Wednesday, March 5, 2008

More on dollar index; Why YOY?

In my post below on Kudlow's latest column, I didn't provide a justification for illustrating the dollar index since 1974 in year-over-year (YOY) terms. It isn't necessarily intuitive, nor is necessarily the "correct" method, but I have some reasons.

The historical dollar index data is given per month since 1973. I could look at the percent change month-to-month (MOM), which would provide an instantaneous rate of change. I decided to use YOY instead of MOM for a couple of reasons.

YOY is often used to account for seasonal fluctuations in data. A classic example is number of houses for sale in a given market. Real estate is a seasonal industry; business usually is twice as brisk in the summer than it is in the winter. Every spring the number of houses for sale increases, and every fall that number decreases. If you want to assess the health of a market like that, the best comparison is to compare houses for sale this year vs the same month last year. This removes any seasonal bias.

The problem is that it doesn't account for any changes that occur since 12 months ago. I wrote a post about this last year about median price for SFH in King Co, WA.

OK, back to the dollar index. There isn't a seasonal pattern in changes within the dollar index. So why use YOY? I felt that I wanted to assess the effect that a presidential policy may have on the index. The force of a policy is somewhat limited, given all of the factors that influence currency rates. Thus, if an administration favors a particular "strong" dollar policy, you would expect to see pattern of positive YOY % changes over the long term.

My point was that during virtually Reagan's entire second term YOY % change in the dollar index was negative. Kudlow claimed Reagan favored a "strong" dollar policy. But the data suggest that either Reagan was impotent to implement a strong dollar policy, or he didn't favor such a policy. Given the Plaza Accords, the evidence suggests the latter.

No comments: