- 1997: Volume 6, No. 3 The Utility Stock Index: Does It Lead The Bond Markets?

Conventional wisdom states the utility stock index leads the bond market. We decided to investigate this relationship to determine if there is evidence to support this long standing belief.

Utility stocks are similar to bonds because they provide the investor with a relatively fixed income. Utilities are legal monopolies. As a result, they are regulated by the government. In their service contracts they are guaranteed rates that will insure a specific level of profit. Historically, utilities payout a consistent percentage of their profits in dividends. A [fixed profit] x [fixed percentage payout] = [fixed dividend] to the shareholder. A fixed income to the investor.

Correlation Analysis

In our first test, we employed correlation analysis. Correlation analysis measures the degree of association between two data series. The product of this mathematical comparison is a number called the correlation coefficient. Correlation coefficients range from -1.0 to +1.0 and identify the direction as well as the strength of a relationship between the two variables. A zero coefficient means there is no relationship between the two variables. A positive coefficient indicates on average that the two series move in the same direction. And the stronger the positive relationship is, the closer the coefficient will be to +1.0. A negative coefficient indicates that the two series move in the opposite direction. And the stronger the negative relationship is, the closer the coefficient will be to -1.0.

Figure 1

If the movements of the utility stock index lead the movements of the bond market, the correlation between the Philadelphia Utility Stock Index and the T-Bond future will be greatest when the T-Bond is lagged a number of periods. Looking at Figure 1 below, series B is lagged five periods behind series A (i.e. what occurs to series A in one period, occurs to series B five periods later). For example, series A made its low in period nine and five periods later in period 14, series B made its low. Series A advanced in periods 10 through 13 and series B did the same five periods later. In period 13 series A put in a short-term high and five periods later-period 18-series B made a short-term high. If a correlation was run on the two series, the correlation coefficient of series A and series B lagged five periods, would be +1.0-a perfectly positive correlation.

In this first test we did a correlation between the Philadelphia Utility Stock Index and the T-Bond future. The analysis was done using daily data for the past one year, weekly data for the past five years and monthly data for the past 10 years. The T-Bond was lagged 0, 1, 2, 3, 4, 5 and 10 periods in the test.

Looking at Figure 2, the correlation coefficients between the price movement of the utility index and the T-Bond was highest when there was no lag. When the T-Bond was lagged 1, 2, 3, 4, 5 and 10 periods, the correlation was lower. The results of this test indicate the day-to-day, week-to-week and month-to-month price movements of the utility stock index do not lead the bond market.

Figure 2
Correlation Coefficients

Lag	1 Year Daily	5 Year Weekly	10 Year Monthly
0 1 2 3 4 5 10	.89 .88 .86 .85 .83 .81 .75	.71 .71 .71 .70 .69 .68 .60	.90 .89 .85 .84 .82 .80 .65

The first test compared every utility index price with the corresponding T-Bond price throughout the test period. More precisely, the conventional belief is the index's reactionary highs and lows lead the T-Bond's highs and lows. So some proponents of the belief might say a fairer test would be to focus on the turning points and see how often the index's turning points led the T-Bond's turning points.

Reactionary Highs and Lows

To address that request, first using the moving average channel we identified the reactionary highs and lows (or turning points) of the utility index and the T-Bond future. [The top of the moving average channel is equal to the 10 week moving average of the weekly high prices. The bottom of the channel is equal to the 10 week average of the weekly low prices.] When the market crossed above the channel, the previous low price was labeled a reactionary low. Then when the market crossed below the channel, the previous high price was labeled a reactionary high. The labeling of successive lows and highs continued this way as each market crossed above and below its channel (See Figure 3).

Figure 3

Second, we looked at the index and T-Bond charts and matched-up the corresponding turning points. Given the two markets are strongly correlated, for each utility index high there was usually a T-Bond high that occurred about the same time. And for each index low there was usually a corresponding T-Bond low.

Third, we examined each utility index turning point and noted the number of times the corresponding T-Bond turning point led, lagged or occurred at the same time. We also recorded the number of weeks the T-Bond's turning points led or lagged.

We followed this three step approach using the Philadelphia Utility Stock Index and T-Bond weekly price data of the past nine years. During the test period the index had 39 turning points. Thirty-five times the T-Bond had a corresponding turning point-15 times the T-Bond's turning point occurred after the index's, eight times it occurred before and 12 times the T-Bond's high or low occurred the same week as the index's turning point. four times the T-Bond did not have a corresponding turning point (i.e., the T-Bond kept trending despite the occurrence of a utility index high or low-See Figure 4).

Figure 4

The Bottom Line

Similar to the results of the first test, these results do not support the belief that utility index turning points usually occur before T-Bond turning points. Only 15 out of 39 times did an index high or low occur before a T-Bond turning point. Twenty-four times the T-Bond's turning point led or occurred at the same time or a T-Bond turning point did not occur at all.

These results indicate the utility index is a coincident indicator. Sure, sometimes the utility index leads the bond market, but sometimes the bond market leads the index and sometimes they move together. However, on average, and that is what determines whether a tendency exists or not, the utility index and bond market trend and reverse together.

Alex Saitta is a technical analyst at Salomon Brothers. Jason Wang, who also works for Salomon Brothers, contributed to the article.