For five years I worked at the [Commerce Department] and did econometric modeling, and you could never get these numbers to load into consumption numbers. They're not leading indicators or coincident indicators. They don't tell you a lot.

There's not a big inflation problem. Under those circumstances, you get a benign yield-curve inversion.

In the meantime, the data on the US economy (especially employment and industrial production) will look very weak in the next few months.

I look at that as really good news for the market; you want something right in the sweet spot. We're just in a nice, smooth acceleration, and now people are starting to think that's likely to be sustained.

I'm more interested in whether the losses we'll see on Friday are a permanent feature on the landscape or just the result of people pushed out of work due to Katrina and Rita. I think the market will look at it and shrug it off and look at what's going to happen in October.

I think this will be more of a company-led expansion instead of one led by consumers.

The biggest thing that would help the markets right now would be a sustained drop in crude oil, but I'm not holding my breath for that. It's drawing attention away from third-quarter numbers that, despite GM, are really pretty good.

That makes it definitely worth watching how bonds are trading. When there are problems at a company, the institutional bond investors tend to figure it out before the rest of the retail market and dump their bonds.

It reminds me of PIO, or pilot induced oscillation, a lag in the plane's response to a pilot moving the stick, ... You push the stick down and the plane doesn't go down right away, so you push it down more, and then it really starts to fall. The same thing happens in monetary policy if you have gradualism in interest rates.