A CSIRO team has spent ten years, it seems, building a model of the
"physical economy" to consider the environmental impact of
alternative population scenarios. In particular it considers the impact on
the availability of natural resources like water, fossil fuels, forests
and fish stocks, on biodiversity, on demand for food and housing, and on a
host of other "physical" inputs in economic activity. It does
this by trying to "keep track of all physically significant
stocks" like "people, livestock, trees, buildings, vehicles,
capital machinery, infrastructure, land, air, water, energy and mineral
resources". The main constraint in solving the model is that the
economy must "obey the known physical laws". The authors
conclude, for example, that the Australian economy could sustain a
population of 50 million by 2100, although it doesn’t nominate an
optimal population target.
Is this a useful exercise? Consider this analogy. The CSIRO model
treats the economy like a pinball machine. The only role for human
decision-making is to put the ball in play. Once in play the ball bounces
around according to automatic, preset mechanisms. If the ball threatens to
disappear down the centre chute – in the language of the CSIRO model,
the economy suffers "tensions" - the controller (policy-maker)
can flip it back into play on a different trajectory and watch it bounce
Economists also use simulation models to consider the impact of future
demographic scenarios. But their models are more like a soccer game. The
referee (policy-maker) puts the ball into play but, once in play, it doesn’t
bounce around according to the predetermined forces of the physical world.
Rather, players (consumers and firms) interact, make plans according to
their goals and targets, and adjust their plans in response to the actions
of the other players. The path of the ball is not preset – it is
continually altered by purposeful human interaction.
Regrettably, the CSIRO model completely misses this fundamental idea.
Consider some examples of how their model can get the effects of
demographic change seriously wrong.
- We know that as the population gets older and the workforce shrinks
as a share of the population, the ratio of physical capital goods to
labour will rise and hence the productivity of additional capital
goods will fall. This will make additional investment less profitable
and so firms will invest less. As a result the call on the nation’s
resources will be commensurately lower. The CSIRO model cannot capture
this because it ignores the basic idea that firms respond to changes
in profitability by making adjustments to their inputs and outputs.
- The above is an example of a more general problem in that the model
assumes no changes in relative prices as a result of demographic
change. The failure to allow relative prices to change means that the
model ignores the fundamental principle that people respond to
incentives to look for cheaper ways of doing things when prices go up.
We know that when oil prices went up in the 1970s people found that
natural gas was a substitute in some uses, and they started to drive
smaller cars. We know that in places where land prices are high,
people economise by living in smaller houses. Similarly, if in the
future the price of water were to rise, people would economise on
water use. If medical procedures become more expensive, people will
tend to look for preventative treatments and perhaps alter their
lifestyle. And so on. None of this is acknowledged in the CSIRO model.
- The CSIRO’s model of international trade assumes that exports are
the "quid pro quo of imports". The documentation is a bit
vague about what this means, but the suggestion is that somehow
exports of physical goods must be found to more or less match imports
of physical goods. Assumptions are then made about the value of
exports and imports of services and the interest rate in order to
calculate a current account balance (CAB). Again the documentation is
short on detail here. But on the face of it, their method of
determining the CAB appears to bear no relation to the way the CAB is
actually determined. For a start, exports are not the "quid pro
quo of imports". That is simply wrong. We do not need exports to
match our imports. Imports can be greater than exports if foreigners
are willing to make a net investment in Australia. The CAB is in fact
equal to the difference between domestic saving and domestic
investment, which are in turn determined by forward-looking behaviour
of households and firms. These decisions are in turn based on the
weighing up of costs and future benefits. The consequences of ignoring
all this is that the model could get the future size of the CAB
seriously wrong which in turn means that Australia’s international
debt will be wrong. That can be important because there are
constraints on the feasible value of Australia’s future
international debt. In a properly specified economic model these
constraints cannot be violated.
Essentially, in trying to keep track of all physical stocks the model
ignores the basic truth that human decisions move these stocks around
according to economic forces. It is not good enough to treat the
"controller" as a pinball player who puts the ball into play and
watches it bounce around according to physical laws. It must bounce around
according to economic laws that are based on what we know about human
behaviour. Economic models try to do this.
Notwithstanding all of the above, there is value in the CSIRO’s
exercise. Economic models could certainly be improved by accounting for
the demands on the physical environment and ensuring that the allocation
of resources does not violate known physical laws. A weakness of existing
economic models is that they tend to ignore these physical constraints.
Most economists would probably have no problem with the CSIRO’s
conclusion that Australia can support a population of 50 million by 2100
– but the economy is likely to look a lot different to that projected by
the CSIRO model. Some other conclusions are very difficult to sustain
given the deficiencies in the modelling approach. For example, the
conclusion that the low-population scenario will result in a higher
physical trade balance cannot be relied upon given the flawed model of the
international trade. The conclusion that without substantial structural
change it could be hard to maintain economic growth also suffers from the
model’s weak economic foundations. Economic growth is driven primarily
by productivity growth. The effect of population on productivity growth is
in fact an issue over which there is considerable economic debate. Suffice
to say that there are plausible arguments that the effect could go either
way. That is, a lower population could either boost or lower productivity
growth. We don’t know. A reasonable position to take would be to assume
a zero net effect, in which case economic growth would be maintained just
as well under a low population as under a high one.
It is a pity that the CSIRO team did not seek input from economists
earlier on in this project. Apparently there were some economists on the
"reference group", and they were not impressed with the
violation of economic laws described above. It would have been better if
they were brought into the team at the model design stage. Perhaps a
positive to come out of this exercise will be a much closer collaboration
between scientists and economists who are both interested in modelling the
future impact of demographic change on people’s lives.
The report can be downloaded here
Discuss in our Forums
See what other readers are saying about this article!
Click here to read & post comments.