Alternative Title: How to Avoid Bad Manager TimingLet’s look at the two types of risks in most investments:
Strategy Risk: If you own a black-box ‘go-anywhere’ hedge fund that invests long and short and uses futures and derivatives at any frequencies, you are mostly exposed to the strategy risk. It doesn’t really matter what is happening to the underlying assets - the strategy either works or it doesn’t. The strategy can be anything: long-short quant, fundamental stocks, futures, or options. In these cases, the long-run performance of the underlying holdings has little impact on the financial outcomes, because the strategy determines most of the risk. In other words, if your manager falls asleep, you cannot hold on to their positions with confidence that over the long term, they should generate a satisfactory return.
Asset Risk: On the other hand, if you buy and hold S&P500 for the long-run, you are mostly exposed to the underlying asset risk. The strategy in this case is very simple, and the only way it breaks is if you don’t stick with it. On the other hand, the asset risk is huge. Will S&P repeat its stellar 20th century performance during the 21st century? Will it also repeat its maximum drawdown of -84%? Is there a small chance that it will go to zero before recovering again? That’s the asset risk.
To set the stage, let’s define risk as an unexpected outcome with unrecoverable consequences. Risk is when investors have to (or decide to) change their approach without a previously specified plan during an inopportune time (more here). Let us also define a Failing Strategy as a strategy that generates random weights that do not have any statistically positive or negative forecasting ability (more here).
To me, an important question in any investment approach is whether the balance between strategy risk and asset risk is right. For example, let’s consider Risk Parity. When properly implemented, its strategy risk is very low. Risk Parity equally allocates the portfolio’s volatility budget to stocks, bonds and commodities, re-balances to a volatility target, and uses leverage to get there. It’s a simple, transparent, rules-based approach. The strategy risk almost does not exist.
On the other hand, Risk Parity has a lot of underlying asset risk. Will global equities under-perform US equities for another decade (triggering the ‘low-return risk’)? Will commodity futures under-perform their observed performance since 1959? Will bonds have a low return following the currently low interest rates? Or, will both stocks and bonds crash at the same time losing their diversification benefit? Risk Parity along with other static asset allocation approaches, like a 60/40, mainly consist of asset risks and not strategy risks.
An asset allocation approach that is rising in popularity is Dynamic Asset Allocation, where the question of a proper balance of strategy and asset risk is paramount. To what extent are you invested in the manager’s dynamic skill vs. the underlying asset classes? For example, if a certain dynamic approach instructs to constantly change the selection of asset classes, go long and short, changes its volatility target, and turns the portfolio over without limits - then investors are mostly exposed to strategy risk. An investor cannot fire the manager (i.e. close the strategy risk), and still hold on to their positions for the long-run with any degree of confidence because a random snapshot of the underlying investments is not a reasonable buy-and-hold strategy. It is inappropriate to call such strategies Asset Allocation, on par with 60/40 or the Endowment approaches, because the underlying assets aren’t driving most of the risks. Many advisers appropriately place such strategies in their Alternatives bucket, which is where most strategy risk driven investments reside.
On the other hand, if the Dynamic Asset Allocation strategy risk is well balanced with the underlying asset risk, then investors can have the best of both worlds. If the strategy stops working – meaning the future weights become ‘random’- then, the investor will still benefit from the exposure to the risk-reward benefits of the underlying asset classes. When the risks are balanced, an investor is betting on two processes to work: the strategy and the assets. If one fails, the investor still has a good chance of winning, if they stay invested and don’t change their approach at an inopportune time. When both asset risk and strategy are balanced, asset class return is what remains if the strategy fails.
While thinking about returns in terms of their alpha and beta does have great benefits for risk management and more efficient fees, it has many downsides as well. It creates a world where some investments are purely asset class driven, while others are purely strategy risk driven. This might sound like a clean solution in theory; however, it encourages investors to continuously switch from one strategy to the other because the risks of any one of the two will never offer lasting satisfaction for investors. Passive 60/40 crashes too often and sometimes has very low expected returns. Also, “pure alpha” hedge funds sometimes crash and sometimes under-perform. In both cases, these outcomes cause investors to keep switching from one approach to one to another, creating a large bad-timing drag (see here).
When asset-risk driven strategies perform well (i.e. static and passive strategies), all the under-performing strategy-risk exposures are reviewed for termination (Hedge Funds, Factor Investing). When asset risk is not performing well (post-2008), investors move into those same strategies for diversification. Understanding how much of the approach comes from its alpha vs beta is very important in order to set the expectations and fees. However, forcing managers to separate the two might not lead to optimal overall behavior.
The same concept of strategy-asset balance applies to active equity managers. At first glance it might appear that the balance between strategy risk and asset risk is solely driven by the strategy’s tracking error but that is not the case. Imagine one active manager investing in a handful of small and risky stocks all concentrated in some niche industry - the strategy risk dominates. The randomized buy-and-hold of the underlying assets is not likely to generate a reliable return over the long-run because specialized and up-to-date knowledge is required in order to keep investing in these assets. Now, imagine a second manager, with the same tracking error and volatility as the first one, who now selects well-financed, large firms with competitive moats from a diversified set of sectors - now the strategy risk and the underlying asset risks are more balanced. If the strategy fails, the investor can still hold on to the underlying holdings and expect to earn equity-like returns.
Why is the ability to hold on to the underlying assets so important?
Because most investors fire their managers following periods of extreme or extended under-performance. However, remembering that the definition of a ‘failing strategy’ is ‘random portfolio weights’, it becomes clear that an investor should not fire their manager during under-performing times as long as there is a conviction that the underlying asset risks are well represented in the portfolio and that these assets have a positive long-run return premium.
In the case of a well balanced strategy-asset approach, an investor is not forced to sell positions during a bad performance period - even if the strategy stops working - because the investor has enough conviction in the underlying assets to mean revert, get out of the drawdown, and then have a choice to re-balance out of the failing strategy at an opportune time. This extra level of patience also gives extra time for the strategy to start working again.
Warren Buffett’s approach to balancing strategy and asset risks gives him the ability to stay very patient even when his investments are not outperforming the index. Even if he is just earning the market returns, he is still compounding his assets at solid returns. By owning great U.S. companies, even if his weights become random for some time (no alpha risk), or significantly crash (drawdown risk), over the long-run he will still earn the average return of U.S. equities. As a result, he does not have to sell during drawdowns or periods of low alpha, helping him avoid the largest cost of investing with the extra benefit of giving his strategy more time to start working again. Because of the conviction in the underlying asset, Buffett does not sell at the bottom, and so his worst case outcome is S&P500 returns with random noise, which is much better than most investors who on average under-perform the buy and hold allocations by 3% per year! Comparing Buffett’s worst case to an average already makes a strong point, but if you include the worst case outcomes for the average equity investors, the point would be even stronger.
Bridgewater’s relatively recent launch of the ‘optimal portfolio’ that combines their pure alpha and risk-parity beta strategies is another great example of trying to balance strategy and asset risks under one umbrella. Why did Bridgewater create this combined strategy when for decades they advocated the benefits of separating alpha and beta? Most of their investors could easily invest in the two components themselves (in fact many already have). They created the combination not just for the convenience of less paperwork, but because they realized the inherit human biases that cause sub-optimal timing behavior, causing performance volatility to transform into risk via poor timing decisions.
Factor Investing is another interesting case study. For decades, factor investing belonged in the strategy risk group, traditionally called alpha investing. With the democratization of quant, factor investing has migrated from strategy risk to premia risk, which via its association with macro risk factors has morphed into the same status as asset class risks on par with the equity risk premia. This shift provided investors the desired comfort, which was well needed after the 2007-2009 quant blow up. Investors wanted a sense of transparency and comfort that holding on to the quant factors for the long-run is the same as holding on to the S&P500. They wanted quant investing to become closer to asset risk rather than strategy risk. I believe that in their long-short formulations, factor investing is still largely a strategy risk approach. However, in the long-only versions, factor investing can be seen as a combination of strategy and asset risks, which means that investors should not panic and sell their smart beta ETFs when they under-perform as long as they have conviction in the underlying asset returns. Let time and either skill or randomness take care of the rest and get the strategy out of the drawdown first, before making any decisions.
Be it in asset allocation, stock selection or factor investing, approaches that balance strategy and asset risks give the investor the extra confidence to hold on during turbulent times and avoid locking in the losses during drawdowns.