Contact Us
Hedgewise advisors are available to every potential client because you deserve to know and trust who is managing your money.
E-mail
Questions / Comments
SUBMIT
X CLOSE
Investments That Outperform
Thanks! Only $1,000,000 remaining.
You can help accelerate launch by referring a few friends.
X CLOSE
PERFORMANCE ABOUT BLOG LIVE
CLIENT LOGIN
PERFORMANCE
ABOUT
BLOG
DEMO
CONTACT US
TOPICS
About Us
|
Account
|
FAQ
|
Investment Strategy
|
Market Commentary
|
The Anatomy of Momentum: Why the Strategy Works, and How to Play the Odds
Posted in Investment Strategy on 2019-03-28

Summary

  • Since the summer of last year, the Hedgewise Momentum framework has experienced a period of significant drawdown and underperformance compared to the S&P 500
  • Though this raises many reasonable questions about its efficacy, such periods are quite consistent with the strategy underpinnings and almost guaranteed to occur an average of twice per decade
  • To understand this, it is useful to deconstruct "Momentum", which is really just an advanced stop-loss strategy. Such a framework is subject to periods of underperformance despite significant evidence that it provides superior long-term risk-adjusted returns
  • Looking forward, the outlook remains excellent, but not because of any "timing" ability or keen insight about the market, but rather because the framework persistently shifts the odds in your favor

Introduction: The Trouble with Intuition

One of the persistent challenges of running a quantitative investment framework is dealing with periods of underperformance. Since the premise of any strategy is to drive superior long-term gains with less overall risk, it may feel like underperformance indicates a failure and suggests that perhaps the strategy has become less effective.

The problem with this intuition is that quantitative frameworks largely function by shifting your odds to a different set of risks than traditional passive benchmarks like the S&P 500. For example, Risk Parity is more vulnerable to issues with cross-asset performance and correlation, and "factor" investing is more vulnerable to various macroeconomic conditions and market microstructure. Even if, over long periods of time, such risks are less persistent and damaging compared to the risks facing the S&P 500, they are still bound to manifest occasionally, and this will naturally drive periods of underperformance.

As a simple example, imagine that the S&P 500 is a bit like a coin with a 60% chance of landing on heads, and a framework is like a completely separate coin with a 65% chance of landing on heads. Given 100 flips, the advantages of the framework will seem objectively obvious. But on any given flip, there's a pretty high chance that the S&P 500 lands heads and the framework lands tails. Despite the odds that the framework will consistently perform better, there will be frequent short-term exceptions.

As with any probability, the only way to converge towards the 'true' odds is to ignore each individual datapoint and keep playing for a relatively long time. Yet, as investors in the midst of losing money, it is natural to fear that the odds have changed, and to consider whether simpler passive strategies would be a better choice. By clearly deconstructing the elements of the Momentum strategy, this article will highlight exactly what has driven its historically superior returns and why this should continue to drive confidence moving forward.

Back to Basics: What is Momentum?

Broadly, the term "Momentum" is used to refer to investment strategies that assume certain trends tend to continue over time. For example, some trend-following frameworks assume that the best-performing stocks over the past year will continue to outperform, or if stocks have done badly recently, they will go on to do even worse. There is a great ongoing debate about how and why a factor like momentum would have such predictive power.

I believe the idea that trend-following strategies have significant predictive ability is simply not credible. That said, I also believe that using momentum as a framework can significantly improve your risk-adjusted returns. How can this be?

This is far easier to explain if you consider that a simple stop-loss mechanism - a rule to sell something after incurring a certain amount of loss - is functionally the same as momentum. If things keep going up, you hold on. If they go down a certain amount, you sell. It would be absurd to argue that such a trigger had any predictive power. Rather, this is more clearly a simple form of risk management, and the relative success of such a framework will be easy to predict. If an asset keeps going down, you avoid losses. If it doesn't, you incur some cost of selling and re-entering (called a "whipsaw" effect). The main benefit of this framework would be to minimize the size of your worst losses, presuming that the worst-case impact of whipsaws should be less severe than the worst-case impact of market crashes.

In terms of efficient market theory, this makes sense. Investors are always considering risk and reward, with the expectation of paying a discount for an underlying asset in the present in order to achieve a positive return in the future. Efficient markets should usually go up, and they shouldn't remain at the same level for very long. Thus, whipsaws shouldn't happen very often. Given that, it's believable that the drawdowns incurred during market crashes may be steeper than those incurred during whipsaws.

The entire validity of 'momentum' becomes a simple trade-off of probabilities. You aren't expecting to predict anything with accuracy, and you aren't trying to figure out whether a whipsaw or a crash is going to occur. You are simply shifting around the risks of your portfolio.

Simple enough: let's make this our hypothesis, and set-up some experiments to test it.

Testing the Hypothesis: First Experiments and Mechanics

To test our hypothesis, we'll first set-up three dead simple strategies that follow one rule: if the S&P 500 hits a certain level of drawdown, sell. Otherwise, invest. Upfront, realize that this first iteration cannot possibly beat the return of the S&P 500 unless you stopped the simulation right in the middle of a big crash. But it's still a useful first experiment to understand some of the mechanics.

We'll set-up three loss triggers to explore: -10%, -15%, and -20%. The goal is to test whether the hypothesis holds up with all of these levels, since they are all consistent with the broad idea.

Performance of Stop-Loss Strategies vs. S&P 500, 1950 to Today

Annual ReturnMax Loss
S&P 50011.15%-54.98%
-10% Stop Loss8.81%-30.77%
-15% Stop Loss9.03%-36.55%
-20% Stop Loss9.41%-40.33%
Model uses index prices and includes all dividends, assumed reinvested. Assumes that portfolios realize end-of-day prices on the day that any drawdown level is exceeded or reversed. Sources: Bloomberg, Federal Reserve Economic Data, Hedgewise

Note that each of the stop-loss strategies has a substantially lower annual return, but also a lower maximum loss. This confirms the hypothesis that the worst-case 'whipsaw damage' tends to be lower than worst-case 'stock crash damage', but you may be still be surprised that a 10% stop-loss can experience a 30% drawdown. This is mechanically important to understand as it demonstrates how whipsaws function. Here's a look at one rough stretch for the 10% strategy.

Understanding the Whipsaw Effect: December 1973

Source: Hedgewise Analysis

Here, the stop-loss trigger is frequently hit, and every time the portfolio liquidates. Then, the day after, stocks often recover and the portfolio buys again, but misses that days' gain. This is the whipsaw effect, which can thus accrue significant losses independent of the performance of the S&P 500. An important takeaway to highlight is that any stop-loss strategy will by definition do worse than stocks during periods where the market whipsaws, so you are likely to experience drawdowns when stocks do not, but you avoid losses when stocks crash.

While this seems scary and random, it is! But remember that's on purpose - you are actively choosing 'whipsaw crashes' over 'stock market crashes' and hypothesizing that it's still preferable.

Given the initial numbers, this doesn't seem like a clear conclusion. To enable a more equal comparison, we need to better take advantage of the stop-loss structure.

Engineering Fairer Odds

One of the substantial hidden handicaps of simple stop-loss strategies is that they will spend very long periods of time invested in nothing at all. For example, if stocks lose 40%, and take a few years to get back above a 10% net drawdown level, the strategy is 100% cash for those years. The period from 2000 to 2007 exhibits what this looks like.

Daily Stop-Loss Performance, 2000 to 2007

Source: Hedgewise Analysis

Spending this amount of time in cash is a bit of an unfair disadvantage. To correct it, we can simply choose a relatively low risk asset that is expected to accrue positive interest over these stretches. Bonds are the obvious choice, but remember that this is in no way timing bonds as the 'smart investment' when the trigger is hit. It's simply a way to deploy a positive yielding alternative when the portfolio is largely in cash. While sometimes bonds might do particularly well if markets go on to crash, they may also do poorly if markets stabilize. The stop-loss framework only cares that the net return is expected to be positive on average, especially when it is out of stocks for a significant duration of time.

If we use 10yr Treasury Bonds as a proxy, here is how it changes the performance of the 10% stop-loss strategy in the same timeframe.

Daily 10% Stop-Loss Performance Comparison, 2000 to 2007

Source: Hedgewise Analysis, Federal Reserve Economic Analysis. Includes all dividends and coupons assumed re-invested.

Observe that bonds significantly underperformed stocks for years beginning in 2003, but it isn't our goal to always outperform the S&P: we simply wanted to do better than cash for a stretch to avoid the stop-loss 'handicap'. It's also worth highlighting how this adjustment reduces the chances of hitting a particularly unlucky sequential drawdown, such as the cumulative impact of the whipsaws in 2001 and then subsequently in 2006.

Here's how the numbers change if we add this bond rule to each of the experimental portfolios and run it again since 1950.

Performance of Stop-Loss Strategies, Add Bonds, vs. S&P 500, 1950 to Today

Annual ReturnMax Loss
S&P 50011.15%-54.98%
-10% Stop Loss, Add Bonds11.27%-23.85%
-15% Stop Loss, Add Bonds10.72%-26.1%
-20% Stop Loss, Add Bonds10.26%-34.58%
Source: Hedgewise Analysis

This is a pretty dramatic effect! This alone makes the 10% Stop-Loss portfolio look more compelling than the S&P 500. However, we'd prefer our hypothesis to hold up consistently across all the portfolios to demonstrate greater statistical significance, and it's very unlikely that the 10% level has some unique special power. It's also still difficult to properly evaluate the 15% and 20% levels, which exhibit both lower annual returns and lower max losses.

The Introduction of Leverage

As you can see in the tables, a buy and hold approach would have experienced a 55% drawdown. It is interesting to consider what the return would be in the stop-loss portfolios if one were willing to tolerate that level of risk. Introducing leverage provides an easy way to do just that. Since you have a built-in mechanism (the stop-loss) to help avoid large crashes, it seems sensible to also amplify your exposure. The following adds a constant 50% 'loan' to the portfolio with an assumed cost equal to the rate on 1yr Treasury Bills.

Performance of Stop-Loss Strategies, Add Bonds and 50% Leverage, vs. S&P 500, 1950 to Today

Annual ReturnMax Loss
S&P 50011.15%-54.98%
-10% Stop Loss, Add Bonds, 150%14.36%-35.57%
-15% Stop Loss, Add Bonds, 150%13.43%-41.45%
-20% Stop Loss, Add Bonds, 150%12.6%-52.42%
Source: Federal Reserve Economic Data, Hedgewise Analysis. At all times, portfolios are either in 150% S&P 500 or 150% 10yr Treasuries, and a daily leverage cost based on either 1yr Treasury Bills (from July 1959 onward) or 3mth Treasury Bills (prior to July 1959) is subtracted from the return.

Now our hypothesis is beginning to look more convincing: every portfolio outperforms the S&P 500 without increasing your maximum loss, yet all of these portfolios still underperform the S&P 500 for lengthy periods of time. Mechanically, this must happen every time a whipsaw occurs. The existence of these periods does not change the idea that your odds were consistently improved over the long-run.

Demystifying "Momentum": Slightly Smarter Engineering

Hopefully, it is clear that all of our work thus far has little to do with secret signals or mysterious predictions. It's just plain financial engineering and probabilities. Introducing a few more intelligent tweaks will illustrate how these same principles apply to nearly all forms of "Momentum" you may hear about in the marketplace.

Some of the most common momentum "metrics" involve the use of moving averages and trailing returns (for example, sell stocks when they fall below their 200day moving average, buy otherwise). These are little more than advanced forms of 'stop-loss' strategies. For example, here's a comparison of a simple -10% stop-loss sell trigger, a 100day trailing return sell trigger, and a 1yr moving average sell trigger.

Stop-Loss Trigger vs. Trailing Return Trigger vs. Moving Average Trigger

Source: Hedgewise Analysis

Notice how closely the dotted-lines often cluster; for a great majority of the time, there's no functional difference between them. The main advantage of the 'advanced' metrics are they build in one important and convenient heuristic: they allow for a re-entry point to equities after a big crash.

Recall in one of our earlier examples how the stop-loss strategies remain can remain in bonds for years at a time, and thus never take advantage of potentially 'cheaper' equity prices. Metrics like trailing returns and moving averages provide a convenient mechanism to re-enter stocks if prices have fallen significantly. To illustrate, here's how the various triggers moved around during the recession in the mid-1970s.

Stop-Loss Trigger vs. Trailing Return Trigger vs. Moving Average Trigger, 1970s

Source: Hedgewise Analysis

Allowing for a lower re-entry point to equities drives some significant potential extra upside, which can be seen by translating these triggers into the correlating actual performance.

Stop-Loss vs. Trailing Return vs. Moving Average Performance, 1970s

Each strategy is either invested in 100% S&P 500 or cash. The 100day Trailing Return sells on any day when the 100day trailing total return of the S&P 500, including dividends, falls below 0. The 200day MA Return sells on any day when the total price of the S&P 500, adjusted for dividends, falls below its 200day average. Source: Hedgewise Analysis.

Note that the strategies only really separate as the recovery ensues between 1975 and 1976. Conversely, it's not as if the momentum metrics are immune to whipsawing, and they'll have other periods of churn where they look worse, as they did at the end of the 70s.

Stop-Loss vs. Trailing Return vs. Moving Average Performance , 1976 to 1980

See disclosures on prior graph.

This illustrates again that whipsaws will occur somewhat randomly around whichever level you happen to choose. That said, it seems perfectly logical that you might achieve some extra upside by allowing for re-entry points after significant equity losses, and these sorts of metrics are a convenient means of doing so. We can return to our identical framework from above to see whether this proves true, starting with the simple case of no leverage, but using bonds rather than cash whenever the triggers hit.

Stop-Loss vs. Trailing Return vs. Moving Average Performance, Add Bonds, 1950 to Today

Annual ReturnMax Loss
S&P 50011.15%-54.98%
-10% Stop Loss Return, Add Bonds11.27%-23.85%
100day Trailing Return, Add Bonds11.76%-25.47%
200day MA Return, Add Bonds12.53%-20.86%
Strategies using an identical set of rules as described above, but only invested in 100% 10yr Treasury bonds rather than cash. Source: Hedgewise Analysis.

This seems about right. If we then inject 50% leverage, all the better.

Stop-Loss vs. Trailing Return vs. Moving Average Performance, Add Bonds and 50% Leverage Performance, 1950 to Today

Annual ReturnMax Loss
S&P 50011.15%-54.98%
-10% Stop Loss Return, Add Bonds, 150%14.36%-35.57%
100day Trailing Return, Add Bonds, 150%15.15%-37.32%
200day MA Return, Add Bonds, 150%16.32%-29.97%
Strategies using an identical set of rules, but invested only in 150% S&P 500 or 150% 10yr Treasuries. A daily leverage cost based on either 1yr Treasury Bills (from July 1959 onward) or 3mth Treasury Bills (prior to July 1959) is subtracted from the return. Source: Hedgewise Analysis

These numbers are already starting to look pretty fantastic, and in some ways, it's easy to see why moving averages might appear to be 'powerful indicators'. But this analysis clearly dispels the notion of any predictive power, and each individual case of hitting a 'trigger' means nothing at all. This outperformance is driven solely through a mix of financial engineering and long-term probability. Thus, ironically, indicators like moving averages can be plenty useful despite meaning nothing at all!

Pressure Testing: Picking a Level, Dealing with Noise

One of the very tricky elements of this framework is that you have to choose a particular trigger, and that will seem like a big deal. Should it be a moving average? A trailing return? How do you choose the best one? What if you choose wrong?

The problem with this is that we've already established that this is not a market timing strategy and that whipsaws around any particular level are random and will occur. On the one hand, that means it shouldn't really matter which trigger you choose. On the other hand, it also means that there's a really, really high chance that some other trigger will be doing better than yours, but even so, there's no reason to think you should use that one instead.

To understand this better, let's expand the scope of our triggers to include many moving averages and trailing returns. If our theory is correct, all of them should exhibit similar performance profiles, though there will also be substantial noise given the randomness of whipsaws.

Various Daily Momentum Metrics Performance, Add Bonds and 50% Leverage, 1950 to Today

Annual ReturnMax Loss
S&P 50011.15%-54.98%
100day Return, Add Bonds, 150%15.15%-37.32%
200day Return, Add Bonds, 150%14.03%-52.7%
3mth Return, Add Bonds, 150%15.18%-41.49%
6mth Return, Add Bonds, 150%16.46%-46.7%
1yr Return, Add Bonds, 150%14.44%-51.45%
50day MA, Add Bonds, 150%15.63%-35.44%
100day MA, Add Bonds, 150%15.96%-53.52%
200day MA, Add Bonds, 150%16.32%-29.97%
6mth MA, Add Bonds, 150%15.73%-53.95%
1yr MA, Add Bonds, 150%16.28%-40.27%
Each of these strategies is using the exact same rules discussed earlier, with the only modification being what triggers a sale of the S&P 500. Source: Hedgewise Analysis.

In another fun twist, this also holds up in the same way if you make all of these monthly triggers - meaning you only trade on them once a month, rather than daily.

Various Monthly Momentum Metrics Performance, Add Bonds and 50% Leverage Performance, 1950 to Today

Annual ReturnMax Loss
S&P 50011.16%-52.82%
3mth Return, Add Bonds, 150%14.71%-37.29%
6mth Return, Add Bonds, 150%15.42%-34.89%
1yr Return, Add Bonds, 150%13.9%-43.39%
3mth MA, Add Bonds, 150%13.0%-38.37%
6mth MA, Add Bonds, 150%15.43%-38.1%
10mth MA, Add Bonds, 150%15.9%-33.81%
12mth MA, Add Bonds, 150%15.61%-36.97%
These strategies evaluate the given metric at the beginning of each month, rather than daily. All other aspects remain the same. Source: Hedgewise Analysis.

It is natural to gravitate towards the better performing raw numbers, but a different view of the rolling trailing ten year cumulative performance of a few of these strategies helps to show the powerful effect of randomness.

10 Year Trailing Performance

This measures the trailing ten year cumulative total performance of select strategies from the prior tables. Source: Hedgewise Analysis.

Every one of these strategies has some good decades and some bad ones, with no discernable pattern. When you look at the summary numbers, the 200day MA stands out as having one of the top returns, but that's only because of one random spike in the early 90s. Outside of that, there's nothing to suggest that it is somehow superior.

The most productive way to approach this choice is to pick a metric at random with confidence that you have an incredibly high chance of beating the S&P 500 over the long-term. It will be pure luck whether you wind up with 2% or 4% outperformance, and that's perfectly okay.

Sorting Through the Short-Term

This story sounds convincing when you are looking at ten-year time horizons, but the implications for the short-term can be quite challenging. When the strategy sells stocks, it will feel like you are trying to time a downturn (though you aren't). When you get hit by a whipsaw, it will feel like your strategy has failed (though it hasn't). Looking forward, you face an unavoidable element of randomness that will determine how well you perform in the future.

The key to sorting through this is to constantly orient your perspective to this very moment: right now, this year, does the framework give me better odds than the S&P 500?

One nice visualization to answer this is the 1yr trailing returns of one of the stop-loss strategies compared to the S&P 500. The 100day MA metric is selected randomly for comparison, but note that every metric from the prior tables returns similar looking results.

1yr Trailing Performance, 100day MA Strategy vs. S&P 500

The 100day MA strategy depicts the 1 year trailing cumulative performance of a portfolio using the 100day daily moving average as a buy/sell trigger, investing at all times in either 150% S&P 500 or 150% 10yr Treasuries. Leverage costs are estimated as discussed in prior disclosures.

What jumps out immediately is that, on net, your returns are higher, more stable, and less prone to severe drawdowns. But you'll still have bad stretches and they'll often happen independently of equities.

The question is, given we can't know what the next 12 months will bring, which set of possible outcomes would you choose?

The Hedgewise Framework

At various point throughout this article, if you've thought "I wonder if you could tweak these rules in certain ways to make them even better", I have great news for you: the answer is yes, and this is what Hedgewise does! There are a couple of additions that are especially useful, such as layering on diversification across assets, and employing differing levels of exposure based on the environment.

While Hedgewise has abundant theory and research to suggest these are effective modifications, it's crucial to understand that the goal is still never to time the market. Instead, it's useful to return to the probabilities: imagine the simple momentum metrics above have about a 30% chance of hitting a whipsaw in any given year, and Hedgewise manages to reduce that probability to 20%. Even given this dramatic improvement, you still fully expect two whipsaws per decade!

By accepting whipsaws as inevitable in any form of this broad framework, it becomes easier to shift focus towards probabilities and long-term returns (rather than on each individual whipsaw and whether it could have been avoided). Using this lens, let's take a look at the relative effectiveness of Hedgewise methods.

First, let's examine the same prior graph of 10yr trailing performance across various Momentum metrics, but add in the specific Hedgewise "Momentum Max" framework as well. The other metrics are still assuming that you have added bonds and 50% leverage.

10yr Trailing Performance, Hedgewise vs. Prior Momentum Metrics

Hedgewise Momentum Max performance prior to November 2016 is based on a hypothetical model that relies on the same algorithm used in live client portfolios. Beginning in November 2016, a composite of live client performance is used instead. The simulation includes an estimate for all fees and commissions and assumes a Hedgewise fee of 0.7%. All dividends included and assumed re-invested. Earliest data available to run the model begins in 1972; as a result, this 10yr view begins in 1982. See additional full performance disclosures at the end of this article.

The Hedgewise Momentum Max portfolio is the brown bar, and there is not a single datapoint in which it has trailed the simpler momentum metrics. However, there is still a significant ebb and flow to the absolute performance, which naturally exists for the reasons already discussed.

Let's zoom in to look at the shorter 1yr trailing performance. This is the identical prior graph examining the performance of the 100day MA strategy, including bonds and 50% leverage, against the S&P 500. The trailing 1yr performance of the Hedgewise Momentum Max model has been added for comparison.

1yr Trailing Performance, Hedgewise Momentum Max vs. 100day MA vs. S&P 500

See disclosures on prior graph.

While the Hedgewise model drives consistently superior returns and a lower probability of loss, the natural variation still results in periods where it looks bad compared to simpler momentum metrics as well as the S&P 500. This is not a sign that anything is wrong, but rather an inevitability of our odds-based framework.

Conclusion: Reflections on Current Performance

Bringing all of this theory back together with our current reality, it's difficult to defend a drawdown of a Momentum framework in isolation when there's such a purposeful element of randomness involved. That said, I'd say it's pretty easy to see why the market conditions of the past few months are fairly exceptional: the jaw-dropping drops in December remain difficult for anyone to fully explain, especially given the enormous reversal in January. Part of this can be attributed to the Fed making one of the most dramatic and speedy policy U-turns in history - which on its own is extremely unique! The point being that it takes a rare combination of factors to drive a whipsaw, and that's precisely why momentum frameworks have worked so well and will likely continue to work.

It's also really important to focus on the performance of the Hedgewise strategy outside of the period of whipsaw, since the whole idea is that it's a great bet so long as it's not whipsawing. In March alone, Momentum Max is currently +6.5%, compared to a 0.9% gain in stocks. From the launch of the strategy in November 2016 through September 2018, Momentum Max also beat equities by 10%. These are relatively common outcomes because markets are generally efficient.

That said, it's also not like the returns are driven by anything dramatic: you primarily make money in the momentum framework because stocks go up (and bonds, to a lesser extent). This is still the bet you are making, same as if you were in a purely passive index. Will another whipsaw occur? Probably! But the whole beauty of the framework is that you don't have to predict that, or try to avoid it, because it hasn't mattered over the long-run. You just have to sit tight and keep playing the game long enough to see.

Disclosure

This information does not constitute investment advice or an offer to invest or to provide management services and is subject to correction, completion and amendment without notice. Hedgewise makes no warranties and is not responsible for your use of this information or for any errors or inaccuracies resulting from your use. Hedgewise may recommend some of the investments mentioned in this article for use in its clients' portfolios. Past performance is no indicator or guarantee of future results. Investing involves risk, including the risk of loss. All performance data shown prior to the inception of each Hedgewise framework (Risk Parity in October 2014, Momentum in November 2016) is based on a hypothetical model and there is no guarantee that such performance could have been achieved in a live portfolio, which would have been affected by material factors including market liquidity, bid-ask spreads, intraday price fluctuations, instrument availability, and interest rates. Model performance data is based on publicly available index or asset price information and all dividend or coupon payments are included and assumed to be reinvested monthly. Hedgewise products have substantially different levels of volatility and exposure to separate risk factors, such as commodity prices and the use of leverage via derivatives, compared to traditional benchmarks like the S&P 500. Any comparisons to benchmarks are provided as a generic baseline for a long-term investment portfolio and do not suggest that Hedgewise products will exhibit similar characteristics. When live client data is shown, it includes all fees, commissions, and other expenses incurred during management. Only performance figures from the earliest live client accounts available or from a composite average of all client accounts are used. Other accounts managed by Hedgewise will have performed slightly differently than the numbers shown for a variety of reasons, though all accounts are managed according to the same underlying strategy model. Hedgewise relies on sophisticated algorithms which present technological risk, including data availability, system uptime and speed, coding errors, and reliance on third party vendors.

2018 Year-In-Review: This Too Shall Pass
Posted in Market Commentary on 2018-12-21

Summary

  • 2018 was an extremely difficult year for investors, with losses unfolding across every major asset class. Typical hedges, such as bonds and gold, were ineffective, and an atmosphere of unpredictable volatility diminished returns for nearly all styles of risk management.
  • Against this backdrop, Hedgewise frameworks lost between 2% - 14% YTD, which remains substantially better than most comparable funds and within expectations given these circumstances.
  • Still, it is natural to wonder whether this environment is radically different than historical precedents, and if it threatens the efficacy of risk-managed frameworks.
  • History suggests that this year's returns are relatively similar to other periods when the Fed was reaching the end of its tightening cycle. This tends to drive fear, uncertainty, and lower valuations across all asset classes, despite many inherent contradictions.
  • Such fear-driven volatility cannot persist forever, simply because you cannot experience inflation and deflation, or recession and growth, simultaneously. So long as high uncertainty persists, large short-term price swings usually remain, but they have little bearing on your long-term outlook or the efficacy of risk management more broadly.

Introduction: A Crummy, No-Good, Typical Year

2018 was not a good year for investors, of any asset, in any style, living in any country. The last 12 months have often felt like a never-ending deluge of worry, including rising interest rates, peak profits, trade war, Brexit, et al. Yet traditional safe-havens like gold and bonds have suffered losses alongside stocks, and recoveries in any asset class have quickly and violently reversed. That said, as ugly as it's been, it's quite a bit more familiar than it might seem, and quite a bit less scary than you probably feel.

First, to dispel the most worrisome economic myth floating around in the media: asset performance trends this year are not radically different than the past, and remain very consistent with what you'd expect when the Fed is aggressively battling inflation and withdrawing mountains of liquidity via higher interest rates. Higher rates force all assets to be more heavily discounted, and everyone also wonders how high rates will need to go, and whether the Fed is driving us straight towards a recession, deflation, hyperinflation, or some mix. Just like that, all assets start pricing in more risk.

Of course, it feels like there's lots else going on, like terrifying tweets and tariff wars. But it's hard to put much weight on those factors when you saw just as much volatility and cross-asset decline in nearly every other year in history that the Fed was behaving similarly. A more likely explanation is that when investors are particularly on edge, they react more strongly to the news, whatever it might be. While this results in very real price swings, it also makes it far less likely that losses persist.

Ironically, this environment can drive moderately high losses in risk-managed frameworks precisely because many of the perceived risks are unsubstantiated. Hedging will be less effective because each individual asset class begins pricing in a worst-case scenario, despite the fact that all of those scenarios cannot simultaneously co-exist. Risk measurements will often fail to signal much chance of a major crash - which is usually quite correct! - but investors might succumb to panic for a while anyway. Financial theory suggests that the most effective path through such turbulence is to wait for the noise to pass, and while this usually seems quite obvious in retrospect, it will almost certainly test your nerves along the way.

Fortunately, this scenario has played out many times in the past, and has always resolved without any breakdown in fundamental economics or the theory of risk management. Though losses this year are certainly unpleasant, Hedgewise has actually navigated the environment fairly well from a historical perspective and relative to competitive funds. While it's natural to wonder how such a strange, crummy year could fit in the bigger picture, the reality is that it is much the same as all the other strange, crummy years we've had, and will soon pass much like the rest.

Review of Hedgewise 2018 Performance

2018 has certainly been unfamiliar in the context of the recent nine-year bull market. Among other fun headlines, there haven't been more radical daily and monthly swings since the depths of the financial crisis, the last time this many asset classes had simultaneous losses was 1901, and the last time investors felt this nervous was the Nixon era in 1972.

Here's a look at year-to-date returns for every major class:

2018 YTD Performance by Asset Class

Data as of Dec 20 2018. Source: US Treasury, Fed Reserve Economic Data, Bloomberg, Hedgewise. Based on end-of-day index prices and includes all dividends and coupons assumed re-invested monthly.

Given this picture, it comes as little surprise that most investment strategies have fared poorly; in years like this one, losses are essentially unavoidable. With that in mind, the primary question becomes how well Hedgewise frameworks have helped to mitigate the damage, and whether this calls into question the effectiveness of the underlying financial theories. Fortunately, there is ample evidence that Hedgewise techniques have continued to be effective.

From a 10,000 foot view, the most obvious explanation for this year's negative returns is that Hedgewise strategies are multi-asset, and frequently use leverage to balance risk and amplify potential returns; when most assets are down, this will result in losses. A simple, static example of implementing such a strategy at a moderate risk level might look like 100% bonds, 60% stocks, and 40% commodities. This year, that portfolio would have lost approximately 15%, depending on its exact weighting of commodities.

For comparison, Hedgewise strategies have incurred YTD losses between 2% - 14%, depending on your product mix and risk level. This is a very compelling outcome, especially in a year when many traditional risk management mechanisms were ineffective. To provide a less theoretical comparison, the Hedgewise Risk Parity strategy has also outperformed all comparable mutual funds throughout the year.

Hedgewise Performance Vs. Major Risk Parity Mutual Funds, 2018 YTD

Data as Dec 20 2018. Source: Morningstar, Bloomberg. Includes an estimate for all dividends and fees. Hedgewise performance is a composite of live client portfolios at the High risk level.

There are a couple of notable highlights from this graph. While it's clear that Hedgewise risk management techniques have largely outperformed the competition throughout the year, no fund was able to escape losses altogether. This adds weight to the idea that losses were inevitable in nearly any form of the strategy framework, but that Hedgewise handled this better than most. While there are some ebbs and flows to each fund's performance, due to myriad differences in the definition of risk and portfolio composition, it's a relatively minor give-and-take that adds up over time (except in the case of Wealthfront, whose performance has continued to raise concerns). Hedgewise will certainly have some months that look better and some that look worse, but the short-term deviations mean much less than the longer-term edge built into each portfolio.

Still, even if Hedgewise has done better relatively, is there something wrong with risk managed strategies more broadly? Why has hedging been ineffective, and how do you know that won't continue? Why can't some of these pullbacks be more nimbly avoided?

To answer these questions, it's helpful to re-examine how the core financial theory is supposed to work and relate that to a few similar historical examples.

The Fed and Peak Uncertainty

When the Fed is trying to figure out how high to raise interest rates - and especially when they are getting close to a level that may cause a recession if they go too high - markets often become quite perplexed. There is simultaneously inflationary pressure, and recessionary pressure, but neither a recession nor high inflation has actually happened. When there are so many different, scary possibilities at once, lots of assets can lose value in contradictory ways while this limbo continues.

From a risk management perspective, this tipping point creates a conundrum. Hedging doesn't work very well, but some assets are becoming inevitably undervalued since only one actual scenario can unfold. Market prices frequently lurch from one worst case projection to another, yet those fears aren't frequently based on any actual events or hard data. Does it make sense to continue operating normally in such an environment, knowing that losses typically result, or is it better to try and avoid it altogether?

There are two key theoretical insights that suggest simply taking the lumps and waiting it out. First, if many assets have begun trading beneath fair value, then it follows that there is a much higher chance of short-term gains than is typical. Second, the longer that losses persist, the more pressure will build to the upside, since value tends to accumulate over time. Together, these ideas imply a fairly rapid but difficult to predict turnaround which easily reverses any net losses. Importantly, note that this does not assume that all assets will recover - oftentimes one bad scenario will unfold - only that some assets are poised for a significant rebound during which risk-managed frameworks will be well-positioned to benefit.

To test this is relatively simple: the pattern you'd need to see is that when many asset classes lose value simultaneously against a backdrop of rising interest rates, you go on to have a relatively rapid and substantial rebound in some of those assets, along with a corresponding rebound in the risk-managed frameworks. Since this should hold true regardless of what event winds up unfolding, this analysis includes every event since 1954 where interest rates were rising while stocks and some commodities were falling. Notice that our current year is included for reference in the final row.

Year Length (Months) Stocks Bonds Copper Gold
1957 17 -5.9% -5.2% N/A N/A
1960 2 -5% -0.73% -6.4% N/A
1962 3 -18.8% -0.7% -6% N/A
1966 5 -12.4% -1.7% -36.3% N/A
1970 8 -13.9% -4.5% -5.7% 0.9%
1974 5 -22.8% -2.2% -46.4% -9%
1975 3 -11.7% 0% 2.2% -13.8%
1978 3 -5.9% -1.9% 4.5% -6.4%
1980 6 -3.5% -6.3% -23.5% 22.8%
1982 2 -5.8% -2.6% 0.7% -5.7%
1984 11 -5.5% -3.2% -14.8% -5.4%
1994 2 -8.1% -10.9% -2% 0.8%
1997 2 -3.2% -3.1% 8.7% 1.7%
2015 9 -0.7% -4.6% -24.9% -9.6%
2018 12 -6.0% -5.2% -17.4% -3.6%
Source: US Treasury, Fed Reserve Economic Data, Bloomberg, Hedgewise, CME. Based on end-of-day index prices and includes all dividends and coupons assumed re-invested monthly.

Note that this covers a very broad spectrum of history. Sometimes the economy was battling hyperinflation (late 70's), other times it was already in recession (early 70's), and sometimes it was just nervous (2015). Sometimes the Fed was in the midst of raising rates too high, and other times it needed to go on to raise rates more. None of that matters much to our thesis.

Here's a look at how each asset class performed just three months after the end of each of the above periods. It would be natural for some asset classes to continue falling, but also to see significant rallies elsewhere. On net, the rallies should appear more common and larger in size than any additional losses.

Performance by Asset Class, Three Months Forward

Source: US Treasury, Fed Reserve Economic Data, Bloomberg, Hedgewise, CME. Based on end-of-day index prices and includes all dividends and coupons assumed re-invested monthly.

Every data point fits the theory, and it's also interesting to note that stocks and bonds both rallied 85% of the time, and losses were relatively light even when they did persist. This traces back to the downward bias of uncertainty; once prices are so cheap, even the asset classes that go on to perform the worst don't have all that much further to fall since such pessimism was already afoot.

Both Hedgewise risk-managed frameworks suffer in these circumstances, but that is by design since these events are quite infrequent, and because this 'rebound effect' provides such an effective antidote. First, let's take a look at how each strategy performed during these same periods. The first column shows the net performance for each strategy, and the second the worst peak to trough drawdown (if worse than the net loss), which gives a better sense of how the 'worst of it' felt.

Year RP Max RP DD MM Max MM DD
1957 N/A N/A -13.3% -16.2%
1960 N/A N/A -8.5% --
1962 N/A N/A -13.2% --
1966 N/A N/A -12% -13.1%
1970 N/A N/A -13.9% --
1974 -16% -- -4.5% -5.8%
1975 -18.3% -- -19.6% --
1978 -15.3% -- -12.2% --
1980 -12.9% -24.5% -0.3% -19.7%
1982 -5.3% -- -2.8% --
1984 -22.9% -- 5.6% -7.8%
1994 -10.8% -- -7.1%
1997 -3.2% -4.5% -4.4% -5.3%
2015 -5.9% -8.1% -1.9% -10.8%
2018 -9.3% -15.8% -13.7% -19.8%
Source: Hedgewise. Momentum performance based on a hypothetical model that relies on the same algorithm used in live client portfolios. Data based on publicly available index or asset price information and all dividend or coupon payments are included and assumed to be reinvested monthly.

Next, let's examine how the strategies performed directly following these months; you'd frequently expect the rebound to be rather fast, but even if not, it should eventually result in large gains since value accrues as you wait. I've separated each strategy into forward looking one month, three month, six month, and one year gains to highlight the speed and size of each recovery.

Risk Parity Subsequent Returns (1972 Onward)

See prior disclosure.

Momentum Subsequent Returns (All Data Points)

See prior disclosure. Data points prior to 1972 include a hypothetical portfolio of only stocks and bonds constructed using the same risk algorithms in place today, though a more limited set of risk information was available at the time.

For both strategies, the average one month return is over 4%, the average three month return is over 10%, the average six month return is over 20%, and the average one year return is close to 40%. Now, the main caveat to these numbers is that they all measure forward starting after the end of the "peak uncertainty" stretch, and it is impossible to say whether this month will mark that point. However, consider that we are already 12 months into our current stretch, which is far longer than the average duration historically and has only been exceeded once, in 1956, when the pain endured for 17 months total. Eventually, the Fed either breaks the economy, or inflation rockets, or the tightening cycle ends uneventfully, and Hedgewise frameworks have handled every scenario with ease. Until then, you usually run into years just like this one, and while that may provide little comfort, it's important to understand that we are not in uncharted territory, nor has Hedgewise performance been any worse than you'd expect.

If these stretches are always so nasty, can't we just find a way to skip them? The data above provides much in the way of a retort. These periods are really infrequent; you see two to four per decade, and on average they last only a few months. Once they end, the same risk management mechanisms that appeared to be failing suddenly work wonderfully. In a way, this is precisely the environment where you choose to accept losses as the outcome: it's an environment with a rebound literally built-in.

Conclusion: Evaluating 2018

On the one hand, 2018 definitely counts as an exceptional year. It joins the handful of events each decade that stand out for their awfulness, and it absolutely feels unfamiliar because you have to go back to the 1970s and 80s to find a year quite this bad. Then again, those were the very decades when the Fed last had to seriously wield monetary policy to fight potential inflation; perhaps it should be of little surprise that is what we are facing again now. When the central bank is the main concern, it's relatively common to hit these stretches of 'peak uncertainty', when many assets fall together despite inherent contradictions, and markets lurch from one panic to another.

Amidst this, risk-managed strategies like Risk Parity and Momentum will have losses, but this doesn't mean that they are working poorly. In fact, Hedgewise has continued to consistently outperform large competitive products. Historically speaking, a similar magnitude and breadth of cross-asset losses has often resulted in losses of 20%, while current YTD losses remain well-below that. Similarly, even a purely passive hedged portfolio would likely be performing worse. These are all signs that the risk management techniques in place continue to work about as well as they should, despite the difficult reality that these circumstances entail a substantial drawdown.

Looking forward, though, this drawdown is entirely about context. It is being driven by a heightened anxiety that has driven down the price of nearly all risky assets, and given that only one economic reality can unfold, this means that one or many of these assets are mispriced. The same risk management mechanisms that have driven losses, like hedging and leverage, will also be positioned to take advantage of the subsequent rallies that have always resulted with time. This is very different than a loss in a single asset class or speculative instrument (like, say, Bitcoin), which explains why recoveries have been so convincing and consistent when they occur.

Certainly, this year feels crummy, but it's about as crummy as you'd expect it to be, and it's the kind of crummy where you usually have the losses that we have. It will probably continue to feel very crummy until it is finally clear which big mistake is being made, if any. Then, suddenly, it will start to look much better, and it doesn't really matter if we are in a recession or high inflation or neither. It only matters that we finally know, after which 2018 will be just another data point of a typical bad year.

Disclosure

This information does not constitute investment advice or an offer to invest or to provide management services and is subject to correction, completion and amendment without notice. Hedgewise makes no warranties and is not responsible for your use of this information or for any errors or inaccuracies resulting from your use. Hedgewise may recommend some of the investments mentioned in this article for use in its clients' portfolios. Past performance is no indicator or guarantee of future results. Investing involves risk, including the risk of loss. All performance data shown prior to the inception of each Hedgewise framework (Risk Parity in October 2014, Momentum in November 2016) is based on a hypothetical model and there is no guarantee that such performance could have been achieved in a live portfolio, which would have been affected by material factors including market liquidity, bid-ask spreads, intraday price fluctuations, instrument availability, and interest rates. Model performance data is based on publicly available index or asset price information and all dividend or coupon payments are included and assumed to be reinvested monthly. Hedgewise products have substantially different levels of volatility and exposure to separate risk factors, such as commodity prices and the use of leverage via derivatives, compared to traditional benchmarks like the S&P 500. Any comparisons to benchmarks are provided as a generic baseline for a long-term investment portfolio and do not suggest that Hedgewise products will exhibit similar characteristics. When live client data is shown, it includes all fees, commissions, and other expenses incurred during management. Only performance figures from the earliest live client accounts available or from a composite average of all client accounts are used. Other accounts managed by Hedgewise will have performed slightly differently than the numbers shown for a variety of reasons, though all accounts are managed according to the same underlying strategy model. Hedgewise relies on sophisticated algorithms which present technological risk, including data availability, system uptime and speed, coding errors, and reliance on third party vendors.

October 2018: Better Times Ahead
Posted in Market Commentary on 2018-10-11

Summary

  • A significant spike in interest rates has caused steep losses across all asset classes thus far in October, but a mix of higher rates and lower stock valuations bodes extremely well for future returns.
  • Periods of Fed tightening are typically quite choppy, but have been consistently followed by rallies in one or more asset classes for structural reasons.
  • Hedgewise continues to minimize losses this year via risk management, which has driven outperformance compared to benchmarks.
  • This loss reduction will eventually add to a significant boost in long-run returns, but patience is required while the market sorts itself out.

Current Outlook: Little Reason to Worry

Current economic conditions - a bear market in bonds and commodities, largely driven by Fed tightening - have historically preceded excellent returns in both Hedgewise strategies. Unfortunately, these periods also usually include significant volatility, like what we saw in January and are now experiencing again in October. But it is extremely unlikely that these losses will persist over the next six to twelve months.

To understand why, let's first take a look at year-to-date returns in each asset class.

Asset Class YTD Returns, 2018

Data as of Oct 10 2018. Source: US Treasury, Fed Reserve Economic Data, CNN Money. Based on end-of-day index prices and includes all dividends and coupons assumed re-invested monthly.

It's fairly rare to see this kind of dual bear market in bonds and commodities, since higher interest rates usually happen alongside a strong, growing economy which buoys the prices of raw materials. Yet this data suggests that the Fed is essentially pre-empting inflation: it wants to cool off growth before it gets out of control, and so far as inflation is concerned, it appears to be doing a great job.

There have only been three other periods since 1954 with similar economic conditions: 1980, 1981, and 1984. In each, the Fed was also corralling inflation, and asset class returns were eerily similar to today:

1yr Trailing Returns by Asset Class

DateBondsCopperStocks
March 1980-18%-6%6%
August 1981-15%-13%5%
April 1984-9%-16%4%
See previous note.

Once the Fed has raised rates enough to produce these outcomes, there are only two logical possibilities for the future: a) the economy is strong enough to sustain it, and stocks do well, or b) the economy is not strong enough, so rates go back down and bonds do well. Here's how the following one-year returns looked for each asset class in the above periods:

1yr Forward Returns by Asset Class

DateBondsCopperStocks
March 19809%1%40%
August 198135%-18%1%
April 198427%-6%15%
See previous note.

This is quite consistent with the theory. Stocks did well in 1980, bonds did well in 1981, and both managed to do well in 1984 (this is known as a a Fed "soft landing"). Perhaps most importantly, though, both Hedgewise products also went on to do fantastically in every period.

1yr Forward Returns, Risk Parity Max and Momentum Max

DateRisk ParityMomentum
March 198044%45%
August 198150%57%
April 198435%37%
Source: Hedgewise. Momentum performance based on a hypothetical model that relies on the same algorithm used in live client portfolios. Data based on publicly available index or asset price information and all dividend or coupon payments are included and assumed to be reinvested monthly. See full disclosures at end of article.

While these numbers are comforting on face value, it's important to emphasize that these periods each came with a ton of volatility, as investors were just as nervous about interest rates and the economy then as they are now. For example, here's how markets looked from April 1979 to March 1980.

Asset Class Performance, April 1979 to March 1980

See prior disclosures.

Note that in October 1979, stocks lost 5%, bonds lost 8%, and copper lost 17% in a single month! Stocks recovered from there, but then went on to fall 10% again in February and March of 1980, while bonds wound up down 18% altogether. Copper dropped 30% in about 2 months. Does it start to feel a bit familiar?

Notably, there was also a huge amount of daily volatility over this stretch. Stocks were down 1% or more on 14 days, including a single day loss of 3%.

Clearly, such years will be stressful, and large single day movements have a high emotional toll. It's hard not to wonder whether it makes more sense to simply sit on the sidelines for a while. But there's a very strong structural case for staying patient, especially as it relates to Fed tightening.

The Structural Case: Interest Rates and Returns

It's very important to differentiate the impact of interest rate movements on asset returns from systemic events (like the mortgage crisis) or economic slowdowns. This is because with the latter categories, volatile markets can sometimes be indicative of more problems lurking beneath the surface. For example, in the mortgage crisis, certain kinds of loans started to default before others, but markets were not yet pricing in the full impact of the issue.

On the other hand, interest rates alone are extremely transparent and the Fed works quite hard to avoid surprising the market much. Interest rates are also not an economic problem in isolation; they only become a problem if companies stop growing sufficiently or generating enough cash flow, etc. As such, if you assume that everything in the economy will stay the same, except for a move up in interest rates, you can calculate quite precisely how much it should impact markets. For example, with the way rates have moved so far in October, you'd expect bonds to be worth about 4% less and stocks to be worth around 3-6% less, simply based on a present value formula and discounted cash flows. As of October 10th, bonds are down 4.04% month-to-date and stocks are down 3.63%.

While those immediate losses are painful, potential gains have been quite literally 'moved' into the future. The expected return on all assets has gone up definitionally, since interest rates are a component of expected return. Little wonder that this is usually a terrible time to sell!

To test this, I've isolated every single rolling twelve-month period when interest rates went up by a similar amount to this year, regardless of what happened in other assets. The following shows the subsequent one-year returns in the Risk Parity Max and Momentum Max products.

1yr Forward Returns in Risk Parity and Momentum, Periods of Rising Interest Rates

Source: Hedgewise. Momentum performance based on a hypothetical model that relies on the same algorithm used in live client portfolios. Data based on publicly available index or asset price information and all dividend or coupon payments are included and assumed to be reinvested monthly. See full disclosures at end of article.

There were exactly three months when Risk Parity and Momentum went on to do poorly over the next year, and all of them were prior to Black Monday in 1987. Even if you include that, it still leaves a 94% chance of positive returns. It is also relatively difficult to parallel our current environment to the one in 1987. Most notably, stocks were up about 40% from January through August that year, with almost no downside volatility.

Outside of 1987, annual returns averaged over 20% for both strategies, and it didn't matter much which asset wound up rallying. Either interest rates came down, stocks rallied, or some combination of the two.

Is There A Better Way to Hedge?

While the forward-looking numbers may provide some comfort, it's still fair to ask whether these losses could have been avoided in the first place. If rates are going up, should we just sit on the sidelines to avoid the possibility of market corrections?

The problem with this logic is that you'd need perfect foresight as to when long-term interest rates are about to go up by 1% or more, and even if you had that, you'd still only be right about a pullback about 30% of the time. To illustrate, here's the same graph as above, but with trailing one-year returns instead of forward one-year returns. This shows you what loss (or gain) you would have avoided if you exited each strategy one year before interest rates hit their peak.

1yr Trailing Returns in Risk Parity and Momentum, Periods of Rising Interest Rates

Source: Hedgewise. Momentum performance based on a hypothetical model that relies on the same algorithm used in live client portfolios. Data based on publicly available index or asset price information and all dividend or coupon payments are included and assumed to be reinvested monthly. See full disclosures at end of article.

Sitting on the sidelines was still a losing bet about 60% of the time - and this was with perfect predictions of rate spikes!

As a result, Hedgewise generally accepts that you'll have a bad year every now and again, but it still seeks to minimize the damage via its risk management techniques. For example, Hedgewise risk indicators for the bond market spiked in January of this year as well as at the beginning of October, and bond exposure was drastically reduced as a result. The net impact of these techniques continues to drive outperformance for Hedgewise versus all major competitors.

Competitive Risk Parity Funds YTD Performance

ProductYTD
Hedgewise RP High -0.7%
AQR -4.8%
Invesco -4.1%
Wealthfront -14.3%
Source: Morningstar, Bloomberg. Includes an estimate for all dividends and fees.

Traditional Diversified Mix YTD Performance

Type (Ticker)YTD
Conservative (AOK) -2.9%
Moderate (AOM) -3.4%
Aggressive (AOA) -3.1%
Source: Morningstar, Bloomberg. Includes an estimate for all dividends and fees.

Wrapping Up: A Typical Volatile Year

While the Great Recession remains fresh on most of our minds, all signs indicate that the economy remains relatively normal. Unemployment is under 4%, interest rates are rising slowly, and markets have all been functioning as they should. While the extreme volatility thus far in October feels exceptional, it is quite similar to the volatility markets witnessed in other periods of Fed tightening. There's also a logical explanation for why markets get re-priced when rates go up, and why this makes future positive returns all the more likely.

Neither a year nor a week like this past one suggests that we are in uncharted territory, or that worse times are imminent. There's a reasonably good chance that markets will continue to be volatile for the next few months, just as they were for most of 1980 and 1984. But there's about a 95% chance that at least one asset class will break significantly higher within the next year. Fortunately, Hedgewise strategies are built so that you don't have to figure out which asset class it will be.

Disclosure

This information does not constitute investment advice or an offer to invest or to provide management services and is subject to correction, completion and amendment without notice. Hedgewise makes no warranties and is not responsible for your use of this information or for any errors or inaccuracies resulting from your use. Hedgewise may recommend some of the investments mentioned in this article for use in its clients' portfolios. Past performance is no indicator or guarantee of future results. Investing involves risk, including the risk of loss. All performance data shown prior to the inception of each Hedgewise framework (Risk Parity in October 2014, Momentum in November 2016) is based on a hypothetical model and there is no guarantee that such performance could have been achieved in a live portfolio, which would have been affected by material factors including market liquidity, bid-ask spreads, intraday price fluctuations, instrument availability, and interest rates. Model performance data is based on publicly available index or asset price information and all dividend or coupon payments are included and assumed to be reinvested monthly. Hedgewise products have substantially different levels of volatility and exposure to separate risk factors, such as commodity prices and the use of leverage via derivatives, compared to traditional benchmarks like the S&P 500. Any comparisons to benchmarks are provided as a generic baseline for a long-term investment portfolio and do not suggest that Hedgewise products will exhibit similar characteristics. When live client data is shown, it includes all fees, commissions, and other expenses incurred during management. Only performance figures from the earliest live client accounts available or from a composite average of all client accounts are used. Other accounts managed by Hedgewise will have performed slightly differently than the numbers shown for a variety of reasons, though all accounts are managed according to the same underlying strategy model. Hedgewise relies on sophisticated algorithms which present technological risk, including data availability, system uptime and speed, coding errors, and reliance on third party vendors.

September 2018: Amidst Volatile Markets, Hedgewise Leads the Pack
Posted in Market Commentary on 2018-09-04

Summary

  • Since the market bottom this spring, Hedgewise Risk Parity has gained 8% and Momentum has gained 10%, demonstrating the resilience of both frameworks and the power of remaining patient regardless of market conditions.
  • Generally, most quantitative and traditional frameworks have underperformed Hedgewise over this period, which can be traced to core theoretical principles that give Hedgewise a persistent edge.
  • While US equities have been the best performing asset class this year, substantial risks such as the trade war and rising interest rates remain. Protecting against these downside risks has resulted in a slight lag compared to the S&P 500, but that is entirely by design.

Few Winners in 2018 Amidst Market Chaos

If you look anywhere outside of US stocks, 2018 has been a pretty terrible year for investors. Both emerging markets and many commodities have entered a full-fledged bear market, and the general environment of heightened volatility led to many funds taking money off the table and missing the ensuing recovery. Bonds failed to hedge the market pullbacks in February and March given fears of runaway inflation, and other typical safeguards like gold were beaten down by a strong US dollar.

Performance by Asset Class, YTD

Source: Bloomberg, Hedgewise. Includes an estimate for all dividends and fees. Hedgewise performance is a composite of all live client portfolios in a given strategy and risk level.

Despite this difficult environment, both Hedgewise strategies have performed quite well, with Risk Parity up 5% (at the Max risk level) and Momentum up 7.3%. Given that both frameworks are frequently exposed to bonds and commodities, these results are quite powerful and add significant weight to the narrative that Hedgewise clients have no need to time the markets.

The challenges of navigating this kind of environment can be seen in the poor year-to-date performance across the majority of quantitative funds and even in traditional diversified portfolios:

Competitive Risk Parity Funds YTD Performance

Mutual FundYTD
AQR -2.9%
Invesco -1.1%
Wealthfront -6.8%
Source: Morningstar, Bloomberg. Includes an estimate for all dividends and fees.

Traditional Diversified Mix YTD Performance

Type (Ticker)YTD
Conservative (AOK) -0.9%
Moderate (AOM) -0.2%
Aggressive (AOA) 1.4%
Source: Morningstar, Bloomberg. Includes an estimate for all dividends and fees.

The outperformance of Hedgewise products can be traced directly to its hyper-focus on avoiding what is known as "asymmetric risk". Simply, this is when there's a chance that some part of your portfolio will perform badly in isolation, and nothing else in your portfolio offsets it. For example, normally a commodity crash would be accompanied by a rally in bonds, since it would suggest lower overall inflation. But this year, commodities have crashed while bonds have lost money as well. Similarly, international bonds would usually rally when international equities crash, but both are negative year-to-date.

The reason for both trends is that the US dollar has had an incredible rally this year, which lowers the value of international stocks and bonds as well as dollar-priced commodities. This is a classic asymmetric kind of risk, and it is exactly why Hedgewise avoids international exposure and has a measure for asymmetric risk built into every asset class.

However, Hedgewise clients are often less interested in comparisons to competitive funds and more interested in performance versus the S&P 500 itself. After all, the goal is to achieve equity-like returns (at the High and Max risk levels) with substantially less risk. Given that, I want to focus the rest of this analysis on how years like this current one fit into the bigger strategic picture, and why underperformance compared to equities is often exactly what you'd expect.

Risk Parity: Stability Above All

Risk Parity is all about balance; it accounts for every possible economic scenario, and constantly builds in a hedge for each. As a result, it will always be holding a mix of bonds, commodities, and equities. Thus, it is somewhat intuitive to achieve a return lower than equities when bonds and commodities are underperforming. But as soon as you hit one period of recession (when bonds usually rally) or high inflation (when commodities usually rally), you easily make up the difference.

The key is that you are constantly trading near-term upside for long-term stability; you'd rather have a boring, steady 8% return every year regardless of what equities are doing. The rub is that you'll probably underperform equities about 50% of the time! You can also run into lots of bull markets where you'll lag the net performance of the S&P 500 for many years. In exchange, you can worry much less about whether next year is going to be a repeat of 2008. Historically speaking, so long as you've waited at least 10 years, you've outperformed the S&P about 85% of the time at the High risk level and 99% of the time at Max.

Still, it is difficult to gauge the strategy's success in years like this one, as you wonder whether a simple stock portfolio might make sense. Fortunately, there is a way to directly measure the 'stability' effect even over shorter timeframes to gauge how well the theory is working.

The following chart shows the distribution of all daily returns of the S&P 500 thus far in 2018. Notice the long 'left-tail' of negative returns; you had to deal with a couple of single days with losses as high as 4%!

S&P 500 Daily Return Distribution, YTD

Source: Bloomberg, Hedgewise

Now let's look at the same distribution for the Risk Parity High strategy. If it is working as it should, the distribution should be much tighter, and have a shorter left-tail.

RP High Daily Return Distribution, YTD

Source: Hedgewise. Hedgewise performance is a composite of all live client portfolios in a given strategy and risk level.

Exactly as the theory predicts, the Risk Parity portfolio achieved a far higher level of stability compared to equities. The portfolio had more positive daily returns, and fewer negative ones; it also protected clients from the worst of the equity volatility. These attributes are what will continue to drive the portfolio's resilience over the long-run, though equity underperformance will very frequently be part of the story.

Momentum: Lean Into Safety, Away From Risk

Unlike Risk Parity, the Momentum framework does not rely on underlying balance. While it can hold various asset classes, it is usually dominated by equities, as its goal is to outperform the S&P 500 at a similar level of risk. To achieve this, it is constantly evaluating the current environment for stocks. When it is deemed relatively 'safe', the portfolio will overweight equities, and vice versa.

Importantly, this means that it will often be underweight stocks in risky environments, since this is what helps protect the portfolio from downside. Given the events of this year, perhaps it is little surprise that equity exposure has generally been lighter than it was in 2017.

The theory behind this is that stocks generally yield a positive return in 'normal' environments, since any reasonable investor demands that. However, once in a while, asymmetric risks appear to the downside (e.g. real estate bubble, dot-com crash, junk bond crisis, etc.). Hedgewise simply behaves more and more conservatively as the risk builds. Roughly speaking, Hedgewise trims exposure as the risk of a systemic event reaches between 20-30%; in other words, Hedgewise expects to be wrong about a crash occurring about 70-80% of the time.

To visualize the impact of this, the following chart isolates every year of gains in the S&P 500, and compares the returns of the Momentum "High" strategy over the same period. The dots under the red line mean the Momentum strategy did worse than equities, and vice versa.

S&P 500 Performance vs. MM High Performance, 1972 to Present (Only Stock Gains)

Source: Hedgewise. Momentum performance based on a hypothetical model that relies on the same algorithm used in live client portfolios. Data based on publicly available index or asset price information and all dividend or coupon payments are included and assumed to be reinvested monthly. See full disclosures at end of article.

It may initially be surprising to see so many years of underperformance! However, there are very compelling reasons to give up these gains. Notice that you tend to make much more in the good years than you lose in the bad ones, and there are also about 2x as many dots above the red line as below. Even more importantly, playing it safe allows you to avoid the occasional catastrophe, as you can see in a similar chart that isolates all of the years of S&P 500 losses:

S&P 500 Performance vs. MM High Performance, 1972 to Present (Only Stock Losses)

See disclosures in previous chart.

Because the strategy behaves so conservatively in risky environments, it has historically avoided about 90% of stock crashes. Essentially, this boils down to a philosophy of being aggressive in good times but cautious in dangerous ones; you lean into safety, but away from risk.

As a result, you'll frequently underperform the S&P 500 in volatile years like 2018, but the amount you give up will be relatively small compared to your outperformance in better years and your ability to avoid significant crashes.

Comparing this to the available live client performance, the Momentum "Max" product had a return of 31% in 2017 compared to 22% in the S&P 500, or a difference of +9%. Year-to-date, it has a return of 7.3% compared to 9.9% in the S&P 500, or a difference of -2.6%. This is exactly consistent with expectations! Last year's outperformance more than outweighs the lag of this year, and while a more significant crash didn't wind up occurring, the risks were high enough to demand caution.

Looking Forward: Not Chasing the Peak

Perhaps the most consistent theme of 2018 is that the world feels much less steady than it did last year. The Fed is threading a nearly impossible needle of controlling inflation without impeding growth, and no one quite knows what to make of the ongoing trade wars. Various emerging markets are on the brink of crisis, and the Chinese economy suddenly appears quite vulnerable.

None of these risks have weighed much on the US economy so far, and it's entirely possible that they never will. But it feels increasingly likely that we are near a peak, and dangerous to try and predict its top. Luckily, there's no need to do so with either of the Hedgewise frameworks. Clients have continued to accrue gains regardless of external market conditions and have vastly outperformed most competitive funds. While there has been slight underperformance compared to the S&P 500, this is quite consistent with theoretical expectations. So long as you stay the course, the odds remain heavily tilted in your favor.

Disclosure

This information does not constitute investment advice or an offer to invest or to provide management services and is subject to correction, completion and amendment without notice. Hedgewise makes no warranties and is not responsible for your use of this information or for any errors or inaccuracies resulting from your use. Hedgewise may recommend some of the investments mentioned in this article for use in its clients' portfolios. Past performance is no indicator or guarantee of future results. Investing involves risk, including the risk of loss. All performance data shown prior to the inception of each Hedgewise framework (Risk Parity in October 2014, Momentum in November 2016) is based on a hypothetical model and there is no guarantee that such performance could have been achieved in a live portfolio, which would have been affected by material factors including market liquidity, bid-ask spreads, intraday price fluctuations, instrument availability, and interest rates. Model performance data is based on publicly available index or asset price information and all dividend or coupon payments are included and assumed to be reinvested monthly. Hedgewise products have substantially different levels of volatility and exposure to separate risk factors, such as commodity prices and the use of leverage via derivatives, compared to traditional benchmarks like the S&P 500. Any comparisons to benchmarks are provided as a generic baseline for a long-term investment portfolio and do not suggest that Hedgewise products will exhibit similar characteristics. When live client data is shown, it includes all fees, commissions, and other expenses incurred during management. Only performance figures from the earliest live client accounts available or from a composite average of all client accounts are used. Other accounts managed by Hedgewise will have performed slightly differently than the numbers shown for a variety of reasons, though all accounts are managed according to the same underlying strategy model. Hedgewise relies on sophisticated algorithms which present technological risk, including data availability, system uptime and speed, coding errors, and reliance on third party vendors.

Best-In-Class Risk Parity Performance
Posted in Investment Strategy on 2018-06-08

Summary

  • The Hedgewise Risk Parity product has outperformed competitive funds by an average of 4.2% in 2018 and 3.3% annually since 2016
  • While 'smart beta' products like Risk Parity share a common philosophy, performance can significantly differ due to the strategic and operational approach
  • By stripping down the core financial theory to its most essential benefits at the lowest possible cost, Hedgewise has emerged as an industry pioneer and generated consistent outperformance for its clients

A Framework for Evaluating Hedgewise

Two core ideas drive the investment product philosophy at Hedgewise:

  • Financial theory can be used to engineer higher returns with lower risk, and
  • Great care must be taken during implementation to ensure the concepts are properly represented at a reasonably low cost

To evaluate the first idea, the focus must be on how a broad theory is supposed to work, and whether those assumptions can be validated in the real world. This is the substance of a majority of the research published here, but can only really convince you of whether 'smart beta' concepts like Risk Parity make sense in general. In that regard, it's not as much of a judgment on Hedgewise itself as on the theoretical frameworks being used.

However, the implementation of these frameworks is far more nuanced than constructing indexes like the S&P 500. Managers face questions like how to define risk, how many assets to include in the portfolio, and how often to trade. If the answers are too complex, they can often drive up costs. If the answers are too simple, they can fail to properly implement the theory. These issues will dramatically impact performance across managers.

Given that, the simplest and most direct way to evaluate Hedgewise is to compare its long-term performance to other competitive funds. This is now easy to do for Risk Parity, as there are three large competitive mutual funds (AQR, Invesco, and Wealthfront). Note that Hedgewise data is based on a compilation of live client portfolios at the High risk level, which had the closest level of overall volatility to the other funds, and includes all costs and fees.

Performance Summary Since 2016

YTD20172016Ann.
Hedgewise1.7%18.2%10.8%12.9%
Avg. Competitor -2.5% 13.1% 10.5% 9.6%
Diff. +4.2% +5.1% +0.3% +3.3%

Breakdown by Fund

YTD20172016Ann.
Hedgewise1.7%18.2%10.8%12.9%
AQR -0.9% 16.2% 11.2% 10.0%
Invesco 1.3% 10.0% 9.7% 9.1%
Wealthfront -8.0% N/A N/A N/A
Data as of end-of-day on June 5, 2018. Mutual fund data from Morningstar. Wealthfront fund launched in January 2018. Includes all dividends re-invested, costs, and fees. Current strategy model use began in 2016; an older model was used in 2015 and performance was close to even with competitive funds.

Hedgewise has beaten the competition by over 3% annually since 2016. Over a ten year horizon, this would lead to additional total gains of over 70%.

However, the key to establishing whether Hedgewise deserves credit for this outperformance is to examine the shape of the competitive performance curves. It isn't enough to simply generate a higher return; this must be accomplished strictly within the Risk Parity framework. If the performance of one manager deviated too significantly from the rest, it would suggest that some driver besides the core theory - like manager discretion, for example - was playing an outsized role. This would naturally diminish the benefits of the underlying strategy framework, and introduce new risks to the portfolio that have no relation to Risk Parity itself.

Performance Curves Since 2016

Data as of end-of-day on June 5, 2018. Mutual fund data from Morningstar. Wealthfront fund launched in January 2018. Includes all dividends re-invested, costs, and fees.

Hedgewise performance is fairly clustered against the competition over the short-term but with a clear edge that widens over the long-term. This kind of pattern is very close to ideal, as it suggests that the Hedgewise approach is successfully capturing the essence of Risk Parity in a superior way.

This is unsurprising since Hedgewise broadly charges lower fees and incurs fewer expenses on behalf of its clients. However, this approach can go quite badly if a manager oversimplifies too much or fails to invest the resources needed to properly define theoretical concepts. Finding this balance is the key to the success of any smart beta product: it must be simple enough operationally, but still conceptually robust.

The relative performance of Hedgewise over the past few years suggests that it has struck just the right balance and provides a tremendous sense of validation. Let's take a deeper look at the core elements of the approach, and how those differ from the competition.

The Difference in Approach

To minimize operational costs, Hedgewise sought to reduce any kind of complexity that would create little or no net benefit. This raised some very significant theoretical questions, like how much value might be gained from investing globally vs. domestically, or from adding more exotic asset classes to the portfolio mix. Research suggested that so long as you had a very accurate understanding of how to define risk itself, you could successfully run the strategy in one single country and with a relatively basic mix of assets. Yet before Hedgewise was founded, this had never been tested and was vastly different from the approach taken by competitive managers.

With the performance now validated, it seems obvious that these concepts are similar to what gave rise to passive investing in the first place. For example, the idea that you don't need to independently value every stock to still include it in a portfolio, or that holding 1,000 stocks instead of 100 doesn't make much difference. Hedgewise is simply refining similar kinds of concepts as they apply to Risk Parity.

However, unlike strictly passive investments such as the S&P 500, smart beta products tend to have more complex dimensions and more theoretical unknowns. For example, the definition of risk plays an enormous role in the Risk Parity framework, and there is no broad consensus on exactly what 'risk' means nor how to calculate it for different asset classes. Defining risk intelligently requires significant research and expertise, and there are many potential performance pitfalls if this is done poorly.

The new Wealthfront Risk Parity fund provides a useful case study. In its white paper, Wealthfront outlines an approach to defining risk that largely equates it with volatility, or how much an asset tends to move up or down every day. However, a key pitfall to this approach is that risky asset classes like equities often have long periods of low volatility. A volatility-driven framework might misinterpret this to mean that stocks have become 'low risk', and then become more vulnerable whenever risk returns.

This pattern would tend to result in losses especially during periods of elevated, choppy volatility - just like the year-to-date pattern in equities thus far in 2018. Though it can't be determined if this is precisely what has happened with Wealthfront since launching this year, its performance is quite consistent with the theoretical outcome. Note that the performance of AQR and Invesco was clustered closely to Hedgewise and omitted for readability.

Wealthfront vs. Hedgewise Risk Parity Daily Performance, 2018 YTD

Data as of end-of-day on June 5, 2018. Mutual fund data from Morningstar. Wealthfront fund launched on January 29 2018. Includes all dividends re-invested, costs, and fees.

Whether this differential was driven solely by the definition of risk, or some other combination of factors, the most important takeaway is that certain assumptions can have a huge impact on what Risk Parity means and how it performs. Even if the theory itself is entirely valid, different managers will still achieve different results. This is a natural hurdle in the smart beta space, since this makes it harder to separate the strategy from the manager.

Yet these challenges have also provided Hedgewise with the opportunity to demonstrate how powerful the approach can be when it is done well. Risk Parity has tremendous theoretical possibility, but that is diminished if the portfolio is burdened with complexity, expense, or misunderstanding.

Looking Ahead: The Evolution of Smart Beta

The idea of smart beta is still in its relative infancy, but one of the clearest themes to emerge thus far is that the broad ideas can be implemented in dramatically different ways. Some of the lessons from the rise of passive management, like prioritizing simplicity and low-cost, continue to resonate and have formed the basis for much of Hedgewise's outperformance over the past few years. However, it's also obvious that the underlying strategies involve some degree of subjectivity. Over the long-run, any smart beta product will only outperform if there is a real theoretical basis for how it works and a fairly accurate understanding of how to capture the benefit.

It's incredibly exciting to be at a point where there is enough data to identify Hedgewise as a clear leader in Risk Parity, and to see its balanced approach yield exactly the kind of benefits that were predicted.

Disclosure

This information does not constitute investment advice or an offer to invest or to provide management services and is subject to correction, completion and amendment without notice. Hedgewise makes no warranties and is not responsible for your use of this information or for any errors or inaccuracies resulting from your use. Hedgewise may recommend some of the investments mentioned in this article for use in its clients' portfolios. Past performance is no indicator or guarantee of future results. Investing involves risk, including the risk of loss. All performance data shown prior to the inception of each Hedgewise framework (Risk Parity in October 2014, Momentum in November 2016) is based on a hypothetical model and there is no guarantee that such performance could have been achieved in a live portfolio, which would have been affected by material factors including market liquidity, bid-ask spreads, intraday price fluctuations, instrument availability, and interest rates. Model performance data is based on publicly available index or asset price information and all dividend or coupon payments are included and assumed to be reinvested monthly. Hedgewise products have substantially different levels of volatility and exposure to separate risk factors, such as commodity prices and the use of leverage via derivatives, compared to traditional benchmarks like the S&P 500. Any comparisons to benchmarks are provided as a generic baseline for a long-term investment portfolio and do not suggest that Hedgewise products will exhibit similar characteristics. When live client data is shown, it includes all fees, commissions, and other expenses incurred during management. Only performance figures from the earliest live client accounts available or from a composite average of all client accounts are used. Other accounts managed by Hedgewise will have performed slightly differently than the numbers shown for a variety of reasons, though all accounts are managed according to the same underlying strategy model. Hedgewise relies on sophisticated algorithms which present technological risk, including data availability, system uptime and speed, coding errors, and reliance on third party vendors.

Can You Time Risk-Managed Strategies?
Posted in Investment Strategy on 2018-04-17

Summary

  • Many clients wonder whether they should adjust their approach depending on market conditions
  • However, risk-managed strategies are constantly responding to the current environment, such that attempts at timing are usually counterproductive
  • After a quick review of why this is consistent with the underlying theory, I'll analyze a few of the most common "timing" questions to see if they have any merit:
    • Should I wait to put cash to work until after a period of large losses?
    • Should I take cash off the table after a period of large gains?
    • Should I invest elsewhere if I think an equity or bond bear market is approaching?

Risk Management versus Timing

If inflation, trade wars, data leaks, slowing global growth, or government instability has given you pause on the investment outlook, you are certainly not alone. It's hard not to wish that you had just sold everything in January, or not to wonder whether you should still sell everything now. The good news is that Hedgewise frameworks have already shifted to account for these market conditions, and the future outlook remains excellent. The bad news is that you will still likely wonder whether you can time it better yourself, but fortunately we can look to the theory and data for guidance.

While Hedgewise runs two different risk-managed frameworks - Risk Parity and Momentum - they are rooted in the same core financial principles:

  • Investors generally expect a positive return on their investments, so in normal environments, markets appreciate over time
  • Diversification across different kinds of assets, like stocks and bonds, reduces risk by offsetting short-term gains and losses
  • As market risk goes up, so does the probability of large gains or losses

Hedgewise then uses financial engineering, like leverage and risk balancing, to most effectively apply these concepts to its products.

Note that none of these principles require you to figure out whether assets are over- or under-valued, which tends to be incredibly difficult. This kind of "timing" usually backfires because it requires that you know the exact right time to enter or exit. One of the main reasons that risk management is effective is because it avoids the need for such precision.

When relying on the relatively simple principles above, the focus shifts away from short-term returns and towards long-term stability and loss reduction. When this is done intelligently, it means your portfolio is constantly maximizing the odds in your favor. Positive returns are always more likely than not, even in the worst of market conditions.

Clients often find this counterintuitive because 1) it seems impossible that return expectations would be positive if markets are about to collapse, and 2) risk-managed strategies still undergo periods of loss, which in theory might be timed. However, because these periods of loss are driven by the relationships between different assets and their volatilities, they often have little to do with bear markets. For example, both Risk Parity and Momentum wound up with calendar year gains during four of the last five major stock crashes.

Put another way, to time a risk-managed strategy, you'd need to be able to predict when fundamental asset relationships and risk indicators are about to breakdown in conjunction with a major market crash - a tall order indeed! These kinds of events tend to be quite sudden, random, and short-lived by definition.

Now, it is true that the probabilities of gain or loss do shift in certain market conditions - for example, you have a better than normal chance that risk-managed strategies will do well after a period of loss. But in waiting around for a loss to happen, you'll probably miss substantial gains in the meantime. No matter which way you cut the scenario, the same theme arises: it's tremendously difficult to beat the simple approach of buy-and-hold.

Scenario 1: Waiting for a Loss

It's far easier to invest with confidence when assets look cheap, and many prefer to wait on the sidelines until after some kind of crash. This is always a tricky topic because, in hindsight, it's true that you make more money if you buy low, and cash also carries great comfort for the risk averse. The problem is that risk-managed strategies are hedged across many different kinds of assets, so a crash in a single market often doesn't result in net losses to the portfolio.

As an illustration, here are the rolling 1yr returns since 1972 for Risk Parity High, Momentum High, and a 50/50 split between the two.

1yr Rolling Returns By Strategy Since 1972

Data based on hypothetical models using end-of-day index data since 1972. All dividends are included and assumed re-invested. Includes an estimate for Hedgewise fees of 0.7%. See full disclosures at end of article.

Depending on your strategy mix, you've historically had about a 6-10% chance of incurring a loss over the following 12 months, or conversely a 90-94% chance of incurring a gain. Most of these losses were also fairly minimal; there was a less than 2% chance of incurring a loss of 10% or more.

That said, you may still figure that waiting for one of these periods still bumps your odds up further, and you'd be right: if you happen to start after a twelve-month loss, your historical odds of a subsequent gain go up to over 98%. But the question is not whether the odds improve, but rather if it makes sense to wait around for that to happen.

A simple way to answer this question is to calculate a "breakeven point" in time. That is, the date where the gains you would have accrued before the subsequent drawdown were greater than the drawdown itself. This assumes that you also had the magical ability to invest right at the exact bottom of each pullback. Performance is drawn from the 50/50 split portfolio (though RP/MM alone show similar numbers).

If you got in at the exact bottom in.. You'd still be worse off than if you started..
Jun 19787 months earlier
Sept 198117 months earlier
Jun 19848 months earlier
Aug 198823 months earlier
Sept 19905 months earlier
Nov 199417 months earlier
Aug 20012 months earlier
Jan 20164 months earlier
Data based on hypothetical models using end-of-day index data since 1972. All dividends are included and assumed re-invested. See full disclosures at end of article.

On average, you lost about 10 months' worth of gains. In other words, to successfully take advantage of an upcoming period of losses, you need to a) correctly identify the 10% chance that losses are going to occur at all, b) be pretty sure it's not more than 10 months away, and c) know exactly when the drawdown hits bottom. If you fail on any single one of these conditions, you'd be better off investing now rather than waiting.

That said, there are understandable exceptions: you may keep a discretionary pool meant for more opportunistic investing, you may have new savings become available at a distinct point in a year, or you may be quite confident in your ability to foresee an upcoming loss. In these cases, you'd like to know when the probabilities for a near-term gain are the most favorable.

One way to study this is to compare the size of any current drawdown in the strategy framework to the subsequent returns. Performance is drawn from the 50/50 split portfolio, but RP/MM alone again show similar numbers.

Forward Return of the 50/50 Split Portfolio for Various Timeframes, by Size of Drawdown

Data based on hypothetical models using end-of-day index data since 1972. All dividends are included and assumed re-invested. See full disclosures at end of article.

Interestingly, even drawdowns as small as 3% substantially increased the size of future returns. While it might seem tempting to wait and capitalize on the bigger losses, they are extremely infrequent: the last 15% drawdown happened in 1988. A more reasonable target would be anywhere in the 5-10% range, which you'll usually see every other year or so. But remember, if you sit in cash for more than a couple of months, you'll probably miss more gains than even a successful timing attempt will recoup.

Applying this analysis to today, we experienced a drawdown in the 50/50 strategy of around 10% from January 26th through February 8th. Historically speaking, there's around a 95% chance we are already past the bottom. While there's been some recovery since then - the current drawdown is now more like 7% - it is still quite an attractive time to invest if you happen to be sitting on cash.

Scenario 2: Selling After Gains

Looking back, it seems obvious that the initial fast gains in January were signs of overheating, and to wonder whether there was a way to identify that beforehand. However, nothing in the underlying theory suggests that gains of a particular size or speed should be worrisome. After all, we expect gains to happen 90% of the time, and the techniques of hedging and risk management are always limiting the impact of an individual asset bubble or crash. To test this, we can compare short-term gains in the 50/50 strategy to subsequent historical returns. To further isolate "overheating" scenarios, this data is also limited to months in which there was no recent drawdown.

Forward Return of the 50/50 Split Portfolio for Various Timeframes, by Size of Prior 1 Month Gain

Data based on hypothetical models using end-of-day index data since 1972. All dividends are included and assumed re-invested. See full disclosures at end of article.

This data shows no indication that large gains typically precede losses. In fact, quite the opposite: forward-looking returns often increase instead! Digging into the numbers, it is unsurprising to learn that you see clusters of great returns in the middle of broad, calm bull markets - years like 1997, 2006, or 2017. In these periods, when you had a fantastic one-month gain, you typically went on to have many more of the same. This same pattern shows up using prior three-month, six-month, and one-year gains as well; there's simply nothing to suggest that big positive returns frequently reverse.

If a risk-managed strategy is implemented well, this is what you'd expect. While individual asset classes like equities may become "irrationally exuberant", such risks are explicitly built into the frameworks and minimized. Though occasional drawdowns are inevitable, they tend to be quite random, and certainly have nothing to do with recent performance trends.

Scenario 3: Timing Bear and Bull Markets

The final timing question that many clients wonder is: what if it is just a bad time to be invested in general? For example, if you knew that bonds and/or stocks were going to do poorly for the next year, wouldn't you be better off exiting?

Even if you had the ability to correctly forecast an upcoming downturn, that wouldn't necessarily mean that risk-managed strategies make a bad investment. Returning to the theory, different asset classes will perform differently depending on the underlying economic environment. For example, in a recession, gold and bonds will tend to rally while stocks will tend to fall. So long as these relationships hold up as expected, risk-managed strategies should be quite resilient against individual asset crashes.

To test this, we can examine the performance of the 50/50 split portfolio in the worst-performing stretches for both stocks and bonds. Starting with equities, the following table displays tranches of every rolling one-year period of losses in the S&P 500 since 1972 and summarizes how the 50/50 split portfolio performed over the same periods. Note that the "% Gain" and "% Loss" columns display the number of data points when the 50/50 portfolio had either a gain or loss out of the total number of data points within tranche.

Rolling 1yr Performance of the 50/50 Split Portfolio During Equity Drawdowns

50/50 Split Performance
S&P 500 1yr Loss Avg.% Gain% Loss
0-5% Loss +5.4% 67.7% 32.3%
5-10% Loss +4.3% 69.6% 30.4%
10-15% Loss +7.6% 87.5% 12.5%
15-20% Loss +9.8% 90.9% 9.1%
20-30% Loss +8.7% 94.1% 5.9%
Over 30% +14% 88.9% 11.1%
Data based on hypothetical models using end-of-day index data since 1972. All dividends are included and assumed re-invested. See full disclosures at end of article.

In every single tranche, you averaged gains in the 50/50 split despite losses in the S&P 500. Your odds of a gain also significantly increased along with the size of the equity pullback. This is because risk-managed strategies tend to minimize equity exposure as losses increase, while safe-haven assets like bonds and gold rally. In fact, the strategies tend to be most vulnerable to "small" losses (under 10% or so) that occur when the market is still trying to "figure things out". For example, in our most recent pullback, bonds and gold have not rallied much despite the pullback in equities because it is not yet certain that a recession is imminent. Of course, these short periods are all the more difficult to time!

Now let's repeat these same numbers for bonds to make sure the story is consistent. Note that the loss ranges had to be reduced compared to equities as bonds are far more stable.

Rolling 1yr Performance of the 50/50 Split Portfolio During Bond Drawdowns

50/50 Split Performance
10yr Treasury 1yr Loss Avg.% Gain% Loss
0-1% Loss +10.2% 87.5% 12.5%
1-2% Loss +13.5% 88.2% 11.8%
2-3% Loss +16% 93.3% 6.7%
3-4% Loss +14.1% 91.7% 8.3%
4-5% Loss +12.9% 83.3% 16.7%
Over 5% +9.2% 74.3% 25.7%
Data based on hypothetical models using end-of-day index data since 1972. All dividends are included and assumed re-invested. See full disclosures at end of article.

Bonds exhibit a similar story, though you do see slightly lower (albeit positive) returns for losses of over 4%. Many of the data points in this upper range come from periods of high inflation, during which both stocks and bonds tend to do poorly while commodities provide a hedge. This presents less opportunity for net upside as commodities are naturally a smaller portion of the portfolios. At worst, though, you averaged a 9.2% annual return and a 75% chance of gains.

Summing up the numbers, it really didn't make sense to avoid risk-managed strategies even if you had perfect insight about upcoming equity and bond pullbacks. Your worst-case returns were around 4-5% during smaller equity pullbacks, while in all other cases you achieved returns of 8% or more. Those are pretty stellar numbers in years of very rocky markets.

Conclusion: Staying the Course, Expecting a Few Bumps

Put simply, risk-managed strategies are really effective at dealing with worst-case scenarios. They assume that bubbles and crashes are part of the norm, but since this logic is already built-in, you can't apply traditional thinking like how to "get out at the top" or "get in at the bottom". Neither a stock crash nor a bond crash will necessarily result in losses; more often, the portfolio is vulnerable to quick pullbacks or markets that have a great deal of uncertainty. Even if you could time these situations, the losses are usually pretty small and not worth the headache of figuring out when to re-enter. If you try to time unsuccessfully, the gains you miss quickly outweigh any potential benefits.

None of this is to suggest that these frameworks are invulnerable, but rather that the probabilities are enormously in your favor if you invest early and stay patient. You will certainly have a few situations where this patience is tried: you may underperform the S&P 500 for stretches of time, hedges will fail to work in certain situations, and markets will sometimes be blindsided by the unpredictable. Despite all of that, there is an extraordinarily high chance that you will go on to perform wonderfully if you simply shrug your shoulders.

For discretionary pools, you've generally found good entry points during drawdowns of 5-10% or during periods of consistently large gains, regardless of surrounding market conditions. That said, if you don't currently have better uses for your discretionary funds, or if in reality you are just trying to optimize the entry point for your broader portfolio, the best path is almost certainly the simplest: invest now at your long-term risk level, and don't worry about the timing.

Disclosure

This information does not constitute investment advice or an offer to invest or to provide management services and is subject to correction, completion and amendment without notice. Hedgewise makes no warranties and is not responsible for your use of this information or for any errors or inaccuracies resulting from your use. Hedgewise may recommend some of the investments mentioned in this article for use in its clients' portfolios. Past performance is no indicator or guarantee of future results. Investing involves risk, including the risk of loss. All performance data shown prior to the inception of each Hedgewise framework (Risk Parity in October 2014, Momentum in November 2016) is based on a hypothetical model and there is no guarantee that such performance could have been achieved in a live portfolio, which would have been affected by material factors including market liquidity, bid-ask spreads, intraday price fluctuations, instrument availability, and interest rates. Model performance data is based on publicly available index or asset price information and all dividend or coupon payments are included and assumed to be reinvested monthly. Hedgewise products have substantially different levels of volatility and exposure to separate risk factors, such as commodity prices and the use of leverage via derivatives, compared to traditional benchmarks like the S&P 500. Any comparisons to benchmarks are provided as a generic baseline for a long-term investment portfolio and do not suggest that Hedgewise products will exhibit similar characteristics. When live client data is shown, it includes all fees, commissions, and other expenses incurred during management. Only performance figures from the earliest live client accounts available or from a composite average of all client accounts are used. Other accounts managed by Hedgewise will have performed slightly differently than the numbers shown for a variety of reasons, though all accounts are managed according to the same underlying strategy model. Hedgewise relies on sophisticated algorithms which present technological risk, including data availability, system uptime and speed, coding errors, and reliance on third party vendors.

February 2018: Why Not to Panic When Markets Go Crazy
Posted in Market Commentary on 2018-02-12

Summary

  • Hedgewise has incurred losses of 5-10% since late January, depending on your product and risk level, but is only down slightly year-to-date and continues to outperform all major competitors.
  • Much of the drawdown has been driven by simultaneous losses across all asset classes, which strongly suggests investor panic and confusion. Such scenarios have never lasted long historically and will likely soon reverse.
  • Even if some of the worst-case scenarios come true, like stronger than expected inflation or a recession, both Hedgewise frameworks have held up well in such environments.

Stay Calm and Carry On: Putting Recent Losses in Perspective

Make no mistake: markets have been pretty wild for the past couple of weeks, and if it's started to make you nervous, you are human after all! It has been especially confusing because the swings are quite hard to explain: not all that much has changed in the economy since January, yet markets are suddenly terrified of inflation, government debt, volatility, and valuations. If you can't explain why people are selling now, it's also hard to predict when they will stop.

Since every investor on the planet has this same logic and fear, it's easy to see how it can all quickly turn into a frenzy. And yet, this story also justifies why short-term market volatility shouldn't worry you much at all. If people are panicking for no good reason, you can be almost certain that they are selling assets too cheaply, and that's really the worst possible time to change your approach.

It helps to return to the basics of investment theory, which I discussed in my previous newsletter. Recall that your expected returns should look something like the following, with the blue line being your realized month-to-month returns, and the orange being the underlying "risk premia" - or "fair value" - that you are accumulating over time.

If you look at the past two weeks or so, we've most likely just experienced a very rapid cycle of this diagram, with assets moving temporarily above their fair value and now back below. The reason this is not particularly concerning is that it has no effect on your expected return over time, so long as you simply wait. By focusing primarily on long-term returns, you also minimize the many pitfalls of short-term timing and active management.

Now, Hedgewise still applies various kinds of risk management, but it is all with this long-term focus. For example, balancing exposures across many different assets, like stocks and bonds, tends to minimize the impact of a crash in any single one. But in the span of a few days or weeks when investors are panicking, it is possible they will all move down together. Likewise, there are certain extreme risk environments, like recessions and hyperinflation, that can sometimes be detected beforehand. But short-term market swings most often have very little to do with the economy at all.

With this perspective, the Hedgewise frameworks have continued to be quite effective. For example, since the beginning of 2018, the Hedgewise Risk Parity framework has lost significantly less than comparable major mutual funds. This continues a clear trend of outperformance ever since Hedgewise was launched. Last year, the Hedgewise Risk Parity and Momentum products both significantly outperformed the S&P 500 at the Max risk level, yet neither has lost significantly more than the S&P 500 so far this year.

Periods like these past two weeks will always be uncomfortable, but short-term losses are very different than long-term risk. To further make this case, let's take a deeper look at recent performance trends and how they stack up against history.

2018 Year-to-Date Performance: Unavoidable Losses, But Better Than the Competition

While most of the news is focused on stock returns since the peak on January 26th, equities were up almost 8% before they gave it all back. Trying to make sense of this fast of a reversal doesn't serve much purpose. The more interesting story is how various asset classes have performed year-to-date overall:

2018 Year-to-Date Performance By Asset Class

Hedgewise data based on various end-of-day index prices and include an estimate for all dividends. Data as of Feb 9th, 2018.

The bond market has actually been in a more significant correction than equities, as long-term yields have jumped about 0.7% since last September and 0.5% in the past two months alone. This makes some sense, given the Fed has started to more rapidly raise rates and reverse the "Quantitative Easing" program, and Hedgewise risk indicators have been frequently spiking as a result, including last month. The effectiveness of this dynamic risk management can be most easily seen by comparing the performance of the major Risk Parity mutual funds.

Performance of Hedgewise RP High vs Major Risk Parity Mutual Funds, 2018 Year-to-Date

Data based on publicly available quotes for AQRNX and ABRYX and include an estimate for all dividends. Hedgewise data is an average of all client performance in the RP High product and includes all costs and fees.

The graph continues to demonstrate a high correlation between the various risk parity products, since they are all investing in the same broad asset classes. The main difference is in how risk is balanced, and Hedgewise has consistently achieved a superior level of performance in the short and long-term, as demonstrated by its comparative performance back through the beginning of 2017.

Performance of Hedgewise RP High vs Major Risk Parity Mutual Funds, 2017 to Current

Data based on publicly available quotes for AQRNX and ABRYX and include an estimate for all dividends. Hedgewise data is an average of all client performance in the RP High product and includes all costs and fees.

While the relative performance is excellent, why hasn't any Risk Parity framework been able to better hedge this equity correction? If you glance back at the year-to-date performance across asset classes, you'll notice that bonds, commodities, and stocks have all incurred losses simultaneously. In such an environment, there's really no way to avoid a loss unless you engage in very short-term timing (quick reminder: all active managers do some form of this, and over 90% of them underperform the S&P 500).

This kind of cross-asset correlation is somewhat exceptional, especially during a 10% equity correction, though not entirely unprecedented. Since 1970, there's been exactly four other scenarios where equities have lost 8% or more while safe havens like gold and bonds also suffered losses.

DateEvent
Jul 1974Beginning of Stagflation
Dec 1980End of Stagflation; Interest rates peak near 20%
Oct 2008Beginning of Great Recession
Mar 2009Great Recession Market Bottom

While these are some pretty scary events, the good news is that Hedgewise frameworks still did fine in all of these scenarios because safe havens eventually kicked in. Let's take a deeper look at how it unfolded.

When Safe Havens Fail: Why It Happens and What It Means

There are only two reasons that investors sell stocks, bonds, and commodities at the same time: either they are in full panic, or they are really confused about inflation. The Great Recession was a great example of 'sell everything' when Lehman went bankrupt. People just moved to cash in a mix of confusion and a need for liquidity. Gold was the logical hedge against a failing financial system, and it went on to rally by 30% by February 2009, but it often won't hold up at the outset.

Stagflation is the other culprit, since it means poor economic growth due to runaway inflation. Both stocks and bonds will lose money by definition (since higher inflation means higher interest rates). While real assets like commodities should do well since the dollar is losing its value, there's often an initial fear that the Fed will pre-emptively raise rates to fight inflation even if it will result in a recession. Ironically, a recession would then mean lower interest rates, so then bonds would actually rally, but you can see how everyone basically gets scared and confused!

In each of these scenarios, equities, bonds, and gold all fell together for a couple of weeks. To see how it eventually played out, though, I looked at the full one year return for each asset class following each event.

One Year Return by Asset Class After Initial Event

Date Stocks Gold Bonds
Jul 1974 15.2% 16.4% 5.6%
Dec 1980 -2.8% -36.6% 11.6%
Oct 2008 -9.2% 14.1% 7.7%
Mar 2009 62.3% 18.8% -2.7%
Data based on publicly available end-of-day index prices and include an estimate for all dividends assumed re-invested.

What's really neat is that you can see all of the different possibilities unfold. While these were all pretty awful economic times, at least one of the asset classes eventually rallied. It's the laws of economics at work.

It's also helpful to look at how the Hedgewise models did over this same one year period after each initial event.

One Year Return by Hedgewise Model After Initial Event

Date RP Max Momentum Max
Jul 1974 9.8% 14.3%
Dec 1980 -4% 41.5%
Oct 2008 20.4% 6.7%
Mar 2009 22.1% 38.2%
Hedgewise models based on hypothetical simulations using end-of-day index prices and assume all dividends are re-invested. See full disclosure at end of article.

The big idea is that markets work themselves out over time, and that is eventually captured in the Hedgewise frameworks. You can still wind up with some bad years, and performance won't always bounce right back. But the odds are always on your side, perhaps even more so right after the scariest kind of market behavior.

Wrapping Up

While hopefully it's clear that there's no need to panic, whether losses have already bottomed or not, these past couple of weeks are still a wonderful opportunity to reflect on your own goals and risk tolerance. Investing presents a natural conflict between logic and fear. Rationally, targeting a higher long-term return seems like the right choice for many clients, despite the warning that a 20-30% loss is basically inevitable at least once a decade (at the higher risk levels). Yet it can feel quite different in the past couple of weeks when you see 10% disappear, and that's not even a particularly severe event!

On the other hand, the RP Max and Momentum Max product returned about 27% and 31%, respectively, in 2017 alone. Even if losses continue and this is the year of a 30% drawdown, you'd only be slightly worse off than if you had held cash the past year. It all depends on your perspective. Whether next month or next year, the drawdown will eventually come. If you remain patient and calm, it will also almost certainly pass. Either way, this most recent experience should help prepare and inform you.

Disclosure

This information does not constitute investment advice or an offer to invest or to provide management services and is subject to correction, completion and amendment without notice. Hedgewise makes no warranties and is not responsible for your use of this information or for any errors or inaccuracies resulting from your use. Hedgewise may recommend some of the investments mentioned in this article for use in its clients' portfolios. Past performance is no indicator or guarantee of future results. Investing involves risk, including the risk of loss. All performance data shown prior to the inception of each Hedgewise framework (Risk Parity in October 2014, Momentum in November 2016) is based on a hypothetical model and there is no guarantee that such performance could have been achieved in a live portfolio, which would have been affected by material factors including market liquidity, bid-ask spreads, intraday price fluctuations, instrument availability, and interest rates. Model performance data is based on publicly available index or asset price information and all dividend or coupon payments are included and assumed to be reinvested monthly. Hedgewise products have substantially different levels of volatility and exposure to separate risk factors, such as commodity prices and the use of leverage via derivatives, compared to traditional benchmarks like the S&P 500. Any comparisons to benchmarks are provided as a generic baseline for a long-term investment portfolio and do not suggest that Hedgewise products will exhibit similar characteristics. When live client data is shown, it includes all fees, commissions, and other expenses incurred during management. Only performance figures from the earliest live client accounts available or from a composite average of all client accounts are used. Other accounts managed by Hedgewise will have performed slightly differently than the numbers shown for a variety of reasons, though all accounts are managed according to the same underlying strategy model. Hedgewise relies on sophisticated algorithms which present technological risk, including data availability, system uptime and speed, coding errors, and reliance on third party vendors.

Related Posts

The Anatomy of Momentum: Why the Strategy Works, and How to Play the Odds2018 Year-In-Review: This Too Shall PassOctober 2018: Better Times AheadSeptember 2018: Amidst Volatile Markets, Hedgewise Leads the Pack Best-In-Class Risk Parity PerformanceCan You Time Risk-Managed Strategies? February 2018: Why Not to Panic When Markets Go CrazyHow to Link an Existing Interactive Brokers AccountHow to Link an Existing Interactive Brokers Account - New AMHow to Link an Existing Interactive Brokers Account - Classic AMAnalyzing Hedgewise 2017 Performance: Benchmarks, Timeframes, Theory and ProofQ3 Update: Hedgewise Outperforming Every Asset Class in 2017Understanding the Theory Behind Better ReturnsApril 2017: A Great Start to the Year2016 Year In Review: Hedgewise OutperformedHedgewise Systematically Avoids Bond CorrectionNovember Commentary: Election Risk Is Just Like Any Other Risk, And It Is Being ManagedHow to Open An AccountHow to Create Leverage Using Options ContractsHedgewise Outperforming Every Major Risk Parity Mutual Fund in 2016Risk Parity Just Got Even BetterWhat's Next for Risk Parity? 2016 Hedgewise Midyear Report: Choppy Markets, Big ReturnsThe Right Way To Invest In Oil Rising Interest Rates And Risk ParityThe Right Time to Buy Oil in 2016The Wrong Way To Invest In OilRisk Parity Outperforming the S&P 500 by 7% in 2016Risk Parity: Year-In-Review and 2016 OutlookWho We AreHow It Works: Faster, Cheaper, and More EfficientOur CustodianFeesWhy Commodities are the Smart Play for 2016Retirement Investing in a Rising Interest Rate EnvironmentComparing Hedgewise Risk Parity to the Competition Improvements to our Risk ModelSeptember 2015: Risk Parity Limits Losses, and the Upside of FearAugust 2015: Perspective on the Commodity CrashIs Gasoline the Smart Oil Play? (UGA vs. USO) 2015 Mid-Year Review: Understanding Unbalanced MarketsNavigating the Bond Market CorrectionFeeling Good About A Down MarketNew Account Set-Up - InvitationThe Optimal Gold Investment Strategy (Switching Between DGL and GLD)How Hedgewise Saved You Taxes in 2014This Is Why You Are Still Diversifying WrongThis Is Why You Are Diversifying WrongRisk Parity: What It Is, How It Works, and Why It MattersThe Financial Revolution Has Begun: 5 Bold Predictions for the Next DecadeMust Bond Investors Fear Rising Interest Rates? Insights From 1958 To 1982 The Oil Futures Curve Reversal: What You Need to Know5 Reasons Why You Should Be Afraid of a Bear Market, and How to Protect Your PortfolioHow to Invest in Oil for the Long Term, Avoiding Contango and Tracking ErrorTax Optimization and Tax HarvestingHedgewise FAQ