Whenever the Fed begins lowering interest rates Wall Street pundits can usually be heard saying, "In 12 to 18 months the stock market is typically higher after the Fed begins expansionary monetary policy." The fact is that this statement does not always prove to be the case.
Knee-jerk stock buying after a rate cut may not turn out well if you don't do your homework first. This article shows that this piece of simple advice offered by Wall Street pundits could leave the investor sorely disappointed. In addition, this article looks into slightly more complex trading strategies regarding a Fed expansionary cycle and demonstrates that investors should take into account the slope of the yield curve and the stock markets valuation before jumping into the market head first. (For background on how the Fed can move the market, read How Much Influence Does The Fed Have? and The Federal Reserve's Fight Against Inflation.)
Broadly speaking, it appears that expansionary monetary policy is beneficial for stock market investors. Yet 12-month and 18-month stock market returns, as measured with the S&P 500, during a Fed expansionary cycle are probably conditioned primarily on the slope of the yield curve and then on the valuation of the market at the beginning of the period. That is to say, that all 12- and 18-month periods after the start of a Fed expansionary cycle are not created equal.
Because stocks are typically higher 12 to 18 months from any start date, it really doesn't take a financial genius to pretend to be a financial genius by uttering that stocks should be higher in the future. So, this is not really what the pundits are saying. Instead, they are implying that the stock market will be up more than the typical amount, not just an uninspiring 1%. Unfortunately for the gurus mentioned above, stock market returns 12 and 18 months after the start of an expansionary cycle are not statistically different from their typical 12 and 18 month returns.
The data being used in this study may not be perfect for the task, but it should be able to capture enough of any 12 to 18 month trends to be considered a reasonable data set with which to examine relationships.
There have been 12 Fed Expansionary Cycles, which are listed in Figure 1 below, which are defined as the next 12 and 18 months following a peak in the Fed funds rate. A peak occurs when the Fed funds rate has risen by at least 1.5% and then falls by at least 1.5% before rising again by at least 1.5
The slope of the yield curve criteria is a way of measuring how tight monetary policy has been prior to the start of an expansionary cycle. Historically, negative sloping yield curves have been pretty accurate indicators of significant economic slowdowns or recessions in the near future. Stocks typically don't perform well heading into a recession. So, adding this variable should be beneficial to explaining stock returns by highlighting periods where monetary action will be overwhelmed by other factors in a recessionary economy. (To learn the basics of yield curve calculation, read The Impact Of An Inverted Yield Curve.)
If the three-month U.S.Treasury bill yield was currently or just recently higher than the 10-year U.S. Treasure Note yield then the cycle period was classified as a "negative slope cycle". If the three-month was below the 10-year, the cycle period was classified as a "positive slope cycle". There are seven cycles categorized as positive and five cycles categorized as negative (see Figure 1).
The valuation criteria is a way of measuring the likelihood that the stock market will rise due to the expansion in the valuation multiple, the P/E multiple in this case. The expansion of the valuation multiple will give a big kick to stock market performance, especially if in concert with rising earnings.
If the current price-to-10-year-average-earnings ratio was greater than its median up to that point in history, the cycle period was classified as "high valuation". If the current price-to-10-year-average-earnings ratio was less than its median up to that point in history, the cycle period was classified as "low valuation". There were only two low valuation periods and nine high valuation periods (see Figure 1.)
|Month the Expansionary Period Starts||Cumulative
|S&P 500 Valuation||Slope of Yield Curve|
|Geometric Average During Expansionary Period||10.8%||20.3%||-||-|
|Geometric Average During Other Periods||9.8%||15.0%||-||-|
|Figure 1: The start date of Fed expansionary cycles and subsequent 12- and 18-month S&P 500 returns. Red text highlights returns below the "Other Periods" average|
|Source:Prof. Robert Schiller\'s S&P 500 data and federal funds data.|
In the early '80s, the Fed funds rate bounced around a lot and triggered several "cycles" in very short time periods. This volatility was not due to a cycle change, but was due to the fact the Fed targeted non-borrowed reserves that made the Fed funds rate very erratic. So, the overlapping of cycle time periods was adjusted for by keeping only the time period with the best 18-month stock return scenario. In essence, giving the resulting 18-month period returns the best shot at out performance.
As Figure 1 illustrates, 12 to 18 months after a Fed expansionary period, stocks are up on average 10.8% and 20.3% respectively. This certainly seems to validate the pundits' point that the stock market will not only be higher 12 or 18 months out, but returns will be bigger than the respective 9.8% and 15.0% averages during the periods not included in the 12- and 18-month expansionary periods.
The problem with this conclusion is that it compares just one data item - an average, between two sets of numbers. In order to properly compare averages you need to look at all the returns in each set. For a simple example, let's take a look at these two sets of numbers: Set A is -20, -5 and 55 and Set B is -30, 20 and 25. The simple average for Set A is (-20-5+55)/3 = 10 and Set B is (-30+20+25)/3 = 5. You can see that just comparing the averages doesn't give you the whole picture. So, in order to determine if the returns during 12 and 18 months of an expansionary cycle are really bigger than the periods that are not included in this group, we need to use some statistical analysis.
This statistical analysis was done by first ranking all the returns from highest to lowest, then categorizing a return into the "expansionary" data set or the "other" date set. A T-statistic for samples of unequal sizes and unequal variance was calculated and used for statistical testing. Based on this analysis we can say that there is no evidence that the difference in the mean of return ranks between the two categories is due to something other than chance. In short, even though the averages during expansionary cycles are bigger, once you look at the whole picture of how the returns are distributed you can't say that the averages are bigger due to anything other than chance.
The Slope of the Yield Curve and Stock Market Valuation Matter
Now, the 10.8% and 20.3% returns are averages, and you can see that there were three 12-month periods and two-18 month periods where returns were negative. So, just because the Fed begins an expansionary cycle does not mean stock market gains are in the bag. If you look a bit more carefully you see that each of these three poor-performing 12-month periods all started with a negative sloping yield curve and a high market valuation. There was only one period out the four periods that started with negative sloping yield curve and a high market valuation that had better than average returns.
|Figure 2: 18-month cumulative total returns for the S&P 500 over 11 Fed easing cycles since 1957.|
Figure 2 shows how stocks are dependent on the slope of the yield curve as well as on valuation:
- Black lines: Returns generated in expansionary cycles that began with a positive sloping yield curve.
- Red lines: Returns generated in expansionary cycles that began with a negative sloping yield curve.
- Dotted yellow line: The expected return for a typical "other period" over the given time frame.
- Triangles: Expansionary cycles that began with low valuations.
- "Xs": Expansionary cycles that began with high valuations (all but two cycles.)
Cumulative Total Returns
As the graph illustrates, the positive sloping yield curve periods (the black lines) are basically above the negative sloping yield curve periods (the red lines.) In addition, the two low valuation periods (the triangles) are basically above the high valuation periods (the Xs.)
Before we make any conclusions let's look at the whole picture and perform the statistical tests as described above in order to compare returns between the four types of periods listed below and the periods which are not included in the respective group.
- Expansionary and Positive Slope
- Expansionary and Negative Slope
- Expansionary and Low Valuation
- Expansionary and High Valuation
The statistical analysis indicate that there is something other than chance that makes returns bigger in expansionary and positive slope periods, bigger in expansionary low valuation periods, yet lower in expansionary and negatively slope periods. Thus,
- The conditions of 1) Expansionary and Positive Slope and 3) Expansionary and Low Valuation are the only time the pundits' verbiage holds any water and investors can expect abnormally large gains.
- Importantly, condition 2) Expansionary and Negative Slope can turn out to be completely contrary to what the gurus sing and investors should expect to experience lower returns and possibly even losses.
With a 12-month holding time frame, it appears:
- There is no difference in average returns between Expansionary "All Conditions" periods and "All Other" periods. This is the same story for Expansionary High-Valuation, too.
- Average returns are larger for Expansionary Positive-Slope periods and Expansionary Low-Valuation periods than for their respective all-other periods.
- Average returns are smaller for Expansionary Negative-Slope periods than for all other periods.
Note: With an 18 months holding time frame, it appears things are pretty much the same as in the 12-month holding time frame except that 3) is no longer true. There is no statistical difference in average returns between Expansionary Negative Slope periods and All Other periods.
Although we only looked at 11 sample periods, the data suggest that simply buying stocks when the Fed lowers interest rates may not be the most productive strategy. It appears that not all Fed expansionary cycles are likely to behave the same when it comes to the equity market's performance over the following 12 to 18 months. Thus, investors should look at a few more variables before jumping into the equity market when the Fed begins lowering interest rates.
This article encourages investors to look at the slope of the yield curve and the market's valuation as further variables to consider before buying stocks after a Fed ease. If the Fed begins lowering interest rates after a downward sloping yield curve, then returns may not be so spectacular. On the hand, if the Fed starts easing with and upward sloping yield curve and/or the market valuation is low, then stock market performance should be more rewarding.
For related reading, check out Bond Yield Curve Holds Predictive Powers.