Wednesday, January 04, 2006

How to make a forecast

It's that time of year when people are cranking out their fearless forecasts for 2006. I have my own:
The peso:dollar exchange rate may be higher or lower by December '06 than it is today. The Philippines' GDP growth for '06 may be higher or lower than it was in '05.
Really, that's the only way to get it right. Say something people can really act on, and you may end up regreting it. Consider the case of Bernie Villegas:

Speaking before members of the Pampanga Chamber of Commerce and Industry, Inc. (PamCham), Villegas, Senior Vice President of the University of Asia and the Pacific (UA&P), presented the strategic plan for Central Luzon, underscoring the need for the Philippines to familiarize itself of the "engines of growth" in the global community.

In his presentation, Villegas said the peso may end up at P55.6 to a dollar by the end of 2005. At the first quarter of the year, the local currency faired at about P55 to the US dollar, which presently shows up to be the strongest currency in the world.

In 2006, he said, a dollar would cost P56.5.


Yes, the same Bernie Villegas, famous "prophet of boom", who predicted a foreign market-powered growth in the 1970s, when the Philippines could still be called an "agricultural economy". However to be a reliable soothsayer, all your misses should be tracked, as well as all your hits. And if you bought dollars in '05, betting on Bernie's forecasts for '06, you're going to be weeping all the way to the bank.

List of good forecasting practices:

1. Don't. You can do this and still look smart. (It's an art.)

2. And if you can't (as often happens), qualify your forecast so that if someone acts on it, you get to cover your ass. (It's also an art, but requires greater skill.)

3. 'Fess up in style! See Brad Setser. Lots of people appreciate how frank he's been, in contrast to some soothsayers out there, whose sole concern is to crank out another set of wrong forecasts for the next year.

4 comments:

cvj said...

i had a teacher in college once who was also a businessman, and he was really mad at Villegas advising the class not to believe anything he says. Unfortunately, never found out what that was all about.

In the december 2005 archives of www.j-bradford-delong.net, delong said this about respected economist robert shiller's 'irrational exuberance' (published 1996) "anybody who followed the portfolio strategy implicit in Shiller's analysis now painfully feels in his wallet or her purse".

Amadeo said...

Even in this age of technological wonders, the “art” of forecasting continues to be a very tender issue. In the political scene here in the US, the users of forecasts now make it a habit of “pooling together” all the forecasts about a certain topic and averaging them out with the ubiquitous phrase, “plus/minus X% of error”.

Still, forecasters show no fears/qualms/reservations about making forecasts. They rely instead on track record.

Thus, in the example given, the question ought to be how many times has Villegas been right?

BTW, the “irrational exuberance” issue is back. This time with the hot real estate market as the focus.

Many of us got singed by the tech bust which todate has not yet fully recovered as reflected by NASDAQ figures. And is expected to never will. Thus those who jumped again into the real estate market are now proceeding with more deliberate steps.

Anonymous said...

With regards to pooling together, there is that axiom/theorem/saying that if you average people's forecasts together, you get a truer forecast. We've done this "experiment" in class a number of times (e.g., guess what the right answer is), and the average had a better track record that any individual person.

I guess another forecasting tool that leverages this is the betting pool. Have people bet on certain outcomes, and this market would make a good prediction tool. I remember the bruhaha before when the pentagon suggested having a terrorism futures market, but quickly shut it down after the expected public outcry.

AT

Econblogger said...

anonymous,

If that last point is correct, then this suggests that forecasting is a public good (i.e. apart from its own informative content, it contributes to a collective pool of forecasts). Since forecasting appears to be private sector dominated, doesn't this mean that forecasting is underprovided?

The axiom should be further clarified - which opinions to be pooled? will it work with any tom and dick and harry? Or should it be limited to experts?

Still it's a fascinating claim, and certainly contains a germ of truth.