How many people can the world take? Are developing countries heading for a population disaster? Will falling fertility play havoc with pensions? Population forecasting pits environmentalist against economist and analyst against analyst, as Warren Sanderson of the International Institute for Applied Systems Analysis found out. He鈥檚 part of a team attacked recently by The Wall Street Journal for introducing as 鈥済ood news鈥� their prediction that global population growth would stop by the end of 2001. Sanderson explains to Liz Else why the troubled discipline needs a radical rethink.
What is your latest forecast for world population?
We鈥檙e saying there is an 85 per cent chance that the world population will reach 9 billion before 2100 and then fall.
Advertisement
And what is the UN鈥檚 forecast?
The UN predicts a peak of 9.3 billion in 2050-that is 300 million more than our forecast.
Why such a difference?
They鈥檙e higher because of a problem with their basic assumptions. The UN assumes that in the long run every country will have the same fertility rate of 2.1, which is the rate required to replace the population. For example, in Austria, the average number of births per woman is 1.3, so the UN is expecting an enormous increase in Austrian fertility, back up to 2.1. The other side of the coin is that many developing countries have fertility rates that are high but falling. In South Korea, it鈥檚 1.6. In Thailand, about 1.8. In fact, we don鈥檛 have any examples of countries coming down from high rates, hitting the 2.1 barrier and stopping there. A prestigious panel set up by the US National Research Council that looked at state-of-the-art population forecasting said they expected fertility in today鈥檚 developing countries would eventually come into the range of today鈥檚 developed countries.
Does the UN have an agenda?
I鈥檓 not sure. We do say that the UN has to give up its assumption about fertility which is so patently wrong. But there is an implicit political element. If you naively take a fertility rate above 2.1 and project it into the future forever, population will explode and go off to infinity. If you take a rate below 2.1, in the long run population will go to zero. The prospect of populations going to infinity or to zero scares politicians. In the real world populations of countries don鈥檛 go to infinity or zero because they adjust to their circumstances. Nevertheless, rather than being confronted with figures showing populations that grow or shrink forever, the UN finds it easier to assume that populations in all countries will eventually stabilise.
But surely a decline in the world鈥檚 population is good news?
Most of the population growth is going to be in the world鈥檚 poorest countries, in south Asia and sub-Saharan Africa. But even their population growth is slowing down and may even come to an end by 2001. I think that鈥檚 good news for those places. You have a combination of poor government and deteriorating resources and famines and wars. In this context any little thing that can help relieve some of the pressure on resources has to be a good thing. But we were criticised in a Wall Street Journal editorial for saying that. The paper invoked the horrors of population control and of some countries running out of people.
Why do you think the paper got upset?
There鈥檚 been a long and unfortunate debate on population between environmentalists on the one side and some economists and some people in the business community on the other. Paul Erhlich wrote The Population Bomb in the 1960s, which talks about the runaway growth and says things like there鈥檚 no hope for India. Well, India has done fairly well, and although there鈥檚 a lot more people since then, there鈥檚 also a lot more economic growth. So there was a backlash in the US against the overly pessimistic forecasts of some environmentalists. Some economists and people in the business community now hold that you shouldn鈥檛 worry about numbers when you think about development. Neither of these extreme positions provides a realistic view of the role of population.
Could you say any more about the UN鈥檚 forecasting techniques?
The UN doesn鈥檛 use probabilistic population forecasting, which is basically the current state of the art in population forecasting. At IIASA, we鈥檝e been using and improving our own version since 1996. The UN relies on what it calls its medium variant, high variant and low variant. These are supposed to give you some idea of uncertainty. The problem is that these are probabilistically inconsistent.
What do you mean?
Its medium projection is what it thinks is the most likely outcome. To get its high variant, it increases the rate of fertility a bit. To get its low variant, it decreases fertility a bit. There are two problems with that. One is that they don鈥檛 do anything with mortality. That matters, because only part of the uncertainty about the world鈥檚 population has to do with uncertainty about fertility-and there鈥檚 also uncertainty about migration. The second problem is that their forecasts take uncertainty into account, but in a very artificial way because there are no probabilities associated with those high and low variants.
And this makes those spans a bit crude?
It鈥檚 worse than that. In the probabilistic sense they鈥檙e meaningless because they don鈥檛 really provide information about the uncertainty of forecasts. There鈥檚 also a problem when they go from countries to regions and then from regions to the whole world. To get the high variant for the world, they add up all the high variants for each country. Now, there鈥檚 no particular reason why a high variant for the world should be high for Brazil and also high for China. When Brazil is high, China might be low. None of this is a representation of what the uncertainty is.
So why does the UN do it this way?
The UN is charged with making forecasts on a regular basis for probably around 200 countries and territories. That鈥檚 a very difficult task, and their method contains a simplification. It turns out that it鈥檚 not a simplification of a very good sort, and we believe that there are other better ones that they could make. The bottom line is that they should move to a probabilistic model, then the high and low figures would have some meaning.
So which is worse, growth or decline?
We鈥檙e working on a new way of thinking at IIASA: population balance. It takes into account the environment, human creativity-in terms of inventing new technology-and the human life cycle, which is very important. The population balance approach emphasises that while rapid population growth can be a problem in some places, rapid population shrinkage and ageing could be a problem in other places. So population balance is a way of thinking about what the population should be or what would be best for us.
Our population forecasts are based on real data, while our balance model is just a tool for understanding and discussion that combines the best parts of the arguments of the environmentalists and the economists. We鈥檙e going to put in some real numbers in about a year, and we鈥檝e a book planned.
What do other forecasters think?
Everybody seems interested. We want to make sure that the concerns about population are heard in a scientific way at the Johannesburg Earth Summit next year. This is a follow-up to the Rio Summit. My feeling about the Rio Summit was that because of the unproductive debate about popualtion growth, people ended up saying: 鈥淵ou guys don鈥檛 know what you鈥檙e saying, you tell us either don鈥檛 worry about human numbers when you think about development, or the exact opposite.鈥� What we should tell people is: 鈥淒on鈥檛 worry about the things you don鈥檛 have to worry about, but do worry about the things that you do have to worry about.鈥�