This is the third post in a series discussing some of the problems associated with investing in “smart beta” strategies. For the previous post on the importance of defining a factor, click here.
A further problem with investing in smart beta funds is that factors don’t always stay the same. As time has gone on, some factors been changed, refined, declared dead, revived, declared dead again, and caused everyone a great deal of confusion.
Let’s look at some examples.
A dead factor…
The size factor was one of the original factors in the Fama-French 3 factor model, and has been massively popular since its discovery. Proponents of the size factor claimed that there were additional returns to be had by allocating to small cap stocks, over and above the extra risk taken. The size factor was ingrained deep into conventional wisdom, being taught in classrooms around the world, and was taken as being one of the premier factors ever since its discovery.
However, recent research (here and here) has suggested that the original Fama-French size factor is not as robust as previously thought. Some have argued that there’s actually no size effect whatsoever once you take the additional risk from small caps and the exposure to the value factor into account.
The fact that the very existence one of the oldest and most well-known factors can be brought into question shows how no factor is safe from change. New research is constantly probing and testing the status quo, so staying on top of the academic evidence is increasingly important if investors want to understand and potentially improve factor exposures in our own portfolios.
A dead factor – revived?
Following the death of the small-cap factor, researchers from AQR found that there was still life in the old factor yet. After adjusting the small cap universe by screening out the lowest-quality small-cap stocks, they found that the size effect was once again robust, and provided statistically significant alpha. But only if the ‘junk’ small companies were removed.
In the words of AQR: “We find, however, that these challenges are dismantled when controlling for the quality, or the inverse “junk”, of a firm. A significant size premium emerges, which is stable through time, robust to the specification, more consistent across seasons and markets, not concentrated in microcaps, robust to non-price based measures of size, and not captured by an illiquidity premium.”
It’s these kinds of revisions that highlight the difficulties of factor investing – there’s no one “perfect” way of measuring a factor. In 50 years’ time, AQR’s junk-adjusted size factor might be found to be a product of data-mining, and a new superior way of measuring the size factor might be found. There are no iron laws of investing, and past truths that were once taken as fact are constantly being tested, updated, and revised as new data becomes available.
A refined factor
Leaving the size factor behind, the value factor has also had its fair share of changes. The traditional way that academics capture the value factor has been to use the B/P ratio using lagged data for both book value and share price – the data lag was anywhere between 6 and 18 months. This had always been the conventional way of measuring value, and had been used this way in academia since the discovery of the value factor in the 1990s. Again, it was very much the status quo.
However, recent research by AQR has shown that using up-to-date share prices in the B/P calculation, rather than lagged data, actually improves the performance of value strategies. This seemingly minor tweak adds over 3% of alpha every year, according to the research findings. They dubbed the adjusted factor “HML DEVIL” to reflect the fact the devil is in the details when it comes to factor exposure. Importantly, this extra alpha only applies to strategies using both value and momentum strategies together – it actually makes a standalone value strategy worse than under the traditional definition of HML.
The value factor formed part of the original Fama-French 3 factor model, and to have such a seminal factor questioned again shows that the only constant in factor investing is change.
A factor imposter?
The low-volatility factor is one of the newer factors to make into the mainstream. It was never a member of any of the Fama-French factor models, but has been shown to be a popular factor not just in academia, but also for practitioners.
However, there’s been some research suggesting that the low volatility factor isn’t a factor in its own right, but is in fact just an amalgamation of the value, investment, and profitability factors. The authors suggest that there’d be no value in allocating to the low vol factor if a portfolio already had exposure to these other factors.
Even the newer factors are being questioned, and the fact that the low-volatility factor may just be an amalgamation of other factors demonstrates the importance of knowing what’s in your current portfolio before adding any additional factor exposure.
P/B – a dying factor?
Whether you look at P/B using a lagged book value/price or not, many people are starting to call into question whether the P/B ratio is still a useful measure of value. It’s come under significant criticism in recent years for its inability to accurately capture the true value of a company.
The crux of the argument for P/B having lost its relevance has 3 parts:
1) Book value excludes intangible assets, which are an increasingly important part of modern companies’ value
2) Intangibles are often understated on a company’s balance sheet, as they exclude R&D and advertising expenses
3) Share repurchases made above book value, but below a company’s estimate of intrinsic value, will reduce book value
Let’s take those points one by one.
1) Book value excludes intangible assets
Book value is calculated as net assets, less intangible assets, less liabilities. The fact that book value excludes intangible assets is a problem for measuring how valuable a company is, as an increasing amount of company value is appearing as an intangible on the balance sheet. Intangible assets include brands, patents, copyrights, franchises, goodwill, trademarks, etc. The fact that these intangibles now account for 84% of the market value of the S&P means that excluding intangibles from the value of a company doesn’t seem to make much sense:
Source: The Economist
All these intangibles do, rather obviously, increase the value of a company. All pharmaceutical companies, for example, rely heavily on their patented drugs, and almost all large food chains rely heavily on franchises. Apple relies heavily on its brand to sell premium-priced hardware, and Disney relies heavily on its trademarks to monetise its film and TV franchises.
Yet none of this is reflected in the book value.
2) Intangibles are often understated
Even if investors incorporated the value of intangible assets into their calculation of book value, accounting rules mean that the value of intangibles is often understated. Researchers at OSAM produced an excellent piece on the erosion of P/B value, and found that many expenses which create longstanding value for a company are not being capitalised on the balance sheet (as most expenses which provide their benefit over multiple years should be), but instead are taken directly to the P&L in the year they’re incurred. As book value is calculated using balance sheet figures, the fact that these expenses are taken to the P&L and don’t hit the balance sheet means that they’re not included in the measure of book value.
In the words of OSAM, “The two largest examples are R&D expenses and advertising expenses. The value that companies create through these investments in brands, patents, intellectual capital, technologies and processes is often held at $0 on the balance sheet. As an example, the $90 billion that Coca-Cola has spent on advertising in its history has no value that is shown on balance sheet. Similarly, Boeing has spent over $100 billion designing aircrafts and all that investment does not create an asset.”
This is a growing problem, as companies are spending more and more on R&D – meaning their book value is more and more understated:
Source: OSAM
Similarly to R&D, advertising expenses help to foster a company’s brand and help it grow. Brands are incredibly important to companies, but are famously difficult to value. Under accounting rules, things like brands don’t appear on the balance sheet until they’re acquired – where the value of a company’s brand appears under the goodwill line the financial statements. This is the difference between the net assets of the company, and the amount the acquirer paid for it.
So if a company cultivates its own brand and isn’t acquired, the value of the brand will never appear on its accounts. According to company balance sheets, the Netflix brand is worth $0 (vs a $155 billion market cap), Tesla’s brand is worth $69 million (vs a $49 billion market cap), and the Nike brand is worth $154 million (vs a $134 billion market cap).
These are obviously hugely understated.
3) Share buybacks can reduce book value
Share buybacks are becoming more and more popular with companies. Despite being almost identical in form to dividends as a component of total shareholder return, investors tend to view buybacks as an added bonus, whereas dividend growth is expected. Companies have therefore started to prefer buying back shares over distributing dividends, as it gives them more flexibility to reduce them in the following years.
However, buybacks at a market price above book value, but below the company’s estimate of its own intrinsic value reduce book value (for a mathematical example, see here). The more popular buybacks become, the less relevant book value becomes.
It’s for all 3 of these reasons that hardcore value investor and long-time proponent of book value Warren Buffett announced in his 2019 annual letter to shareholders that he’d be abandoning the use of book value altogether. The FT commented that, “For Warren Buffett to abandon value investing feels like the Pope renouncing his faith: an act so grand it is large enough to cause a crisis of faith in even the most committed of devotees.”
For those of us whose interest in the P/B ratio lies in how it’s used to create value factor funds and multifactor funds, P/B’s weaknesses cause us a major problem, given that 99.8% of value ETFs (by assets) use book value:
Source: Eric Balchunas
If you believe that book value is a flawed measure, then finding value exposure in a smart beta fund that excludes book value is more difficult than you’d think.
Conclusion
Overall, the fact that factors can (and have) changed over the years makes selecting smart beta exposure for portfolios more difficult. New evidence is constantly emerging that puts conventional wisdom to the test, and revisions to previously canonical factors shows how nuanced factor investing can be.
For me, the changing nature of factors highlights the need for investors to stay on top of the academic evidence, and ensure their portfolios are in line with the most recent and robust research. For those who choose to outsource their investment management, ensuring their manager has an evidence-based philosophy and a sound research process will likely be key in making sure their factor exposure remains relevant.
The next post in the series focuses on how factors can decay after they become widely known, and how that makes allocating to smart beta strategies more difficult.