[ show plain text ]
I would still argue that it is a very strong statement to say "any
empirical calculations are more of the "back of the envelope" variety." The
quality of empirical work depends on:
1. the quality of theory underlying the empirical work,
2. the availability and quality of the data,
3. the availability and quality of statistical procedures, software, and
4. the imagination and ability of the analyst.
When all four of these factors are solid, the empirical work is solid, that
is to say, it is something more than a "back-of-the-envelop" calculation.
Consider Thomas Sowell's work on race and labor market outcomes. Sowell is
a strictly "back-of-the-envelop" economists. None of his work ever uses
anything more sophisticated than descriptive statistics. He argues that
since African American immigrants from the Caribbean earn at or near the
national average and that since Asians earn at or near the national
average, there is no racial discrimination in the labor market. Of course,
he believes this regardless of the data. But, that's beside the point. His
back-of-the-envelope analysis is also deeply flawed. Using the same
descriptive statistics, a better theorists would simply say that the
distribution of income is "mediated" (to use Alfredo's language) by
geographical location. Those who live on the coasts of the US face greater
price levels than those who do not. Hence, the average level of income for
residents of New York, Boston, LA, Hawaii, etc are higher regardless of
race. Caribbean and Asian immigrants are bi-coastal residents in the US, or
they tend to live in cities such as Chicago (which also has a very high
cost of living). Hence, the correct comparison is between the incomes of
Caribbean and Asian immigrants and the average income of the populations in
the cities where they reside, rather than for the average for all of the US.
The point here is that measurement matters. Some measurement is better than
others. All measurement has some flaws. Indeed, all theory has flaws. So,
if we say that all empirical work is strictly "back-of-the-envelop" because
it has flaws then we must make the same statement about high theory.
Economic measurement is linked to economic theory.
At 08:36 AM 5/25/00 -0700, you wrote:
>I did not want to give the impression that all empirical research is a
>time. I only intended to say that the pretension of precision is foolish. I
>tried to explain -- perhaps not very well -- that we do better to realize
>any empirical calculations are more of the "back of the envelope" variety.
>"Patrick L. Mason" wrote:
>> But, your point is fundamentally different from Mike and others. They are
>> arguing that any type of empirical analysis is a waste of time. You are
>> simply saying that there are some theoretical problems with the
>> Neo-Ricardian approach to values and prices.
>I was also making the anti-neo-Ricardian poing.
>California State University
>Chico, CA 95929
This archive was generated by hypermail 2b29 : Wed May 31 2000 - 00:00:12 EDT