high and low population growth in Australia. These scenarios embody population projections that are in the range 2.5-97.5 per cent of the distribution of the forecast median.
In other words, each group represents a 95 per cent confidence interval of the population forecast in the corresponding regions. In these policy scenarios, population growth rates in other regions remain the same as in the baseline scenario (the median of the forecasts).
Generally, the differences in population forecasts tend only to have significant effects beyond 2010.
This is because the forecast error bands are small in the early years of the projections and because the growth process is driven by capital accumulation—the effects of which appear with a lag. All the simulations show a positive relationship between regional GDP and regional population growth. More population means more labour to combine with the existing capital stock and hence more output overall. However, there is a negative relationship between population growth and per capita income growth.
The multi-region structure of the model also captures interactions between population growth in one region and income growth in others. Because saving rates and labour force participation rates are held constant in the simulations, these interactions are of two types: (1) slower population growth in the rest of the world leads to higher per capita incomes there and hence a shift in consumption toward the more income-elastic products; and (2) slower population growth in the rest of the world reduces labour supply growth in the rest of the world relative to Australia, raising the return to investment in Australia and accelerating Australia’s capital accumulation, especially in manufacturing, which is comparatively labour-intensive. This shifts the international terms of trade in favour of manufactures.
From the simulations incorporating slower population growth in countries other than Australia, all regions, including Australia, have higher GDP per capita; however, only Australia’s total GDP rises while it falls in all other regions. Factors that contribute to higher GDP per capita in the rest of the world are different from those that contribute to a similar outcome in Australia. For the former, it is the combination of slower population growth and the Solow-Swan structure of the model; for the latter, it is the expansion of output with the same size of labour force. Lower total GDP in the rest of the world is the result of slower population growth, which reduces labour supply and therefore output.
Higher total GDP in Australia is due to more capital inflows from overseas.
In this kind of scenario, Australia is the only region that has higher investment. This outcome can be explained by the changes of factor ratios in Australia versus the changes in the rest of the world.
Slower population growth in the rest of the world increases capital per worker relative to that in Australia. As a result, investment is drawn to Australia, where the rate of return on capital is now relatively high as its capital/worker ratio is relatively low. Due to the increase in the capital/labour ratio in the rest of the world, the rate of return on capital falls world wide, which brings down total world investment (-8.7 per cent by 2030). The decline of investment in the rest of the world is far more than the increase of investment in Australia.
These effects on supply and demand in Australia and overseas ultimately change the terms of trade and export demand from Australia. When other regions in the world grow more slowly as a result of lower labour supply, Australia’s economy benefits from capital inflows from overseas and its output increases. Higher domestic production requires more imports by Australia, while demand from overseas for Australian exports falls. As a result, Australia’s terms of trade deteriorate. Slower global population growth affects Australia’s exports adversely via the relative decline in demand abroad for its energy-minerals exports and for its less income elastic commodities.
The impacts of the different population scenarios on Australian food production and exports are quite significant, especially when we consider the total ranges generated by the 95 per cent confidence intervals of the population forecast. For example, faster/slower world population growth increases/decreases Australian exports of vegetable and fruits by about 50 percentage points, or about per cent deviation from the baseline scenario.
Finally, does it matter where the population forecast errors occur? Interestingly, the impact on individual commodity sectors in Australia can be quite different, depending on the region of the world which the faster/slower population growth occurs. For example, the volume of paddy rice exports from Australia falls by over 10 per cent (from the baseline scenario) when the population growth lower end of the error band in the OECD, while it increases only slightly (by 0.3 per cent from baseline scenario) when the population growth rate is at the lower end of the error band in the developing countries. Among agricultural products, the worst hit export sectors when the population growth rate in developing countries is lower than the median are forestry, vegetables, fruit and nuts, fishing, and bovine cattle, sheep, goats and horses. The worst hit export sectors when the population growth rate in OECD countries is lower than the median include paddy rice, oil seeds, and forestry.
|
