Misleading Divergence Between Leading Indicators

Back in the days when traders used phones to place orders, I formulated the rule that if you didn't know where the market was going, you were sure the person to whom you were speaking did. We've lost this valuable layer of telephonic groping in our headlong rush to computerized trading.

A corollary applies to economic analysis: If you don't know where the economy is headed, there must be a reliable indicator that does. The Conference Board has been publishing its monthly index of leading indicators since January 1959; they include:

1. Average weekly work hours
2. Average weekly initial jobless claims
3. New manufacturing orders
4. Vendor performance
5. Manufacturers’ new capital orders
6. Building permits
7. The S&P 500
8. M2
9. The yield curve spread and
10. Consumer expectations

The Economic Cycle Research Institute has developed a weekly index going back to 1967 with a high degree of overlap in its components:

1. Growth rate in M2 plus long-term household mutual funds
2. Growth rate of Journal of Commerce-ECRI industrial price index
3. Initial unemployment claims
4. Moody’s seasoned BAA corporate bonds
5. NYSE Composite index
6. BAA Corporate -- 10-year Treasury spread and
7. Mortgage applications, purchase index

Over time, the two indices have mapped each other closely, until recently. The ECRI index dropped much faster during 2008, and has turned lower while the Conference Board’s index has continued to move higher. Clearly, both indices can't be right simultaneously.



But we're not here to engage in methodological bashing or any quasi-theological discourse. No, we're traders and investors! As the ECRI index’s latest dips and doodles look suspiciously like our beloved stock market, let’s map it against the weekly S&P 500 index on a logarithmic scale with the ECRI index led one week.



Your eyes aren't lying. The r-squared, or percentage of variance, explained for the ECRI index by the S&P 500 is 0.97. We can start playing around with other components of the index to solve those large explanatory gaps before 1983, but the gain is small for the effort involved. Moreover, the recent fit is suspiciously tight.

Where does this leave us? In an uncomfortable position, that’s where: If all of these indicators can be collapsed into the S&P 500, ol’ Mr. Flash-Crash himself, we're left wondering whether the stock market is a forecasting device. The mind rebels and the stomach surely follows. Paul Samuelson quipped, “The stock market has forecast nine out of the past five recessions.” Or, as I have said, “These are not GDP futures.”

We're left with an unknowable stock market looking like a forecasting device for an unknowable economy. Yes, they give you a round ball and a round bat and tell you to hit it square. And as far as that person on the other end of the phone, forget it. All that voice wants to know is whether you'd like to continue the conversation in Spanish; if so, please press 2.