Political Calculations really is unlike just about every other blog you might find out there.

For a lot of readers, we suspect that we fit the description of being "that blog with all those interesting tools." But what we really are would not necessarily seem to be a lot different from that description. We ask questions for which we don't already know the answers and we answer them. We make tools along the way because they help us answer those questions and we go the extra mile and make them available to you too. The payoff is that we all know more at the end than we did when we first asked the question and we can often take what we found out and put it to practical use.

But what makes us really different is that we bring you into our process of discovery (see this post for a really good and recent example!) We do a good portion of the analysis we present "live" and in many cases, we're no more than maybe a day ahead of what we write!

The trouble with all that starts because the path of discovery is almost never a straight one. For every seemingly simple solution, there are twists and turns that we never see coming and we'll find ourselves having gone down a blind alley before we know it.

We've gone down blind alleys like these before during the course of our major projects. And as we discovered over the weekend, we've just gone down one again!

Here's how it happened. We have developed a method to measure the level of distress in the stock market. There are two major pieces of data that go into the method's calculations: the year over year rate of growth in stock prices and the year over year rate of growth in stock dividends per share. The key driver in the relationship between these two factors is the dividend rate of growth, which is significant because when it approaches a value of zero, the level of distress in the stock market peaks.

Back on 16 January 2008, following a series of dividend cuts in major mortgage and financial companies that triggered a breakdown in the order that had existed in the stock market since July 2003, we turned to dividend futures as a source of data that we believed we would provide the most up-to-date and real-time view for tracking the progress of building distress in the stock market.

In doing that, we made a rookie mistake. We had interpreted the data as being the value of dividends per share that S&P 500 stocks would pay out in total for the first quarter. Instead, it turned out to be that the data really represented the value of dividends per share that would be paid out between the current date and the end of the quarter.

As such, the value of dividends per share will always decrease as the companies paying out dividends actually pay them during the course of the quarter. Over time, and by the end of the quarter, these dividend futures will eventually hit zero, not because corporate dividends are being slashed, but because they've already been paid! Because that's the case, we didn't pick up that things weren't quite right early, because the numbers weren't that far off from what they should have been. But, as more time went by, the discrepancy became greater.

Here's an example of that effect. Back on 16 January 2008, near the beginning of the quarter, the level of dividends per share was high enough to look like the reduction in dividends to be paid in the first quarter of 2008 was the result of the financial companies in the S&P having cut their dividends. So we ran with them. And unfortunately, we kept running with them.

Ultimately, we finally figured out that we had the error in our data because we kept asking questions. The question that directly led us to finding our mistake was this one: "If dividends are being cut so much for the current quarter, why aren't they changing more in the farther future?" And the reason, as we discovered, is because the dividend futures work in the way that we've described above.

The good news is that things aren't getting as bad as fast as it might have otherwise appeared using our previous analysis. The bad news is that whatever distress is now occurring in the market is more likely to be drawn out over time than what the later portion of our erroneous data-driven analysis had suggested.

Beyond this, Standard & Poor has confirmed that its forecast for dividends per share for the S&P 500 will still likely come in at $30.30 for 2008, a 9.3% year over year increase. This increase is expected even with several major financial companies having cut their dividends as other healthier companies have acted to increase theirs. The practical upshot of all this is that we'll be returning to what S&P is forecasting in our models. And now that we understand what the dividend futures are really communicating, we can make more effective use of them as we go forward as another tool in our analytical toolbox.

Perhaps no one really understands this, but dealing with this kind of problem is really kind of thing we live for. It's like the adrenaline rush you get from going down a blind alley that no-one else knows you've gone down, working out how you got there and how to get out, and then the satisfaction of getting back onto the real path to discovery after having been diverted from it. That's where the real fun in the art of analysis is to be found.

As they say, a straight line may be the shortest distance between two points, but it is far from being the most interesting one.

Labels: none really

- posted by Ironman at 8:39 AM | Permalink | Home |