Active Management: Diluting by concentrating
Concentrated portfolios can look like a proxy for high-conviction and high-alpha portfolios. Martin Steward asks if the two things necessarily follow one another
A portfolio that holds only 30 stocks benchmarked against the Russell 1000 index is almost certain to exhibit higher active share and higher tracking error than a portfolio holding 250 stocks. Investors looking for active managers who earn their fees with genuine active risk might, therefore, use a portfolio-concentration heuristic to select them.
“Concentration is one of the few ways that you can generate meaningful alpha,” as Liad Meidar, managing partner at the consultant Gatemore Capital Management, puts it.
Of course, there are exceptions to this rule. Is a portfolio manager who selects 30 stocks at random and weights them equally earning her management fee? Presumably not. But even with non-random stockpicking, we need to ask questions. If managers of 30-stock portfolios all chose the same 30 stocks this wouldn’t justify a fee, either – we could simply buy the same 30 stocks alongside them. And what if they didn’t hold the same 30 stocks, but the stocks they did hold exhibited a very considerable overlap in systematic risk exposure? While we would have to choose from among the dozens of stocks in these portfolios, if we did so in a way that systematically overlapped with their style bias, we could replicate a great deal of their performance.
How probable is it that concentrated portfolios do indeed exhibit big overlaps in stocks, a common risk factor or a style bias?
IPE asked Morningstar to gather all of the UK-registered mutual funds on its database that hold 30 stocks or fewer and list the stocks they held in common. There were 58 products, and the commonly-held stocks are shown in table 1. GlaxoSmithKline is held by 21 of the funds. Ten stocks are held by 11 funds or more.
A quick glance suggests that this has a lot to do with size. But is there also a hint of ‘quality’, too – multinational growth generated from strong brands, resulting in high levels of free cash flow and low debt? Set aside some names like Tesco, BP and Shell and the idea doesn’t seem so far-fetched. One would not be alarmed to find GSK, Unilever, Diageo, Microsoft, and the other big tobacco and healthcare stocks in self-styled ‘quality’ portfolios.
“More than half of these stocks are, I would suggest, in the quality or defensive style buckets,” says Rich Dell, global head of the equity boutique at Mercer.
Even some of the fund names tell a story: ‘Worldwide Leaders’, ‘Global Franchise’, ‘Global Brands’, ‘Blue Chip Income & Growth’, ‘Outstanding British Companies’, ‘Buffettology’.
Stuart Gray, senior consultant in equities at Towers Watson, conducted a similar sweep using eVestment’s database of global equity portfolios, and found 37 holding 30 stocks or fewer. Here there was less overlap but the 10 stocks held in more than six of the portfolios make for similar reading – Google, Oracle, Nestlé, Microsoft, Diageo, Johnson & Johnson, Novartis, Time Warner, GlaxoSmithKline.
Risk analytics specialist Axioma took the Morningstar list and compared its characteristics to the broader universe of global stocks (table 2).
“It looks like this list of stocks is more sensitive to changes in exchange rates than the average stock in our database, is slightly more levered, less liquid, has worse momentum, is larger, more expensive and is less volatile,” observes Melissa Brown, Axioma’s senior director in applied research. “The average beta is also lower than that of the broader universe.”
Large size and marked FX sensitivity would be characteristic of multinational brand leaders. Low volatility, low beta and even low liquidity speak of stocks that investors buy and hold with little sensitivity to price – like high-quality franchises. Even higher-than-average debt levels might be explained by the fact that stable cash flows allow these firms to take on leverage, and that debt levels at lower-quality companies have been falling since the financial crisis. Higher-quality firms may also hold cash and debt in different jurisdictions to minimise tax exposure. Moreover, this indebtedness profile is much less pronounced if we isolate the top 10 most commonly-held stocks. Similarly, high valuations might have been a little surprising five years ago, but no longer: Nestlé’s free cash flow yield was nearly 7% in 2006 and has since gone as low as 4% as these quality names have attracted big investor flows.
Gray at Towers Watson suggests that the overlap of concentration and quality could simply be coincidence, as both have been in vogue recently. “These holdings may completely reverse as the trend away from quality, which we may already be seeing, takes hold,” he suggests.
But one can come up with sound theories that explain the naturalness of the overlap.
First, the more concentrated a portfolio is, the higher its level of stock-specific risk. To keep that risk within manageable bounds, it may be tempting to tilt towards quality franchises – and this apparent tilt towards quality, defensive stocks is quite remarkable, given the general bias among active managers to choose stocks with higher volatility and higher beta. According to this theory, the concentration causes the tilt to quality.
Second, the number of genuine global franchises is limited. Demand a sensible valuation for them and the universe becomes still more limited. “Most high-quality managers will tell us that their global universe is a couple of hundred companies at the most,” says Gray.
At FOUR Capital, one of the managers in the Morningstar list, this seems to be an important impetus behind the concentration of its ‘Stable’ product.
“This portfolio would fit with the quality characterisation, because it concentrates on the lower-volatility part of our universe – and only about 20% of the market will fit,” explains senior investment manager Colin McQueen. “It’s not a choice to concentrate, it naturally falls out from the way the strategy works.”
As McQueen suggests, according to this theory, the tilt to quality causes the concentration.
A similar phenomenon can be seen at PGGM, where head of long-term equity strategy Felix Lanters manages a 16-stock portfolio that he would like to get up to 30-40 holdings. The fact that this is taking time is an indication of how selective his stock-selection criteria are; his mandate is to keep up with the FTSE All World index over the longer term while investing in companies that have a meaningful positive societal impact as well as generating sustainable financial returns.
“We are looking for many of the things associated with quality companies,” says Lanters. “Low leverage, high free cash flow – preferably cash returns in excess of the cost of capital, which is one of the reasons any company manages to survive over the long term. Essentially we are looking at sustainability in both societal and financial terms. We want to establish significant relationships with our companies and it’s easier to do that with companies that have reasonable prospects of staying around. Both the selection and the maintenance processes are quite labour-intensive, so they are both reasons to run a concentrated portfolio.”
Having set out the case for a significant overlap between quality and concentration – and suggesting that such a marked systematic factor bias could be replicated cheaply – let us now run through the caveats.
First, we should not discount the possibility that many concentrated quality portfolio managers are doing something unique.
Fundsmith is another manager on the Morningstar list. Founder and CEO Terry Smith says that the sectors that do or do not qualify for his fund have remained “pretty stable” – “consumer staples, medical equipment, elevator and escalator manufacturers, franchises” are in, “banks, insurers, technology, chemicals, airlines, mining, steel, oil and gas” are out. These sector tilts are common to self-styled quality portfolios – and even within technology and chemicals the fund has quality holdings like ADP and Sigma Aldrich.
Smith insists that he has no interest in benchmarks, peer-group or factor analyses. But, pushed on the question of style bias, he reminds us that his team of analysts earns its money by selecting portfolio holdings from its universe of eligible stocks.
“We do track ourselves against our investable universe of 64 companies, so I am able to say that an index that replicates our investable universe will underperform our portfolio – during 2013 it would have been 10 percentage points of underperformance,” he says. “Either that must have something to do with some skill we have, or we have simply been lucky.”
Second, it is possible to run a successful quality portfolio that is not concentrated, and to run a successful non-quality portfolio that is. This point is made by factor-investing specialists AQR Capital Management, which has built portfolios focused on profitability, growth and defensiveness and found them to highly-diversified. Managers who reduce diversification must feel they have skill, he reasons.
“The only way to make sense of concentration as a characteristic of quality portfolios would be if we found that a smaller sub-set of quality turned out to be less scary than the full universe of quality stocks, because if it were a safer place to be concentrated and you held skill to be a constant, quality would naturally be a place where active managers would feel comfortable with concentration,” explains Cliff Asness of AQR. “However, this is not a premise that we have tested.”
Asness’s colleague, Antti Ilmanen, adds that concentrated long/short equity portfolios tend to be long momentum and short defensiveness. And let’s take another look at that Morningstar list: alongside ‘Global Franchise’ and the like, there is the odd ‘Value Equity’ fund, and even ‘Aggressive’ and ‘High Octane’ products.
Think about the value style. While the lighter version might benefit from diversification because these are risky stocks that, even when successful, may have to wait years for their re-pricing catalysts to come through, deep value benefits from concentrated portfolios because it is even more risky. Practitioners like to know their companies inside-out to avoid value traps, which narrows the universe of possibilities, and they want that hard work to be rewarded by meaningful risk contribution.
At a glance
• Concentrated portfolios are often seen as proxies for highly active, idiosyncratic portfolios.
• But there is some evidence that they hold a lot of common stocks and share common factor exposures.
• Theory may also support the idea that concentration works well for ‘quality’ strategies.
• The relationship is not cut-and-dried, but fund selectors should be aware of the risks.
In Europe, Mandarine Gestion’s president Marc Renaud comes close to this style in his 30-40 stock, contrarian-value Valeur fund.
“I like to run a concentrated portfolio because, otherwise, my successful positions will not add much to the return of the portfolio,” he says. “When the market isn’t really cheap it is not easy to find many names with 40-50% upside, so at these points the portfolio would have 40-45 names. But through a period like 2011 the portfolio certainly becomes more aggressive, which often does lead to more concentration.”
Perhaps the last word on this should go to Tom Howard, CEO and director of research at AthenaInvest and an academic expert on active manager performance. He believes that high-conviction is vital to success, but that it has less to do with the number of portfolio holdings and more to do with how closely they agree with the manager’s stated style.
To explore this, he has tracked almost 3,000 US-domiciled funds for nearly 20 years and grouped them into strategy ‘pools’ such as ‘Valuation’, ‘Economic Conditions’ and ‘Future Growth’. Pools we might expect to contain quality stocks include ‘Competitive Position’, ‘Opportunity’, ‘Profitability’ and ‘Social Considerations’. Does Howard notice any style-specific patterns in levels of concentration?
“By and large, if you disregard the diversified portfolios in the ‘Quantitative’ group, there is no relationship between the strategy pool and the number of stocks in constituents’ portfolios,” he reports.
On balance, the findings from such an extensive sample should allay concerns raised by the more modest and anecdotal study this article opened with. But even if quality and concentration are not such an obvious overlap as they may seem at first, investors should remember that concentration is not a proxy for ‘activeness’, and that while it is possible to achieve good diversification between genuinely active, highly-concentrated portfolios, this should not simply be assumed. Style biases should be rigorously assessed.