|
|
|
|
Arguments - Prediction Error |
| Diet |
While the general trends of the change in diet seem rather predictable in
China, the specific diet pattern is not. Moreover, it is the specific diet pattern that
can greatly affect overall grain demand. In the case of China, diet may be more important
for the overall food demand than population growth. |
| Crucial
Issues |
| Can we
predict diet change? Let us focus on two issues: |
 |
The background: Why is
it difficult to predict diet change? |
|
What methods are
available to predit diet change? |
 |
|
|
| Discussion |
| Background: Why is it difficult to predict diet change? |
|
Predicting
diet change in China does not seem too difficult, because the range of possible change is
limited and most of the change has probably already occurred. For any further change,
there are only two likely possibilities. If the country follows its current path of
economic reform and development, its diet will most likely become similar to those of
other developed Asian countries. It will become more diverse, it will probably include
more animal products (milk, fat), and it will have a higher percentage of foods with
high sugar content, such as sweet beverages, ice cream, and cakes. People will also drink
more alcohol. However, if economic crises strike or if a return to command-and-control
agriculture should occur, China would probably change back to the staple diet of rice,
wheat, and tubers that was typical in the 1950s and 1960s. While these two general
scenarios seem obvious, many uncertainties exist in the details. Will Chinese consumers in
the future prefer pork or poultry, freshwater fish or sea fish? How much will consumption
of milk products increase? How much will consumption of alcoholic beverages and stimulants
increase?
In our in-depth analysis, we have tried to model China's future food demand as it depends
on diet change and other factors. The bottom line of this analysis is that diet matters.
In fact, in the case of China diet change could have a bigger impact on overall food
demand than population growth (for details, see the in-depth analysis). The analysis shows
that overall food demand could vary considerably depending on the specific diet
pattern (see Table 1). |
Table 1 |
| What methods are available to predict diet change |
| The
simplest method for predicting change in diet is to extrapolate current trends.
Unfortunately, this method produces unrealistic results if applied to more than a few
years. For instance, if we extrapolate China’s steep per capita increase in meat
consumption for the next 30 years, we get a completely meat-based diet by the end of the
projection period. Obviously, diets can change only in a relatively narrow range: it is
unlikely that China’s population would eat, on average, more than 60 or 70 kg of meat
per person per year. |
|
| Alternatively,
we can use the current food consumption patterns of other, already developed Asian
countries as a target pattern for China. One can imagine that China’s average diet
will approach the current pattern of Hong Kong, Singapore, or South Korea. However, there
are deep ethnic and cultural differences between countries that must also be taken into
account. |
| Finally,
we can develop a model of change in diet that predicts food consumption patterns from
related social and economic factors. For instance, we could calculate the elasticity of
meat consumption by income or level of urbanization. With growing income or urbanization,
consumption of meat would increase. Unfortunately, the income elasticity for most food
items is already low in China. For instance, an income increase would not greatly
affect average meat consumption in China, because it is already very high and the
small marginal increase from income growth would not make a big difference. Income growth
was certainly important for China's change in diet over the past few years, but in the
future, food preferences will depend more on changes in lifestyle (which are hard to
predict) than on income. |
| Related Arguments |
|
Diet Change: Trends
Impact Data Quality Prediction Error Intervention Possibilities Intervention Costs
|
|
|
Revision 2.0 (First revision published in 1999)
- Copyright © 2011 by Gerhard K. Heilig. All rights reserved. (First revision: Copyright © 1999 by IIASA.) |
|
|