U.S. Economic Trend Update | 9.28.12

With a full set of July indicators in hand, along with a nearly complete set of August numbers, the data is telling us that the economic trend has weakened. But the decline in the trend is nowhere near the danger zone for recession risk. Nonetheless, the 3-month moving average of the Capital Spectator Economic Trend Index (CS-ETI) has dropped modestly for each of the three months through August and so the potential for trouble down the road can’t be ignored in the current slow-growth climate.

Leaving guesses about the future aside for a moment, here’s what the recent past looks like based on the numbers published so far:

August witnessed another indicator that slipped over to the dark side: oil prices rose relative to year-earlier levels last month (based on monthly average prices; price changes are inverted for CS-ETI). That’s the first year-over-year jump since March for crude prices. Even so, the majority of indicators are still trending positive, as the table above shows.
Taking the 3-month average of these indicators and tracking the changes through time still leaves us with a comfortable margin of safety relative to where recession risk has historically become a problem. As you can see in the next chart below, the August reading of CS-ETI (3-month average) fell to 75.5%. In other words, 75.5% of the indicators are trending positive. But the level has been slipping lately? Is that noise, or a sign of a new round of weakness? No one really knows at this point, but it’s clear that a 75.5% reading is still far above levels linked with economic downturns. It’s another story if CS-ETI continues to retreat, but for now it’s premature to argue that the economy is caught in a fatal cyclical swoon.

For some perspective on what the near-term future may bring, let’s turn to ARIMA forecasts for each of the indicators and then aggregate the individual predictions for a read on where CS-ETI may be headed in the near-term future.1 Any one forecast is likely to suffer error, of course, but predicting all the indicators in a robust econometric framework should minimize the risk a bit. With that in mind, the next chart suggests that the trend is set to modestly weaken again in September and then stabilize in October. That’s a sign that we should manage our expectations for the economy, but it’s still well short of an argument for expecting the worst. (Note: the “forecasts” for September’s market numbers (stocks, interest rates, and oil) aren’t really forecasts. Instead, I’ve taken the average prices for each market in September based on available daily data to date. Accordingly, the ARIMA forecasts for market data begins with October. For the economic series, the ARIMA forecasts start with September.)

As for what we do know, it seems clear that August wasn’t the start of a recession, despite what you might have heard elsewhere. Short of a dramatic and broad downward revision in the data for last month, the business cycle managed to eke out another month of growth. But the trend is weakening, which is no trivial matter in a world still struggling with lots of risk factors, including slow growth. Based on what we know today, however, the case for slow growth still looks like a reasonable forecast.
Will September data tell us another story? The first round of clues are scheduled to arrive next week, starting with the ISM Manufacturing Index report for September, which will be released on Monday, October 1. The consensus forecast anticipates a slight decline, according to Briefing.com. If so, that would leave the ISM index just under a neutral reading of 50 for the fourth month in a row. Not good, but not particularly fatal either. (For some perspective on the ISM index for manufacturing, see my report on the August data.)
Finally, a quick update on the methodology for CS-ETI. Starting with this report, I’m taking the average of four employment indicators (the blue cells in the table above) and using that average as the representative variable for the labor market. The adjustment doesn’t change much in terms of CS-ETI’s historical record. Ok, so why the change? Primarily it’s an issue of using one data series to represent the labor market rather than four. As a result, CS-ETI’s fluctuations move a step closer to an equal-weighted measure of the economy’s various signals—a change that should deliver more robust readings of the cyclical trend.

* * *

1. The ARIMA forecasts are calculated in R software, using Professor Rob Hyndman’s “forecast” package, which optimizes the prediction model based on each data set’s historical record.^