From: hal@finney.org
Date: Wed Nov 01 2000 - 00:31:49 MST
Nick writes:
> A few possibly clarifying points:
> 1. As Robin points out above, we factor out uncertainty, growth etc. and
> focus on the remaining tendency that people have to discount the future.
I don't think you can separate out discounting so easily from these
other effects. An analogy I might suggest is our taste for foods.
In some sense this is ideosyncratic and comes from within. However it is
determined in large part by our evolutionary heritage. We like foods that
are sweet and fatty because they were good energy sources in an era when
starvation was a threat. We don't sit down and analyze a food rationally
to determine whether it tastes good, so it may appear that whether a
food tastes good or not has nothing to do with how healthy it is.
In the same way, discounting may still appear even in experiments where
growth and risk of loss have been eliminated. But that doesn't mean that
it is independent of those effects, any more than the fact that unhealthy
foods taste good means that tastes have nothing to do with the health
value of foods. Rather, in each case I would suggest that evolution built
into us an instinct which serves us well in the majority of situations we
encountered at the time. We like sugar and we discount the future for the
same reason, because it was a good strategy for our ancestors to follow.
Put us in a situation where sugar is unhealthy and we'll still like it.
Put us in a situation where there is no risk in postponing consumption,
and we'll still want to consume the majority today. It's not because
discounting has nothing to do with risk, it's that discounting is an
instinctive behavior that works well in a risky world.
It also appears to me that the authors have not tried to separate
discounting from issues of uncertainty about the future, as when they
write,
: Beyond revealed preference, the strongest theoretical argument in
: favor of discounting, involves what Bohm-Bawerk referred to as the
: "brevity and uncertainty of human life." According to this view,
: people discount future felicity because they may not be around to
: enjoy it. An individual weighs pleasures at date t by the conditional
: probability of living until date t.
Rather than factoring out uncertainty, they propose uncertainty as a
theoretical justification for discounting. I suggested earlier that
in addition to the possibility of death, there are other forms of
"uncertainty of human life". Given that life is uncertain, it seems
very plausible to me that evolution would program discounting into
living creatures.
Put another way, I suggest that if we could somehow imagine creatures
evolving in a situation where there was no uncertainty, and goods saved
would be guaranteed to be available in the future, those creatures would
not discount the future. They would have no incentive to consume more in
the present. Rather, they would take a long term perspective and allocate
their goods equally across each period of their life. Of course this
is purely hypothetical but it will hopefully make my position more plain.
We see discounting today as a psychological effect, but the fact remains
that it must be there for a reason. The question is, then, is discounting
rational, and are discount rates optimal under at least some conditions.
In other words, is discounting purely endogenous as the authors seem
to assume, a matter of taste only, or is it exogenous, determined by
outside conditions and external constraints?
If the latter, proposals to artificially change discount rates must
cause inefficiencies and impose costs. I don't see that the paper's
analysis considers this possibility. Social planners do not have the
freedom to shift discount rates around at will in order to benefit one
time frame over another, any more than they can shift prices at will in
order to benefit one group or another. Making these changes will have
consequences for the welfare and efficiency of the society as a whole.
Any proposal to take such actions must look at the larger picture.
Hal
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:31:48 MST