Skip to main content

You are here

The US small cap premium and the business cycle

  • The US economy’s current expansion has lasted more than 85 months and, if history is any guide, may continue for some time to come. Nonetheless, it is natural for market participants to wonder whether we may be approaching the end of this economic expansion and what might happen to the stock market if and when the business cycle reaches a new peak and an economic downturn occurs again.
  • This is especially relevant for small company US stocks (“small caps”) as they tend to be more directly connected to the US domestic economy than their larger brethren, which often have global sources of revenue. Our representation of US small caps is the Russell 2000 Index, which incorporates the smallest 2,000 stocks of the broad market all-cap Russell 3000 Index. These 2,000 stocks account for roughly 8%-10% of the total market capitalization of the Russell 3000.
  • The main focus of our analysis is on the behavior of the small cap “premium” over the business cycle, which we will measure as the return difference between the Russell 2000 and the Russell 1000 indexes. The Russell 1000 Index incorporates the largest 1,000 stocks in the Russell 3000 Index, representing 90-92% of total market capitalization. The designation of premium for this return difference is based on extensive academic literature identifying a positive return difference between small caps and large caps over the “long run.” Academics have termed this a risk premium, a reward for shouldering the additional risk associated with smaller companies. One must quickly add that the premium is often negative for periods shorter than the long run.
     

Previous research and summary of current findings

How have US small cap stocks as represented by the Russell 2000 performed historically at the end of an economic expansion? Previous research has shown a tendency for small caps to move into positive territory in advance of economic troughs, before recessions end. In the spring of 2009, for example, in the midst of the Great Recession, a Russell paper reported that across the four prior US recessions, both the broad equity Russell 3000 Index and the small cap premium (Russell 2000 Index – Russell 1000 Index) tended to recover prior to the trough, but the small cap premium tended to move earlier.  In 2012 these findings were updated to include the Great Recession. FTSE Russell has also published research on the historic performance of the Russell 2000 during periods of rising Federal Funds rates as well as periods of slow economic growth. In the present paper, we add to this body of work by examining the historical performance of the Russell 2000 around the five expansionary peaks of the US economy identified by the National Bureau of Economic Research (NBER) since 1979.

We first look at the US broad equity market as represented by the Russell 3000 Index during the last five expansionary periods, and then examine the historical performance of the US small cap premium during these same periods of economic growth. We find that over these cycles:

  • Both the broad market Russell 3000 and the small cap premium (difference between the Russell 2000 and the Russell 1000) turn negative on average 9 months before the peak, the peak being when the expansion ends.
  • Both the Russell 3000 and the small cap premium return to positive territory on average 3 months before the peak but the small cap premium on average is substantially above the average Russell 3000 return at this point (6.53% compared with 2.21%).
  • The Russell 3000 returns to negative territory on average at the peak itself, but the small cap premium, although reduced, stays positive at this point.
  • The period after the peak until the next trough encompasses the contraction/recession phase of the business cycle. The average small cap premium stays positive and is substantially above the Russell 3000 broad market return throughout the 12 months after the economic turning point. This is true even though the broad market return on average rises to 16.4% at +12 months: the average small cap premium was at that point over 7%. Both averages at the +12 month time point are dominated by the outcome of the July 1981 cycle.

Our findings support the position that the broad equity market is a leading indicator of the US economy. Perhaps the most surprising finding is the average forward return strength of the US small cap premium just before and during a recession.

Methodology

Historical performance data for the Russell US equity flagship indexes—the Russell 3000, 2000 and 1000 Indexes—begins in January 1979. Our sample period for these indexes spans five complete expansionary cycles of the US economy as defined and dated by the NBER. The US is currently in a sixth expansionary period, which began in June 2009. The NBER has not yet announced a peak as of September 30, 2016.

Table 1 reports the details for these last five expansionary periods from trough to peak, in other words, from the beginning to the end of each expansionary period, including the duration in months of each expansion. Finally, we include the date when the NBER identified the peak and the lag between the peak (end of expansion) and the NBER identification date. The current expansion which began June 2009 is the 3rd longest growth period in our sample. By far the longest of these periods was from March 1991 until the tech bubble burst in 2001, and the shortest was the brief 12-month period 1980-1981.

To assess the performance of the broad market as represented by the Russell 3000 Index, we computed the next or forward 12-month index return at 12, 9, 6, 3 months before and after the economic peaks, as well as at the exact market turning point itself. For the small cap premium of the Russell 2000 minus the Russell 1000, we calculated the 12-month returns to the small cap premium at the same months-from-peak points: 12, 9, 6, 3 months before and after the economy peaked, and at the peak itself. We report the average outcomes at each time point. We also add another performance measure that we call the “hit rate.” The hit rate at each time point is simply the percentage of the five periods where the Russell 3000 Index posted positive returns over the ensuing 12 months at each time point.

The broad US equity market (Russell 3000)

Exhibit 1 displays the results for our analysis of the broad US equity market based on the performance of the Russell 3000 Index. On the left hand side of the chart are 12, 9, 6 and 3 months before the economic peak, while the results for the periods of economic contraction 3, 6, 9, and 12 months after the peak are on the right.  The bars display the forward-looking 12-month total returns to the Russell 3000 Index averaged across the five periods in our sample at each months-from-peak point. The brown line charts the hit rate of positive returns, i.e., the percentage of the five periods where the Russell 3000 total return was positive over the ensuing 12 months.

Prior to the peak, on average the Russell 3000 produced low to negative returns as the economy ran out of steam. The hit rate declined precipitously from 80% at -12 months (i.e., in four of the-five instances at the -12 month point the returns were positive) to 20% at the -9 month point (in only one of the five instances did the Russell 3000 record a positive performance). Index returns improved only slightly at -6 months, with low or negative returns on average and a hit rate of 40% (two out of five) through the peak itself.

Subsequent to the economic high point, we enter the period of contraction/recession.  The average return levels improved after the peak, and by 12 months into the recessions on average the Russell 3000 recorded returns of over 16%.  With the exception of the +9 month mark where we marked a hit rate of 40% (only two out of five times was the ensuing 12-month return positive) — the hit rates during this part of the cycle were all at 60%— 3 out of 5 times, the Russell 3000 experienced positive performance in the ensuing 12 months.

All periods of expansion were not identical, obviously, and in Table 2 we report the forward-looking 12-month return at each period point for each of the five expansionary periods as well as the average and the hit rate.  For the expansionary period that peaked during March 1991, for example, all returns at all points were negative, although the loss was smallest at the peak itself. The July 1981 results accounted primarily for the magnitude of the positive average for the +9 and +12 month points. The pattern of average outcomes, while informative, were no guarantee of what would have happened during a given (or the next) expansionary period.

The small cap premium (Russell 2000 – Russell 1000)

Turning our attention to the small cap premium, we see a notably different pattern from that of the Russell 3000, particularly in periods after the peaks. Exhibit 2 displays the results for the small cap premium averaged across the five historical periods in our sample, centered on the economic peaks. The small cap premium has tended to be positive during recessions (the right side of the chart) while mixed during the periods of expansion prior to the peaks (left side of chart). We see that prior to the peaks (the center point on the chart), the average small cap premium was  slightly negative at the -9 month point, turning positive at the -3 month point, but then falling back to close to neutral at the peak itself. The hit rate (the percentage of the five expansions where the forward-looking 12 month small cap premium was positive at any give months-from-peak point) during expansions was 60% (three out of five) with the exception of the -9 months-from-peak point where it was  40% (only two of the five periods have a positive small cap premium).

The small cap premiums for the periods of contraction on the right were substantially positive. The average size of the premium increased notably at 3 months after the peak, after the economic downturn began, and was positive in four out of the five instances with a hit rate of 80%.  The hit rate increases to 100%—five for five—at the +6 month mark, and the average premium size was only slightly lower than at the +3 month point.  At 9 months, the hit rate was still high—80% (four out of five)—though the premium size on average declined.  The hit rate declined to 60% but the average was still positive—over 7%—at the 12 month mark.

As with the Russell 3000 analysis above, we caution that each period in our sample is unique, and all in some way deviate from the average outcomes charted above. The results for each of the five periods are reported in Table 3. Returns for the March 2001 period, for instance, were positive at each point and declined to close to zero at the +12 mark; the January 1980 experience was all positive except at -9 months. The July 1990 sample set had negative premium up to the +3 month time point, while around the July 1981 peak the premium was negative at the -3 and +3 month marks as well as at the peak itself. The high positive average return at the +3 and +6 time points were driven by the January 1980 index returns. Again, while we have identified a marked pattern in the average outcomes, we note there have been important differences in each expansion/retraction period.

Comparisons of the average outcomes for the Russell 3000 and the small cap premium at each months-from-peak point around the peak are the focus of the chart in Exhibit 3. This aids the reader in evaluating what is common to the broad US equity market and the small cap premium, as well as what sets the small cap premium apart. Across these five most recent expansionary periods on average both the Russell 3000 and the small cap premium turned negative in advance of the peak, at the -9 month point, but returned to positive territory at -3 months, where we observe a strong small cap performance on average. The broad market receded back at the peak itself, with a notable negative average outcome, but the small cap premium remained in positive territory at the peak. The Russell 3000 rejoined the small cap premium in positive return space at the +3 month mark and both stay positive through the last inception point in our analysis, at +12 months.

Conclusion

Market participants can never know in advance when an expansion might end and a new contraction might begin, nor how that might impact the equity market. A look back at history can be useful as history doesn’t repeat itself but it can rhyme. Looking at historical averages of the last five economic expansions and contractions shows:

  • The broad equity market as measured by the Russell 3000 has on average turned negative several months before a contraction/recession.
  • The small cap premium as measured by the difference between the Russell 2000 and the Russell 1000 has on average been positive for the forward 12 months beginning 3 months before the peak and through 12 months into the contraction/recession.

That the broad market is a leading indicator of recessions is nothing new. Many economists incorporate broad market returns in their forecasting models of the business cycle. However, that the forward 12 month small cap premium has on average been positive – meaning the Russell 2000 outperformed the Russell 1000 – from 3 months prior to the peaks through the recessions may be surprising to some. We hope market participants find these insights useful.

—Download the full paper—