This is the second post in a mini-series on how factor returns from academia can be different to those from smart beta investing. For the previous post on how long only smart beta exposure compares to long/short factor exposure, click here.
A sad truth for investors is that size really does matter. But contrary to most other areas of life, when it comes to factor investing, the smaller the better.
In reality, most of us will construct our portfolios using large funds which already have millions, if not billions, of assets invested in them before we invest. These large funds are excellent for keeping costs low, because they can spread their fixed costs across a large pool of AUM, but their size can also be a weakness.
There are many investment strategies whose performances diminish with every marginal pound allocated to them. For example, if you run a fund and your investment strategy revolves around investing in illiquid micro-cap stocks, then the strategy may work well with a small amount of capital, but as your strategy does well and attracts new capital, the total capital you must invest increases and you start encountering a couple of problems.
Firstly, your liquidity needs increase. With more money to invest, you need to buy more shares of the companies, and so need more people who are able to sell the company’s shares to you. When selling a company’s shares, you also need people who are able to buy them from you. With small-cap stocks, their markets are more illiquid – they have fewer buyers and fewer sellers – which makes investing increasing amounts of capital difficult. This problem of fewer people to transact with means that you’ll have to accept a worse price for buying and selling to entice others to trade with you, as you have no choice but to buy/sell in high volume because you have more capital to invest. If your capital continues to increase, you may become such a large part of the market in particular stocks that it becomes impossible to buy/sell with anyone else in the volumes that you require – trading in those stocks then becomes impossible.
The second problem you encounter is that with larger amounts of capital, you start having to buy larger stakes of the companies you’re investing in. If someone with a £100bn portfolio decided to start investing in micro-cap UK stocks, then they would find they’d be buying majority stakes in the companies they’re investing in – if not the entire companies themselves. The problem with owning more of a company is that the investor is expected to take more of an active role in the company’s management. “With great ownership levels comes great responsibility”, as Spiderman famously said. Not all investors want this – they’re investing in the public markets, after all, and are not looking to be activist investors or investing in private equity.
What this means for us as investors is that the large funds we’re investing in will almost always be investing in large-cap companies. You can’t have a multi-billion pound fund investing in small-cap stocks because it just doesn’t work. And this becomes a problem when we’re trying to invest in factors.
Alice’s Adventures in Factorland
When we look at factor returns from academia, the returns are not from only large-cap stocks, but are from stocks across the whole cap spectrum. In a recent paper entitled “Alice’s Adventures in Factorland: Three Blunders That Plague Factor Investing” from Research Affiliates, the authors note that when Fama and French constructed their original factors, they equally blend the long–short factor portfolio constructed in the large-cap space and in the small-cap space. This is problematic, because the vast majority of academic factor research is drawn from the original Fama-French definitions, and many papers even use the same factor data from Ken French’s data library.
The authors find that, “Although their stated motivation in factor construction is to separate the factor effects from the size effect, the important consequence of this method is that it significantly boosts the theoretical factor returns. Because small stocks are more risky, less liquid, and potentially more prone to mispricing due to investor inattention, the factor premia are much stronger in the small-cap space. Many factor investing products completely miss this point of the Fama–French construction approach.” [emphasis mine]
This point that factors are stronger in the small-cap space is reinforced by research from Morningstar, who take the value, momentum, low-volatility, and profitability factors and see how each of them performs in portfolios of differing market-caps.
The tables below show the returns on 25 portfolios of U.S. stocks formed on the basis of stocks’ size and its factor score. For example, the large cap portfolios with the highest quintile value score returned 11.57% annualised.
The column labeled “5–1” shows the return spread between the portfolios of the cheapest and most-expensive stocks across five different size strata. So, for example, the cheapest fifth of U.S. large-cap stocks outperformed the most-expensive fifth by 1.93% annually. You can see that the return gap between deep-value and high-growth stocks increases dramatically as we move down the market-cap ladder – the value factor performed better in small-cap stocks.
And the same can be said for low volatility, momentum, and (to a lesser extent) protfitability:
Whilst the effect is much less pronounced for the profitability factors, the other factors show clear outperformance for factors in the smaller cap arena. The article goes into explanations for why the factors may be stronger in small caps (including less analyst coverage, increased exposure to other factors which are strongest in the small-cap space, and the fact that news is incorporated into stock prices more slowly). The author re-ran these tests for global ex-US portfolios, to test whether it was a phenomenon confined to the US, but found that the results were consistent in international markets.
Research from Factor Research reaches a similar conclusion, looking at data for the 2000-2018 period. They noted that allocating to the value factor via smart beta ETFs generated negative excess returns over the period, but a long-short value portfolio as featured in academic research would have generated positive returns:
Source: Factor Research
The majority of smart beta ETFs are invested in market-cap weighted large-cap stocks, where factor effects are weaker. If investors wish to capture more factor exposure by allocating more of their portfolios to small-cap factor funds, where factor effects seem to be stronger, then they must realise that this comes with a few drawbacks:
- Smaller cap funds have higher transaction costs than their large cap counterparts, so have a higher performance hurdle to clear
- Smaller cap funds are generally riskier than their large cap counterparts, which makes the returns on offer more difficult to capture (as the funds are more difficult to stay invested in)
- Since shorting small cap stocks is expensive and sometimes impossible, the exposure is even more likely to be long-only (the drawbacks of which we looked at in the previous post)
- Depending on how far down the cap spectrum the fund is investing, there may be liquidity issues in the underlying stocks
- The investors are taking an active bet away from the market portfolio (which might contradict the ethos of an index-tracking purist)
All in all, investors can’t expect to achieve academic levels of factor exposure by investing in the (apparently very popular) large-cap smart beta funds. Whilst it can tempting to believe in the possibility of having the comfort of a large-cap portfolio combined with the performance benefits of factor exposure, it doesn’t appear that we can have our cake and eat it too. Higher factor exposure comes with higher risks, and investors need to understand these risks before chasing factor performance.
The next post in the series looks at how trading costs in academia differ from those in real life, and what this means for smart beta investors.