Market Timing: Should You Attempt It?: Re: Re: Market Timing on the QT

» Normxxx - Re: Re: Market Timing on the QT

In response to message posted by Kirk:

Thanks for the thanks.

I can't say that a good market timing model improves the results of a well balanced portfolio over a long period of time, but it does seem better at letting you sleep nights. As I said elsewhere, I have been out of the market (except for a small number of shares) since May 19, and so missed out on much of the summer fun. But on the other hand, I was out all last summer until October, and so missed out on that peak adrenaline time. Also, my REITs, gold, and emerging markets have done very well, so I can't complain.

You are right about having to become (at least semi-)expert in each of the classes. You have to know enough to find several advisors for each class who are not feeding you some version of their private truth or worse, and you must constantly monitor that these gurus do not develop some fixed idea, so that they warp while the asset class woofs. It seems to be only natural, from time to time, that all gurus and systems fail. You have to be conversant enough with the class (and have some "circuit breakers") so that you are not dragged down with them, but not so sensitive that you get out just before the system kicks in again. That's why you need several gurus. Don't expect them to agree, but when they all don't agree with each other or they all do agree with each other, that should alert you to examine what is happening closer and look for other info sources.

I am familiar with Ed Yardeni's Fed model and, in fact, use a version of it to determine overall stock and bond risks. However, there are problems with the model.

1. There has been a lot of discussion of whose or which earnings to use. Before about 1996, operational earnings were about the same as GAAP earnings, so do we use estimated GAAP or estimated operational earnings going forward?. Maybe we should use trailing earnings, since Wall Street's estimated earnings seems to have about a 15% growth bias built in (which, like the stopped clock, is sometimes, but rarely correct).

2. You certainly can't use it for timing (which is why I use it in determining asset class risk); it has large periods when it is out of step with the market. It is really a warning of when stocks or bonds, as a class, are getting overbought or oversold (relative to each other), in the long term. But you already knew that.

3. Models do not necessarily work if some of their input parameters are artificially constrained or if they are at the extremes. Currently, you can argue that our interest rates are being held artificially low (not just by the Fed, but by China, Japan, etc.) and, in fact, are abnormally low on a real basis (currently, the real short term interest rate is about a negative 1% or so (shades of Japan).

4.1 The Fed model works well in a world where recessions are caused by the Fed shutting down demand through raising interest rates because of the economy overheatng. When rates are subsequently lowered, pent up demand in housing, autos, and CAPEX zoom. It is hard to see much pent up demand in housing and autos, which have gone gangbusters throughout this downturn. Or in CAPEX, where capacity utilization is still hovering around 75%. INTEL just came out with a relatively glowing report (although, if you read the fine print, you will see they are not so certain of 2004). Yet INTEL just cut their CAPEX budget. And several of the other biggies have spent less than 50% of their 2003 IT budgets. (Unlike the government, I don't see them scrambling to spend the rest in the last 3 months of the year.)

4.2 Currently, we are in a(n over)supply side recessionary environment (which is why deflation is a threat and the job market is so poor). Our last experience with such an environment was in the '30s, and we didn't cope very well. (Nor has Japan in the '90s.) Although currently we seem to be doing much better, if the economy doesn't kick start here, we are in serious trouble. We are out of both fiscal and monetary stimuli (regardless of what Fed governor Bernanke says). The acid test will be in 1Q 2004 when the echos of the current stimuli have faded away.

5. Earnings are being (artificially?) maintained by drastic cost cutting; so far, there has been little improvement in revenues.

6. The model may be saying that bonds are way overpriced (which they probably are, since interest rates are being held artificially low).

I believe that the market is doing a repeat of late 1999 and for the same reason: a vast (over?) supply of money and credit which has to find a home. It will tank again in 2004 if it looks like a double dip, or if the financial system otherwise locks up. At this point, I think a repeat of 9/11 would be much less of a stock market event, except to let off a lot of the pressure. If it happens before next year, or if next year looks good (we muddle through once more), then I think it would be a great buying opportunity. But if it looks like a double dip, look out below, and I don't think it will wait for May to crash (as it didn't in 2000).

I have much more to say on this, but this is more than enough for one message!

WHAT'S WRONG WITH THIS PICTURE?

-- posted by Normxxx