Personal Consumption Expenditures and the Stock Market
Posted in Economic Indicators
August 13, 2010
A reader, citing the book Ahead of the Curve by Joseph Ellis, inquired about the hypothesis that consumer spending drives economic cycles and is therefore a leading indicator for the stock market. For example, Mr. Ellis presents a chart relating quarterly annualized change in Personal Consumption Expenditures (PCE) to quarterly change in the S&P 500 Index and states: “Most bear markets…proceed as (the rate of growth in) consumer spending continues to slow, and are largely over by the time recessions…are under way. Conversely, most bear markets end…when consumer spending and S&P 500 profits are at, or even prior to, their worst Y/Y comparisons.” Does PCE usefully predict stock market behavior? The Bureau of Economic Analysis releases seasonally adjusted PCE monthly with a lag of about one month via Line 24 of Table 2.6, “Personal Income and Its Disposition, Monthly.” Using this series and S&P 500 Index data for January 1959 through June 2010 (618 months), we find that…
The following chart compares the behaviors of seasonally adjusted PCE and the S&P 500 Index over the entire sample period. Although both series generally rise over time, the large mismatch in volatilities makes it very difficult to discern any tradable relationship between the two series.
For a closer look, we compare monthly changes in the two variables.
The following scatter plot relates the monthly change in the S&P 500 Index to the same-month change in PCE over the entire sample period. The Pearson correlation between the two series is 0.05 and the R-squared statistic is 0.004, indicating that changes in PCE explain less than 1% of S&P 500 Index returns at a monthly horizon.
Using “core” PCE (excluding food and energy expenditures) slightly weakens the relationship (R-squared 0.002).
Do changes in PCE, as indications of the economic trend, reliably lead changes in the stock market? Or, might changes in the stock market, as indications of fluctuations in personal wealth, reliably lead changes in PCE?
The next chart shows the Pearson correlations for various lead-lag scenarios between monthly changes in PCE and monthly changes in the S&P 500 Index, ranging from stocks lead consumption by 12 months (-12) to consumption leads stocks by 12 months (12). Although there is a little evidence that the stock market leads rather than lags PCE, all correlations are small.
What if the relationship is cumulative, visible not with monthly changes but perhaps with quarterly changes?
The next chart shows the Pearson correlations for various lead-lag scenarios between quarterly changes in PCE and quarterly changes in the S&P 500 Index, ranging from stocks lead consumption by six quarters (-6) to consumption leads stocks by six quarters (6). A reasonable interpretation of results is that there is some tendency for the stock market to lead PCE by one to three quarters, but there is no tendency for PCE to lead the stock market.
As a final test that discovers non-linearity in a relationship, we rank S&P 500 Index returns by lagged changes in PCE.
The final chart shows the S&P 500 Index average monthly return by quintile of monthly change in PCE lagged by two months to avoid using data that would not be available to an investor. For example, the Bureau of Economic Analysis typically releases PCE for January at the end of February, so an investor has the January PCE to make a decision about investing for March. This offset corresponds to that indicated by “2″ in the second chart above.
The chart presents results for the overall sample period (123 instances per quintile) and, as a robustness check, for two equal subperiods (61-62 instances per quintile).
Results suggest that the lowest (most negative) changes in PCE, as announced, indicate relatively strong stock market returns the following month (the average monthly return for all months in the sample is 0.6%). A possible interpretation is that the stock market is rebounding after poor returns that preceded the worst PCE announcements. However, lack of monotonic progression across quintiles and inconsistencies between subperiods undermine this conclusion. Omitting the five strongest stock market months from the 123 in the lowest PCE quintile for the total sample brings that quintile’s average return close to the average for all months.
In summary, evidence from simple tests does not consistently support a belief that changes in Personal Consumption Expenditures (PCE) usefully predict stock market returns at monthly and quarterly horizons. What evidence there is suggests that a strong monthly contraction in PCE, as announced, indicates relatively good stock market performance the next month.
