[Note for bibliographic reference: Melberg, Hans O. (1997), Why culture should not
be treated as a residual, www.oocities.org/hmelberg/papers/971025.htm]
Why culture should not be treated as a residual
Positive proofs vs. Proofs by default
by Hans O. Melberg
Introduction
In his book How Russia Became a Market Economy Anders Aaslund writes:
"Many Russian peculiarities have been inferred as objections to reform. One
objection is the backwardness of Russian political, economic, or legal culture. With my
strong belief in modern institutional economics, I treat culture as a residuum that can
explain what cannot be explained with other means" (p. 10).
I interpret this to imply that explanations based on rational self-interest are always
better than explanations based on culture. Hence, if he had to choose between two
explanations of the same phenomena, Aaaslund would always choose the "rational"
explanation, not the "cultural explanation" (a lexicographic preference). Only
if he cannot find a rational explanation is he willing to consider a cultural explanation.
The quotation from Aaslund is not isolated. In fact, most economists and many political
scientists argue that culture should be treated as a residual (see for instance Elkins and
Simeon's article. Culture as a residual cited in the bibliography of my "The
Cultural Approach to Russian Politics" at www.oocities.org/hmelberg/papers/papers.htm).
I disagree with this, and in this short essay I want to explain why.
The arguments
I have three arguments. First, rational choice does not win by default: I want concrete
evidence of the mechanisms involved, the linkages and the assumptions in the rational
explanation. This can be used to judge the reliability of the rational explanation, for
instance by examining the empirical plausibility of the assumptions and the empirical
strength of the links. Second, culture does not win by default i.e. the fact that you
cannot find a rational explanation does not imply that a cultural explanation is true.
Once again you need positive evidence; a demonstration of the mechanisms involved. And,
once again the plausibility of this explanation can be judged from empirical examination
of these links. Third, and finally, it is perfectly possible that the probability of a
rational explanation being true is less than the probability of a cultural explanation
being true. Hence, when faced with two phenomena it is wrong to always choose the rational
explanation no matter how implausible and outlandish it is. In short, sometimes a cultural
explanation, or some other explanation, is more plausible and should be given preference.
Before I go on to illustrate these three arguments (and I mean illustrate, all my major
points have already been made, the rest is only examples!), I want to make one
qualification: I am not against explanations based on rationality. People usually try to
do what they think is best for themselves (rational selfishness). However, my point is
that I want positive evidence for this in every specific case (to avoid the fallacy of
composition), and this evidence takes the form of tracing and empirically testing each
step in the chain, each mechanism in the explanation. This testing gives rise to a
probability of an explanation being true, and this probability determines which
explanations is the best. One cannot, and should not, give lexicographic preference to the
rational explanation.
With this qualification in mind, I shall give some examples to explain what I mean,
starting with the third claim that even if we can find a rational choice explanation it is
not always superior to alternative explanations.
Are explanations based on rational choice always superior to other explanations?
Consider the claim that religion can be explained by parents wanting to reduce the rate of
discounting in their children (in other words, make them more patient). One
"rational" strategy for doing so, might be to make your children religious
(suggested by Gary Becker). The continued existence or religion, the continued upbringing
of children by parents in a religious way, is explained by a rational choice: The parents
want their children to be better off, to be better off they have to be patient, being
religious makes them more patient - so parents give their children a religion. And, this
is only one way of explaining religion rationally. One might also argue, as Becker does
(according to Jon Elster) that the upper-classes use religion to pacify the lower classes
- to protect themselves from revolutions. In short, religion can be explained by rational
choice.
My immediate feeling is well captured by Jon Elster who writes that:
I find this analysis intrinsically implausible and devoid of empirical support. Had
Becker looked at the historical literature, such as the chapter on "The Transforming
Power of the Cross "in E.P. Thompson's The Making of the English Working Class,6
he would have found that for religious indoctrination to work the indoctrinators have to
believe in the religion themselves.7 Self-consciously manipulative attempts to inculcate
religious and political ideologies invariably fail.
(from Elster's review - "More than Enough" - of Becker's Accounting for
tastes, available at the Jon Elster Page: www.oocities.org/hmelberg/elster.htm)
To claim that an explanation based on rational choice is always superior to alternative
explanations, is to ignore that the plausibility or an argument depends on many factors.
Of course, the theoretical possibility that religion is simply a tool of the oppressor to
keep the oppressed down, or a tool used by parent to induce more patience, does exist.
However, the mere possibility that religion can be explained in this way, does not make me
believe it: I want positive evidence that this is actually the way religion is used.
When faced with two explanations for religion, the correct approach is then not to ask
"which is based on rational choice" and "which is based on a non-rational
conception of man" (for example, a psychological explanation based on the need for
meaning). Nor is it correct to always prefer the explanation based on rational choice. The
correct procedure is to test various explanations, to examine the links empirically (and
logically) and decide which explanation is most plausible.
Rational choice does not win by default
In his book Rational Expectations Steven Sheffrin repeatedly reports tests which
could not reject the theory of rational expectations (which is the null hypothesis in this
case), and this is interpreted to support the theory of rational expectations (see p.17,
21, 52-54, 142-3, 156). In my opinion, and I think Sheffrin would agree, this is not a
very convincing or powerful argument in favour of rational expectations. By convention the
null hypothesis is very difficult to reject (at the 5% significance level), and the
failure to reject it does not constitute strong evidence in favour of it.
There are, in fact, quite a lot of micro evidence that people often do not have
rational expectations. (See, for instance, Kahneman & Tversky & Slovic: Judgement
under uncertainty , H. Simon (bounded rationality) and D. Hausmann on preference
reversal in The Inexact and separate science of economics). Now, I do not want to
turn this into an essay about the truth of rational expectations in general. My point is
only this: To believe in rational expectations we need positive evidence, not simply the
failure to establish an alternative (and equally general) expectations model. This
fallacy, I believe is exemplified by some of the articles considered in Sheffrin's book.
One might say that you cannot beat something by nothing - that you should believe in
rational expectations as long as you do not have an alternative. I disagree. It is
perfectly possible to be agnostic - do admit that we do not know. The choice is not
between the extremes - yes or no. rather it is a question of more and less faith, as well
as the intellectually respectable option of "uncertainty": I do not know!
Culture does not win by default
Faced with a puzzling phenomena that apparently cannot be explained by rational choice, it
is no necessarily true that a cultural explanation is most plausible. Once again, only
positive evidence can convince me of the plausibility of the cultural explanation. By way
of illustration, consider the following: Some might argue that employees are guided by
norms about fairness when they do not reduce the wage to the lowest possible given the
market: In other words, the "irrational" and high wage level is interpreted as a
failure of rationality, and as evidence in favour of a norm-based explanation.
However, the same phenomena, it turns out, can be explained by rational choice. Given
that the employees are humans with emotions - so they might work less hard if they feel
their pay is too low (and they are more likely to shirk if the potential penalty is lower
i.e. the efficiency wages argument) - it becomes rational for employers to keep a wage
above the market clearing level. The cut in wages is not worth the gain since productivity
will also fall if you cut wages.
The lesson is this: The apparent failure of simple and rational profit maximization,
does not constitute good proof in favour of a cultural explanation. It only teaches
us to look further for better explanations of whatever type that the data tell us is most
plausible - rational, cultural or some other explanation.
One should also note that there are many alternatives to a rational explanations, and
culture is only one of these alternative. We could use psychology (cognitive failures) to
explain irrationality, we could use norms to explain irrationality, we could use ideology
to explain irrationally and so on. This implies that the rejection of on, does not
constitute good evidence in favour of another approach - since we do not know which
alternative is the most likely to be true. If there had been only two mutually exclusive
alternatives the inference (rejection implies confirmation of alternative) would be
correct. For instance, if I reject the hypothesis that you are pregnant, then it is safe
to infer that you are not pregnant. However, in our case a cultural explanation is only
one of several alternative to a rational explanation.
Conclusion
The arguments made in this essay may seem obvious (and repetitive). However, the points
are continually being ignored by many social scientists, as I have tried to show with some
examples. Aaslund wants to treat culture as a residual (as does Elkins and Simeon), Becker
and many other economists (e.g. Steven Landsburg: See the first pages of his book
The Armchair Economist seem to give rational choice not only a favourable position (as
a first working assumption), but a lexicographic preference: Explanations based on
rational choice are by assumption superior to other explanation, no matter how implausible
they are empirically. In real life, imperfect calculative abilities, emotions, culture and
norms also do play a role. Even Aaslund, for instance, admits that the outcome of reforms
in Russian (under Gorbachev) was different from reforms in China because, amongst other
reasons, there were many "entrepreneurial" Chinese people - a factor which might
be viewed as cultural (p. 15). (It implies that the market economy requires more
than rational selfishness to work, it also seems to require some kind of entrepreneurial
spirit). Now, as for the degree to which people are rational, emotional, selfish, guided
by tradition and so on, I remain agnostic. All I know is that the only way to find out is
to do empirical studies, as opposed to assuming that one alternative is true or false.
[Note for bibliographic reference: Melberg, Hans O. (1997), Why culture should not be
treated as a residual, www.oocities.org/hmelberg/papers/971025.htm]