Brian Vuong wrote:
>
> Hello everyone,
>
> I would like to build a regression model that explains aggregate savings
> behavior as a function of the level of interest rates. Should I use the
> sample during a period of fuluctuating interest rates or a period in which
> interest rates are constant? Any help is greatly appreciated.
You had better - if this project is advisable at all, which I rather
doubt- use a _long_ period, during which interest rates fluctuate
through the same values several times with other variables varying in
the meantime.
A sample with interest rates constant will give you _no_ information
about what happens when they change. A short sample during a period of
increase or decrease will not permit you to distinguish between the
effects of that change and the effects of other changes occurring at the
same time.
On top of that there will be effects due to rate of change of interest
rates (8% after a period of tight money may look low and encourage
spending, the same rate after a period of very low rates might look high
and encourage saving), alternative investments (8% during a bear market
looks much better than 8% in a bull market), inflation (5% interest with
5% inflation is neither the same as 5% interest with 0% inflation nor
10% interst with 5% inflation), and so on.
My own feeling is that by the time you have enough data for a model,
something will change to upset the model that wasn't available during
your data-gathering period. But good luck anyway...
-Robert Dawson
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
