Have you read any research about mid caps lately? No!? Many market professionals claim that mid caps are under-researched and therefore assume they are under-represented in portfolios. Moreover, the same professionals claim that this seemingly un-loved segment of the market capitalisation spectrum offers superior returns. However, most of the existing body of research on the matter is US-focused and therefore is not necessarily applicable for every geography or investor.
In this article, we try to dig up the truth about mid caps and give some insight into whether these statements are true for the US only (if at all) or if they apply globally. Moreover, we will look at different market environments and try to identify conditions in which mid caps might perform particularly well.
Mid Caps Defined
Before delving into the topic, let’s first define the different size segments. Morningstar defines large-cap stocks as those that account for the top 70% of the capitalisation of the Morningstar domestic stock universe; mid-cap stocks represent the next 20%; and small-cap stocks represent the balance. As of this writing, the cut-off point in the US is $10.9 billion for large caps, $2.4 billion for mid caps and $709 million for small caps.
Most of the research on mid caps is US-focused and uses the Russell 1000 Index or the S&P 500 Index as a proxy for large caps, the Russell Midcap Index as a mid cap benchmark and the Russell 2000 Index as a proxy for small caps. This approach has, however, a few drawbacks. First of all, the Russell Midcap Index is a subset of the Russell 1000 Index, representing about 800 stocks and 31% of the market capitalisation of the Russell 1000 Index; thereby creating considerable overlap. Secondly, only about 530 of the constituent stocks of the Russell Midcap Index actually have a market capitalisation of between $2.4 billion and $10.9 billion. The remaining stocks are either large caps or small caps as per Morningstar’s definitions; hence one should ideally use the Russell Top 200 Index as a proxy for large caps, though many research papers use the S&P 500 Index as their large cap benchmark.
Using different index providers for these purposes has some inherent flaws. Only about 50% of the components of the S&P 500 Index have a market capitalisation above $10 billion; hence this index cannot be seen as a pure large cap benchmark by our definition. In addition, the S&P 500 Index has more than 20% overlap with the Russell Midcap Index.
On this side of the Atlantic, the current cut-off points using Morningstar’s definitions are $8.1 billion for large caps, $1.5 billion for small caps and $329 million for small caps. Despite some deviation from this definition, the STOXX Europe Large 200 Index, the STOXX Europe Mid 200 Index, and the STOXX Europe Small 200 Index are the best proxies for the different market capitalisation segments in Europe.
Do Mid Caps Offer Attractive Risk/Return Characteristics?
Having highlighted some issues investors should keep in mind when evaluating research of this kind, we will now take a look at whether mid caps actually offer superior risk/return characteristics to either large or small cap equities. Despite the aforementioned drawbacks, we will use the Russell Indices for the US market and the STOXX Europe Size Indices for the European market as they offer the best proxy for their respective markets.
As mentioned previously, the problem with most of the existing research on mid caps is that it is very often US-focused. As we will see, the US findings are not universally relevant, as the European example shows.
As we can see in the table, US mid caps have indeed offered better absolute and risk adjusted returns compared to small and large caps over the past two decades. In Europe, large caps would have been the preferred investment choice based on a risk adjusted return over the same span.
When looking at the performance of the different size indices, it seems that the US market differs from other regions. In the article: Small Caps: Ripe for Further Outperformance?, we showed that US small caps outperformed large caps whereas European small caps underperformed their respective large cap benchmark. We showed that this tends to occur mainly during periods of negative real interest rates of which there were more in the US than in Europe during the period under consideration. Examining the performance of mid caps through the same lens yields similar results; meaning that mid caps have also tended to outperform large caps in negative real interest rate environments. This could be one potential reason for mid cap outperformance in the US and underperformance in Europe.
Are Mid Caps Under-Researched And Under-Represented?
Many argue that the mid cap universe is not widely followed by analysts which in part leads to them being under-represented in portfolios.
Some research papers compare the coverage of mid caps with that of large caps. These papers conclude that only a few companies with a market capitalisation of between $2.4 billion and $10.9 billion are followed by an analyst; hence they argue mid caps are under-researched. Goldman Sachs, for instance, stated in their white paper “The Case for Mid-Cap Investing” that “mid-cap stocks have less sell-side research coverage than large-caps.” In another bit of research “The Case for Mid-Cap Equities” issued by Allianz Global Investors, the authors argue that “Like small cap, the mid-cap size segment is under-research by Wall Street analysts compared to larger companies, providing opportunities to profit from informational inefficiencies.” Both of these research papers focused on the US-market. However, according to Morningstar Data, almost every US domiciled stock is covered by at least one analyst, which would appear to undermine this argument. Again, the case for Europe is different; only about a quarter of stocks are followed by at least one analyst. As the table below indicates, mid caps represent about 15.4% of the overall market capitalisation, whereas this segment represents only 10.4% of the number of stocks covered by at least one analyst.
Based on the data shown here, it seems mid caps are relatively under-researched by sell-side analysts in Europe versus the US. In fact, small caps are even less followed by sell-side analysts in Europe, making this market segment the least followed, and not mid caps as claimed by many professionals.
After showing that mid caps are relatively under-researched by sell-side Analysts in Europe versus the US, we now move on to attempt to gauge buy-side interest in mid caps to get an indication on whether mid caps might be under-represented in portfolios. The idea behind our approach here is that buy-side analysts would only analyse these stocks if there is demand for them in mid cap funds. Hence, if the market of mid cap funds, in terms of total AUM, is relatively small compared to the overall market (as measured by the market capitalisation of the relevant index), it could be argued that mid caps are indeed under-represented in portfolios.
This next table compares the AUM of mutual funds with the market capitalisation of the different size indices in their respective markets.
As we can see, mid caps indeed appear to be under-represented in portfolios on both sides of the Atlantic relative to large caps. It is once again important to note that the different Russell Indices do not follow the definition for mid caps used in this article. However, when taking the Russell 3000 Index and applying our definitions for large, mid, and small caps, we find that the “Funds AUM to Index Market Cap”-ratio for large caps drops to 23.5% and increases for mid caps to 12.7% Nevertheless, this would still indicate that mid caps in the US are under-represented relative to large caps. In fact, this would be inline with the European market.
Conclusion
Some market professionals claim that mid caps are under-researched and often under-represented in portfolios. As a result, mid caps are believed to deliver superior returns.
As we have seen, mid caps have generated superior returns compared to large and small caps in the US during the period we examined, but not in Europe. The reason, however, is not necessarily that mid caps are under-researched or under-represented. If outperformance was in any way related to the fact that mid caps tend to be under-researched, one would assume that European mid caps should have outperformed their large cap benchmark in much the same way that the Russell Midcap Index outperformed the Russell 200 Index during the period in question.
The argument that mid caps are under-represented in portfolios is difficult to prove or dis-prove. However, with the approach used in this article, we get a good sense that mid caps might be indeed under-represented in portfolios relative to large caps. This fact could therefore be a valid reason for mid caps’ outperformance. But then again, if this factor did indeed have any part in explaining differences in returns, we would assume that mid caps in Europe should have outperformed large caps over this period.
One factor potentially contributing to the performance differences is the sector allocations of the different size indices as each sector is driven by different macroeconomic factors.
The Russell Midcap Index is overweight financials with respect to its large cap benchmark. Generally, one would assume that financials underperform during negative real interest rate environments as they derive a large part of their income from their net interest margin (borrowing short and lending long). However, looking at the returns, we can sees that the largest part of mid cap outperformance in the US occurred during the housing bubble as mid cap financial firms were making a fortune from mortgage lending.
On this side of the Atlantic, we discovered only a slight underperformance of mid caps relative to large caps. The graph below* demonstrates that the STOXX Europe Mid 200 Index has outperformed the STOXX Europe Large 200 Index during periods of increasing inflation. These periods are often coupled with raising commodity prices. During the 1990s many commodity prices collapsed; especially oil, which dropped from a high of $41.20 per barrel at the end of 1990 to a low of $10.44 in early 1999. Gold suffered a similar fate, losing about 50% of its value over the same time span. Depending on the sector, some companies will be more affected (positively or negatively) by low commodity prices than others. Utilities, for instance, derive the largest part of their income by providing gas and electricity to companies and households. Prices for end-consumers are far more regulated and therefore less volatile then commodity prices. Therefore, if commodity prices drop, profit margins for the utility sector tend to raise. Basic resources, in contrast, derive their income from digging up commodities and/or processing and selling them. Therefore, if commodity prices drop, profit margins decrease. Although costs of raw material for commodity processing companies drop, the price of finished products also drops whereas operational costs--like labour or machinery--will remain high as they are far less volatile.
Now, looking at the sector allocation of our STOXX Europe Mid 200 Index, we see that utilities were underweight in 1999 relative to their large cap benchmark; representing only 6.2% of the mid cap benchmark versus 23.5% for the large cap index. Moreover, basic resources were overweight (8.2% vs 2.8%) in the mid cap index at the same time. Therefore, we can argue that the underperformance during the mid- to late 1990s is likely a result of unfavourable developments in commodity prices over this span. In return, as commodity prices picked up after having bottomed out in 1999, the mid cap index started to outperform its large cap benchmark.
Therefore, we can argue that the difference between the performance of mid cap equities on either side of the Atlantic can be attributed mainly to differences in sector exposure relative to their large cap benchmarks.
*The light brown line represents the cumulative performance of mid caps versus large caps in Europe. As the underperformance is very small, the graph shows magnitude performances, e.g. 1% underperformance represents 10 points on the left axis.
