[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Simulations



At 06:28 AM 8/29/96 -0400, Duncan Frissell wrote:
>At 10:05 PM 8/28/96 -0800, jim bell wrote:
>
>>To me, the most obvious one is GIGO:  Simulations, especially 
>>political/social ones, might depend heavily on assumptions that are 
>>programmed into them.   A trivial, yet interesting example is the computer 
>>game "Sim City" which allowed you to adjust the "tax rate" but problems 
>>always cropped up the further away you were from 7%.   The libertarians were 
>>frustrated that we were unable to drop the tax rate and still get a 
>>well-functioning, happy society.  
>
>I was taking economics back in the Armonk Iron days and we played around with
>an economic simulation program written in Fortran.  One was supposed to
>adjust government spending and taxes to find an optimum level.  I set both
>taxes and spending to zero.  We got a lot of economic growth and a lot of
>inflation (this was not a monetarist simulation).

Yes, it does sound a bit inaccurate.  Let's see, while I only took one 
economics course in college (macroeconomics) as I recall inflation is 
generally the product of either the monetarization of the deficit or an 
increase in wages unmatched by productivity increases.  The latter is 
unlikely to be caused extensively by government policy or spending 
(Davis-Bacon is an obvious exception to this in the private sector), and if 
the government spending and taxes are zero then obviously the former can't 
be the cause.  


> But we were happier.

That reminds me of a related flaw in Sim City.  When the program is queried as to
what the public's main complaints are, the information is presented by the program as 
the percentage of the population considering various problems to be the 
"most important."  However, it did not explicitly rank the SERIOUSNESS of 
these problems in absolute terms.  Since it is practically an axiom of human 
nature that we'll FIND something to complain about if nothing comes 
immediately to mind, the danger is that you'll over-react to a non-problem 
with a "solution" (often involving more government spending) that is even 
worse.

Another bias was the fact that the program seemed to lump all 
infrastructure-type spending into government, as opposed to a privatized 
solution.   In reality, if you drop government spending very low problems 
will begin to be solved by private means.  In Sim City, however, the system 
did not compensate in this way,  leading to low or even negative growth, and 
extreme dissatisfaction, etc.

Jim Bell
[email protected]