|
Response to a Reader:
Jim Campbell, in a response to the Washington Post Blog, Stumped, which cited Part 2 of this commentary, made the observation that "Every president inherits an economy from his predecessor. If you consider that it takes some time for a president to formulate a policy, for Congress to process the policy, and then additional time to implement it and for it to be felt in the economy and measured, it seems reasonable to expect a lag of about 18 months from inauguration until it is that president's economy."
This is a reasonable observation that in the interest of fairness and accuracy should
be checked out. Fortunately, there is a technique called lagging, which is helpful
in picking up effects that show up after a specified time. The classic example is
trying to predict how many mufflers will be need to be replaced on a given model
of car. If you look at the car repair data, very few mufflers will be replaced the
year the vehicles were sold. In fact not until the third year will a large number of
them will be needed. This makes sense because it takes about three years to wear
out a muffler. By the fourth year most of the mufflers have been replaced, leaving
a few random sales thereafter. Thus, using car sales today you can get a
reasonably accurate estimate of how many mufflers will be sold in three years, but
a much weaker prediction of how many will be needed in other years.
It could be argued that a similar situation occurs when we change presidents.
However, unlike Mr. Campbell, I will argue that presidential policy changes are a
process and not an event. This process begins when financial markets react as
soon as it becomes apparent that one candidate will win and ends sometime
between the first and second year as legislation is passed and the law put into
effect. Thus, it is important to make more than one measurement to see if a party
effect is truly there.
In the case of my data set, the unit of analysis is the year, thus, it can only be
used in increments of one year. This means that I can't test his hypothesis
precisely at 18 months but I am able to do so at one and two years.
The results were facilitating. At a one year lag, the results were almost exactly
the same, literally within rounding error at one significant figure. The Republican
averages were 2.832 percent of GDP for the current year and 2.780 for a one year
lag. The Democratic average was 4.434 for the current year and 4.370 for a one
Page 1 of 2 Next Page 
|
| |
| |
