« Disney Earnings Preview: Last Slow Quarter Before Growth Accelerates | Main | Disney Comes Through »

February 03, 2006

February Model Signals

For February, Northlake's Market Cap model continued to flash a Mid Cap signal as has been the case for every month beginning in September 2005. As a reminder, this model measures ten factors which I group as economic, interest rate, and stock market indicators. Each indicator has historically had predictive value for the relative performance of small caps vs. big caps. The model uses a weight of the evidence approach so that if the indicators generally favor small or large caps that is the signal that will be flashed. A mid cap signal is flashed when the indicators are split as they are currently. One interesting observation from the current status of the underlying signals is that most of the economic and interest indicators favor large caps while the stock market indicators which include especially the trend and measures favor small caps. Decelerating economic growth and moderately rising interest rates historically have favored large caps but for now investors still are supporting small caps. A key question is when will investors rotate decisively toward large caps on a relative basis if fundamentals continue in their favor?

Northlake's Style model shifted back to value for February after a one month sojourn in growth territory. This change led to the sale of all holdings in Russell Growth ETFs (IWF and IWO) and the purchase of Russell Value ETFs (IWD and IWN). Prior to January, this model had flashed a value signal for three consecutive months. The underlying factors in this model generally favor value across the economic, interest rate, and stock market indicators. However, most of the signals are weak which explains the monthly fluctuation in the overall signal. For February, three underlying indicators shifted. First, the valuation indicator moved in favor of growth. This indicator measures the relative P-E of the Russell 3000 Growth index vs. the Russell 3000 Value index. Historically, the growth P-E has average 1.45 times the value P-E. After three plus years where value looked cheap, recently, the relative P-E has normalized on a historical basis. Last month, growth moved very slightly into the cheap mode, thus the indicator switched....

The second change for February occurred in advisory service sentiment which measures bulls vs. bears using Investors Intelligence data. This is a contrarian indicator that shifts when sentiment moves toward extreme readings. This month the indicator shifted from growth to value as bullish sentiment rose considerably. The model reads this a reason to move into the lower beta, less risky category which in this model is value.

The final shift for February was in the technical indicators. These indicators measure intermediate trends covering 6 to 12 months. This month the indicators actually went from a reading favoring growth, acknowledging the trend mostly in place since last spring, to a reading favoring value, picking up on better value performance over the past 4 to 6 months.

The indicators offered mixed accuracy for January. The mid cap signal was a good one as using the price only performance of the tracking ETFs, the S&P 400 (MDY) gained 5.3% against just 2.4% for the S&P 500 (SPY). However, the most accurate signal would have been small cap as the Russell 2000 (IWM) gained 8.4% in January. The mid cap signal that has been in place since September has proved accurate so far as MDY is up 8.8%, IWM is up 8.8% and SPY is up just 4%.

The Style model did not produce an accurate signal in January when it favored growth. I look at several indices to determine the accuracy of this model. Last month, the Russell 3000 Growth (IWZ) and Value (IWW) and the Russell 1000 Growth (IWF) and Value (IWD) saw better performance for value. Only in small caps was the signal accurate as the Russell 2000 Growth (IWO) gained 9% vs. 7.8% for the Russell 2000 Value (IWN). My implementation of the growth signal for January was weighted 75% Russell 1000 and 25% Russell 2000. This produced a return in excess of the S&P 500 but left a fair amount on the table.

Posted by Steve Birenberg at February 3, 2006 01:11 PM in Models