“What goes up when stocks go down?” is one of the most important questions an investor will have to find the answer to.
But the answer isn’t a simple one.
Diversifying away from pure equity exposure is critical for those who don’t have the willingness or ability to invest in an equity-only portfolio. We’ve seen the reasons why diversification is so important in the posts: ‘What is diversification and why do I need it?’ and ‘Smooth is what we aim for’.
But figuring out how to diversify is a much trickier question.
The purpose of this post is to examine each of the asset classes which are commonly used as so-called “portfolio insurance” (they insure against equity market crashes), and to figure out which diversifiers seem like sensible options to include in a portfolio.
This post has taken me a while to put together, simply because of the weight of evidence for and against each option.
I’ve distilled the pros/cons for each option at the start of each section, followed by the evidence for each. I’ve summarised each piece of research into key charts and key quotes. To wrap it all together, I’ve put my thoughts in the conclusion.
I’m intending for this post (along with many of the other evidence-heavy posts on this blog) to be a live document, which I’ll come back and revisit as I come across more research relevant to the topic.
Without further ado, let’s take each of the following options in turn:
The crash protectors
- Long vol
- Equity factor strategies – trend and quality
- Government bonds
A couple of my own charts
- Strong positive performance during previous market crashes
- Not reliant on asset class correlations
- Difficult to invest in for retail investors
- Negative long-term returns
- Protection is not guaranteed
- Drawdowns rarely coincide with option expiration cycles
- Difficult to stick with
- Relatively expensive TERs for the few accessible funds
- Better results can be achieved through diversification
Puts: The Evidence
AQR’s paper: ‘It Was the Worst of Times: Diversification During a Century of Drawdowns’
‘As with everything in investing, there is no perfect solution to addressing the risk of large equity market drawdowns. However, we find using nearly a century of data that diversification is probably (still) investors’ best bet’.
AQR’s paper: ‘Pathetic Protection: The Elusive Benefits of Protective Puts’
‘The divested portfolio nearly universally has better drawdown characteristics than does the protected portfolio. For example, the worst 1% peak-to-trough drawdowns over a 20-day period are −9.6% for the protected portfolio versus −6.6% for the divested portfolio. Arguably, investors should be more concerned about longer-term drawdowns. Over 250-day windows, the results are even worse for protection: −32.1% for the protected portfolio and −20.9% for the divested portfolio.’
‘Many of us naturally expect that those who are willing to pay the cost will obtain meaningful downside protection. I find that this simply is not so. It is not safe to assume that a protective put will protect your portfolio against large drawdowns.
For those who are concerned about their equity’s downside risk, reducing their equity position is significantly more effective than buying protection. Sized to achieve the same average return, divesting has lower drawdowns, lower volatility, lower equity beta, and a higher Sharpe ratio than does buying put options. The one case where buying put options shines relative to holding a reduced equity position, even if options are priced to include volatility risk premium, is when a very large crash occurs prior to the options’ expiration.
The results are clear. Buying protection more often than not and on average leads to worse drawdowns than does divesting the equity position to match the average return.’
Commentary on AQR’s ‘Pathetic Protection’ paper by the CFA Institute
‘The surprising result that put options often do more harm than good comes about because drawdowns rarely coincide with option expiration cycles. The longer a put option’s maturity, the more likely it is that a stock can rise far above the strike of an out-of-the-money put and then crash. Even though the stock has declined appreciably, the put does not hedge the decline.
The key takeaway is that put options are effective in reducing drawdowns only in the rare circumstance when options are priced with no volatility risk premium and equity drawdowns precisely coincide with the option holding period.’
Man Group’s paper: ‘The Best of Strategies for the Worst of Times: Can Portfolios be Crisis Proofed?’
‘The most reliable defensive tool, continuously holding short-dated S&P 500 put options, is also the costliest strategy (-7.4% return over all periods). While it performs well during crashes (earning a 42.4% return on average), it is very costly during the “normal” times (losing 14.2% on average), which constitute 86% of the sample, and expansionary (nonrecession) times, which constitute 93% of the observations. As such, passive option protection seems too expensive to be a viable crisis hedge. They noted that options are also expensive to trade, and thus returns would have been even worse after implementation costs.’
‘Once a crisis has begun, it can be observed that subsequent rolls of the options become more expensive as implied volatility rises, increasing the cost of the hedge. This effect then requires accelerated price movement to produce the same hedge return.’
NB: This table is from a slightly modified version of the original paper, and appears on the Man Group’s website.
AQR’s paper: ‘Tail Risk Hedging: Contrasting Put and Trend Strategies’
Unlike Trend, long Put strategies have had persistently negative returns despite, or perhaps because of, their gains during crashes (cost for a valuable insurance service). This jibes with the balance of economic theory, which would suggest a negative risk premium for insurance-like strategies. Yet, the documented return drag may be mitigated if the tail hedge allows an investor to take more equity risk and earn a premium for it, or if active tail hedge managers can offer “alpha” over Put.
Both Put and Trend strategies have good hit rates in most equity market drawdowns, with important differences. Put strategies have offered more reliable tail insurance in fast market drawdowns, providing more “bang-for-the-buck” than Trend in gapping markets like October 1987. That said, Trend does tend to make money in the largest equity market monthly drops as they typically do not occur out of the blue. Moreover, Trend is better suited to slower, protracted bear markets, for obvious reasons. In these scenarios, Put is hampered by more sensitivity to the exact path of negative returns (e.g., a slow drawdown in which puts continuously expire out-of-the-money) and the general negative return being relevant over longer horizons.
Ultimately, the long-term cost argument tips the scales in favor of Trend. This view is reinforced by the inevitable investor impatience during the dry spells when tail insurance costs are paid year after year before the tail event materializes. A good strategy is one that you can stick with; Put-based tail hedging too often fails this test. Trend strategies do not offer as direct or explicit tail protection, but they have a strong empirical record, and you have a better chance of sticking with Trend.
Verdad Capital’s research email: ‘Tail Risk Hedging’
‘Tail risk funds are merely one extreme point in a broader menu of options. Investing solely in tail risk funds forgoes the benefit of diversification. As shown below, a simple equal-weighted basket of the above hedges would have significantly outperformed tail risk hedge funds alone.’
‘The basket of tail hedges would have offered a similar quality of hedging at lower cost. The basket approach also offers more opportunities for alpha generation by rotating into the cheapest hedge when relative valuations diverge.
Adjusting for the favorable endpoint of this historical sample leads to an estimate of the long-term cost of hedging of around 2% per year, a steep cost given an expected 3-5% equity risk premium. Even after cheapening the cost of hedging through diversification, the cost of insurance is still very high.
In some sense, tail risk hedging is just asset allocation on steroids: investors should think through the costs and the full range of options rather than sticking only to the most popular and hence expensive hedges like VIX futures or tail risk hedge funds.’
- Gold has performed reasonably well as crash insurance in the past. It might not have always gone up when equity markets went down, but it almost always fell less than the market.
- Easily accessible for retail investors through ETFs.
- Cheap to gain access to when using ETFs.
- Lower opportunity cost for holding gold when interest rates are low.
- Gold has been a poor shorter-term inflation hedge, having suffered an 80% real drawdown between 1980 and 2000.
- Gold has also done a poor job of protecting against unexpected inflation.
- Requires a relatively large allocation before its crash protection and uncorrelated returns have a meaningful impact on a portfolio
- May be difficult for investors to stick with, given large real drawdowns and unclear drivers of demand
Gold: The Evidence
My previous post: ‘Should you invest in gold as a UK investor?’
- Well documented negative correlation between volatility of equity prices and stock returns
- Volatility exposure should protect portfolios when markets crash
- The value of protection during times of market stress is quickly negated due to the roll costs associated with the rebalancing strategy of the VIX
- Very negative long-term returns
- VIX ETFs:
- Do a poor job of replicating the VIX index
- Are designed for holding periods of days, not years
- Are far more volatile than both stocks and bonds
- The combination of negative long-term returns and high volatility makes VIX ETFs difficult to stick with
- Limited investment options for a retail investor
- Relatively expensive TERs (c. 0.90%)
Long Volatility: The Evidence
Verdad Capital’s research email: ‘Tail Risk Hedging’
‘Long volatility was helpful on average, with the current crisis a standout winner. Also, as discussed earlier, the long-term cost of buying options is much higher than other asset classes. For instance, the VIX future return has been a whopping -50% per year since inception in 2006.’
Larry Swedroe’s commentary on two volatility papers
‘The apparent ex-ante benefits have been more than offset by the high roll costs (implied volatility has been greater than realized volatility) for these products.’
‘VIX ETPs have very poor average returns which are more negative than reported in previous studies. This is not a surprise as the VIX futures term structure has mainly been in contango.
For short-sales constrained (long-only) investors, the value of protection during times of market stress is quickly vaporized due to the roll costs associated with the rebalancing strategy of the VIX.
The standard deviations indicate that the VIX ETPs are far more volatile than both stocks and bonds.
Except for one period (2009-2010), the economic value of diversifying with VIX ETPs is negative.
Even accounting for the simulated market crash is not enough for the VIX ETPs to add economic value to the investor.
An investor who wants to insure against corrections in the equity market faces the trade-off between paying the higher premiums of shorter-dated products and then getting larger payouts in times of market distress versus paying lower premiums but also getting lower potential payouts from the longer-dated products.
They cited the specific example of a long position in one ETP (VXX) which, because of the roll cost, from its inception date until September 14, 2018 would have lost 99.9% of the initial investment. The longer dated product (VXY) would have lost less as roll costs are lower on longer-dated products. The authors concluded that with their return profile, “it seems obvious that these VIX ETPs are not suitable for buy-and-hold strategies which is also stated in their prospectus.”’
A good brief overview of VIX ETFs by Fidelity:
‘By any measure, VIX futures indexes—and therefore VIX ETFs—do a lousy job emulating the VIX index. The VIX index is truly uninvestable, and over periods of a month or a year, the return pattern of VIX ETFs will differ radically from that of the VIX index.
VIX ETFs are at the mercy of the VIX futures curve, which they rely upon for their exposure. Because the typical state of the curve is upsloping (in contango), VIX ETFs see their positions decay over time. Decay in their exposure leaves them with less money to roll into the next futures contract when the current one expires. As time goes on, this process repeats itself multiple times, and most VIX ETFs end up losing money over the long term. These funds almost always lose money long term.’
Equity Factor Strategies: Trend and Quality
- Easily accessible for retail investors
- Trend has performed well in hedging against slower, protracted bear markets
- Easier to stick with due to positive expected returns during normal market environments
- Trend and quality show low correlations with each other, so can be used as a diversified hedge
- Subject to all the problems of smart beta, including:
- Variations in factor definitions between funds
- Factors can change
- Factors can decay
- It’s impossible to know when a factor stops working
- Academia is not real life
- Subject to rebalance timing luck
- Research results are often ex-post scaled to volatility targets, which is not relevant for retail investors without access to leverage
- Trend tends to focus on 12-month trends, so may suffer whipsaws during sharp selloffs
Equity Factor Strategies: The Evidence
AQR’s paper: ‘Tail Risk Hedging: Contrasting Put and Trend Strategies’ [Trend]
‘Both Put and Trend strategies have good hit rates in most equity market drawdowns, with important differences. Put strategies have offered more reliable tail insurance in fast market drawdowns, providing more “bang-for-the-buck” than Trend in gapping markets like October 1987. That said, Trend does tend to make money in the largest equity market monthly drops as they typically do not occur out of the blue. Moreover, Trend is better suited to slower, protracted bear markets, for obvious reasons. In these scenarios, Put is hampered by more sensitivity to the exact path of negative returns (e.g., a slow drawdown in which puts continuously expire out-of-the-money) and the general negative return being relevant over longer horizons.
Ultimately, the long-term cost argument tips the scales in favor of Trend. This view is reinforced by the inevitable investor impatience during the dry spells when tail insurance costs are paid year after year before the tail event materializes. A good strategy is one that you can stick with; Put-based tail hedging too often fails this test. Trend strategies do not offer as direct or explicit tail protection, but they have a strong empirical record, and you have a better chance of sticking with Trend.’
Verdad Capital’s research email: ‘Tail Risk Hedging’ [Trend]
‘Trend following was disappointing. It provided meaningful protection in only the 2000 and 2008 crises. Trend following strategies tend to focus on 12-month trends. As a result, they may get whipsawed in sharp selloffs and W-shaped recessions.’
Man Group’s paper: ‘The Best of Strategies for the Worst of Times: Can Portfolios be Crisis Proofed?’ [Trend and Quality]
‘First, futures time-series momentum strategies – which add to winning positions (ride winners) and reduce losing positions (cut losers), much like a dynamic replication of an option straddle strategy – performed well during both severe equity market drawdowns as well as recessions. Restricting these strategies from taking long equity positions further enhances their protective properties, but at the cost of a lower overall performance.
Second, strategies that take long and short positions in single stocks, using quality metrics to rank companies cross-sectionally, have also historically performed well when equity markets have sold off and in recessions, likely a result of a flight-to-quality effect. We analysed a host of different quality metrics, and point out the importance of a beta-neutral portfolio construction, rather than using the dollar neutral formulation that is more common in most published papers.
Futures time-series momentum strategies and quality long-short equity strategies are not only conceptually different, but also have historically uncorrelated returns, meaning that they can act as complementary crisis-hedge components within a portfolio.’
‘The futures time-series momentum strategies (1-, 3-, and 12-month momentum with equity positions capped at zero) demonstrate negligible correlation with any of the quality stock strategies (profitability, payout, growth, safety, and the grand quality composite). According to the research we would therefore expect the two strategies to help complement each other, potentially providing crisis hedge properties with added diversification benefit.’
NB: The following tables are taken from a slightly modified version of the original paper, which appears on the Man Group’s website.
- Have tended to rise when stocks fall, although this wasn’t the case before the year 2000
- Don’t suffer from as many credit quality/duration/credit type changes over time as aggregate bond indices, which makes managing overall portfolio risk easier. They therefore demonstrate improved volatility dampening characteristics over aggregate indices.
- Because of convexity, bonds do still have diversification potential at very low or even negative rates
- Pay contractual income payments during market crashes, which dampen drawdowns
- Extremely liquid
- Cheap to access through ETFs
- Bond correlations with equities have only been negative post-2000. Bonds were positively correlated with equities in the UK between 1764 and 2000.
- Bond returns were negative during the worst equity crashes pre-2000
- Historically, higher higher inflation equated to a positive stock bond correlation and lower inflation to a lower/zero/negative stock bond correlation
- A large part of bonds’ ability to protect from a crash protection has come from their income, rather than capital appreciation. With such low yields today, bonds may not provide as much crash protection as in the past.
- Bonds provided negative inflation-adjusted returns from 1926 to 1981. The performance streak since 1981, created by falling yields, has created two unrealistic expectations. One, that bonds earn high rates of return. Two, that if some duration risk is good then more is better. Due to today’s incredibly low interest rate environment, returns from bonds are highly likely to be lower than in the past
Government Bonds: The Evidence
Man Group’s paper: ‘The Best of Strategies for the Worst of Times: Can Portfolios be Crisis Proofed?’
‘A strategy that holds long positions in 10-year US Treasuries performed well in the post-2000 equity crises, but was less effective during previous crises. This is consistent with the negative bond-equity correlation witnessed post-2000, which on further analysis appears atypical from the longer historical perspective. As we move beyond the extreme monetary easing that has characterized the post-Financial Crisis period, it is possible that the bond-equity correlation may revert to the previous norm, rendering a long bond strategy a potentially unreliable crisis hedge.’
Panel A shows that, although post-2000 the correlation was negative, it was positive for most of the 100 years before that. This is in line with studies that argue that common fundamental factors would typically imply a positive bond-equity correlation (see, for example, Baele, Bekaert and Inghelbrecht (2010)). Funnell (2017) provides a similar long-term perspective of the bond-equity relationship for the UK.
Panel B shows that, consistent with the positive bond-equity correlation before 2000, a long bond position does not provide a drawdown hedge before 2000. In fact, bond returns are negative in quintile one (the worst periods for equities) for both the 1960-1979 and 1980-1999 periods.’
Man Group’s paper: ‘Fire, then Ice’
‘Furthermore in all three cases the peak positive stock bond correlation roughly coincided with the peak of the yields, but importantly the trough in the stock bond correlation occurred almost exactly on the trough in the bond yield. In other words, as bond yield fell from high or relatively high levels, the stock bond correlation fell and as bond yields troughed out the stock bond correlation troughed. Now these are just three episodes but it does seem crystal clear from the trends that, historically, higher bond yields equated to a positive stock bond correlation and lower bond yields to a lower or zero or even negative stock bond correlation.’
Meketa Investment Group’s paper: ‘Investment Grade Bonds’
‘From a strategic perspective, a government-only index has historically displayed more desirable characteristics, such as a higher overall Sharpe ratio and a lower correlation with equities (see Table 1). A government-only index will typically yield less, but may offer increased volatility dampening, diversification, and liquidity. On the other hand, concern about the potential for rising government credit risk may give some investors pause.’
‘The primary disadvantage of using a market-weighted index like the Bloomberg Barclays Aggregate as a benchmark or passive approach is an inability to manage the allocation’s overall risk level. That is, as the composition of the aggregate bond market changes, the fundamental characteristics of the index change as well. These changes do not necessarily align with the needs of most investors.’
Alpha Architect’s blog post: ‘Will Bonds Deliver Crisis Alpha in the Next Crisis?’
‘Historically, bond performance during Crisis Alpha months have been dominated by the income return component. The income return component delivers more than 100% of the Crisis Alpha month performance and the income return component is positive almost 90% of the time.
Historically (and surprisingly), the price return component of bond returns is negative during Crisis Alpha months!
Today, we have lower yields and the duration of bond indices is much longer than average. These changes may impact how bonds perform during Crisis Alpha months in the future which will affect the diversification benefit of bonds:
- Because bond yields are lower than historical average this decreases the income return we can expect from bonds. Given that the income portion of returns has historically comprised more than 100% of the return of bonds during Crisis Alpha months, this means that bond performance during Crisis Alpha months going forward is likely to be more muted.
- Bond index durations are longer than they have been historically. This means that the price component of bond returns will likely be a larger component of returns going forward.
Having price changes compose a larger portion of bond returns going forward isn’t good for a couple of reasons:
- Bond price performance isn’t likely to be very large (historically it is only 33% of bond total performance).
- Bond price performance during Crisis Alpha months has actually been negative historically!
- Bond price performance during Crisis Alpha months has only been positive about 52% of the time.
This analysis shows that bonds have provided some diversification benefit (positive performance during Crisis Alpha months) historically but that going forward, we may want to lower our expectations of bond performance during Crisis Alpha months.’
Movement Capital’s blog post: ‘Analyzing bond performance in stock corrections’
‘Bond performance since 1981 is anything but normal. Bonds provided zero inflation-adjusted returns from 1926 to 1981. The performance streak since 1981 has created two unrealistic expectations. One, that bonds earn high rates of return. Two, that if some duration risk is good then more is better.’
Winton’s blog post: ‘How Safe Are Your “Safe Haven” Investments Really?’
UK gilts (gilt futures closing prices that Winton has back-extended to 1975) vs FTSE 100:
Ben Carson’s blog post: ‘Allow Myself to Contradict…Myself’
“Here are some reasons for owning bonds in a portfolio:
- High-quality bonds can provide an emotional hedge against reactive behavior due to stock market volatility.
- They can act as a shock absorber when the stock market crashes, and dry powder to rebalance into the pain.
- They can provide valuable diversification benefits which allow investors to earn stable income over time.
- For those with a large enough portfolio, why take excess risk when you don’t need to?
You could have the safer side of your asset allocation in cash or cash-like products. Anyone in the shortest duration vehicles has seen their fixed income allocation hold up nicely as rates have risen over the past 27 months or so. But there is a cost to this strategy as well.
You could also own treasury-inflation protected securities (TIPS) to account for inflation. This is one of the few assets that explicitly protects your investments on a real basis.
The truth is there are a lot of strategies that can work but what really matters is being comfortable with the drawbacks of the strategy you choose because there’s no way to perfectly thread the needle.”
Ben Carlson’s blog post: ‘Do young investors need bonds?’
‘Bonds may not earn a high return going forward but if they can help you from making a bad decision during volatile markets they serve a purpose. And for those investors, young and old, who want to take more of a barbell approach to portfolio management (this would work for cash equivalents as well) they can diversify your risk buckets.
If bonds, or any other behavioral hedge, can give you a higher probability of sticking with a long-term investment plan then they’re worth holding.
It really comes down to your tolerance for volatility and large losses.’
JP Morgan’s quarterly ‘Guide to the Markets’
Portfoliochart’s blog post on ‘The Benefits of Bond Convexity’
‘This chart is one of my favorites that I’ve made in a while, as not only does it contain a lot of interesting information but I also learned a lot by making it. Here are a few of the most important takeaways:
- At high interest rates the coupon is most important, and at low rates capital appreciation is king.
- Short and intermediate term bonds (typically capped at about 10 years) are much less sensitive to interest rates at all levels than long term bonds.
- Low-interest 30-year bonds are very volatile! In fact, the range of returns is similar to what you might expect from the stock market.
- Note that the spread of total returns for long term bonds is not symmetrical. Because they are increasingly more sensitive with every drop in rates, for the same +/-1% change they actually have more upside than downside.
- One thing that’s not obvious from the chart is that interest rate sensitivity declines as bonds age. A new 30-year bond will start on the red line. When it only has 15 years left, it has the volatility of the green line. And when it only has 5 years left it has the predictable tight range of the purple line. Just like people, bonds get less active as they mature.
But if you take only one point away from this post let it be this:
Because of convexity, bonds have way more income potential at very low or even negative rates than most people realize.’
- Guaranteed to outperform in a market crash
- Positive nominal return
- Liquid, free to hold, and accessible
- Low opportunity cost in low yield environments
- Will lose value against inflation over longer term
- Safeguards may be required to avoid spending any cash allocation
- High opportunity cost in high yield environments
Cash: The evidence
Movement Capital’s blog post: ‘The case against Treasury bonds in 2020’
‘People say that cash has an opportunity cost. But that opportunity cost is a function of what comparable investments can earn. Rarely has cash (without rate risk) had this low of an opportunity cost compared to Treasury bonds (with rate risk).
I’m not in the camp that thinks interest rates will mean revert to levels from decades ago. But bond investors should acknowledge the risk they’re bearing. For example, 10-year Treasury bond prices would fall 19% if their yield rose to 3%.’
A couple of my own charts
Just in case you’re not sick of charts by now, I put together a couple of charts for us sterling investors.
Most the research above has been USD-focussed. One benefit of US research is that US asset classes tend to have the most data – often going back to the 1920s. Us GBP-based investors have less data to work with. But I think it’s still interesting to put a sterling spin on things.
The chart below shows how each of the different asset classes mentioned above has done during the 8 equity market drawdowns over 10% since 1998. All returns are in sterling, and any international bond index returns are hedged back to sterling.
Given the amount of data that’s compressed into each of the charts below, they’re probably best viewed on a desktop rather than mobile.
The two options which stood out as being the least reliable crisis hedges were the two corporate bond indices – the sterling corporate bond index, and the global hedged corporate bond index. This is probably unsurprising given the conventional wisdom that corporate bonds have similar risk factors to equities (i.e. when markets fall, the companies which issue the bonds are more likely to go bust).
The remaining options all looked like they provided good amounts of diversification during each of the crashes, each to differing degrees.
However, we’re hampered by data only going back to 1998.
Luckily we have an extra 13 years of data going back for a few of the most promising candidates crash protection. Alongside equities, we have data going back to 1985 for hedged government bonds, gold, and cash. This increases the number of drawdowns over 10% we have data for from 8 to 12:
Gold doesn’t look quite as good after included data for 1985-1998. Although only posting a 3% loss during the 1987 market crash, it fell more than the market in the 1991 drawdown, fell in line with the market in the 1985 drawdown, was down 13% in the in the 1989 drawdown.
Cash and hedged global government bonds, however, looks just as good with the extra 4 crashes included. They both posted positive returns in each of the 1985-1998 drawdowns.
Cash is often overlooked as a diversifier, because although it generates positive nominal returns, it’s guaranteed to lose value in real terms over the long run. But it’s also guaranteed to outperform in a market crash. And with rates as low as they are at the moment, the opportunity cost of holding cash is very low – bonds aren’t giving you much in the way of extra return. As we saw in the cash section above, it also outperformed bonds in 60% of 20-year periods before 1980.
An asset with positive nominal returns which is guaranteed to outperform in a market crash – I’ve seen worse products marketed as portfolio insurance.
For the chart above I annoyingly don’t have data either for gilts or for hedged government bonds going back before 1985, so we can’t see how well they fared in the pre-1985 rising interest rate environment. For that, we’ll have to rely on all the US data in the previous section.
Whether or not you believe the advantages outweigh the disadvantages for each option is up to you.
For me, I think a few options are more easily ruled out than others.
I’m happy to rule out owning put options due to the difficulty of investing in them (there are only 1 or 2 retail funds available, and all are US-focussed), their cost, their negative long term returns, and their unreliable crash protection (due to drawdowns rarely coinciding with option expiration cycles).
I’m also happy to rule out owning long volatility funds due to the extremely negative long-term returns, limited investment options, and their high expense ratios.
Finally, I’m happy to rule out owning equity factor strategies, due to all the problems surrounding the implementation of smart beta strategies, along with the inability for retail investors to gain access to leverage.
That leaves gold, government bonds (both nominal and inflation-linked), and cash.
All three asset classes are cheap, liquid, transparent, accessible, and have shown varying levels of outperformance when equity markets fall.
However, all three have their downsides.
Gold is a poor shorter-term inflation hedge, so there are no guarantees it provides positive real returns. It may be difficult for investors to stick with, given its history of large real drawdowns and unclear drivers of demand. It also provides no contractual cashflows, as bonds and cash do, so investors can’t rely on income to dampen the effects drawdowns.
Government bonds have much lower expected returns than in the past, due to today’s low interest rate environment and the extremely strong correlation between a bond’s starting yield and its expected return. Lower yields mean lower cashflows from bonds, which reduces the income received during drawdowns, so reduces the ability of coupon payments to offset drawdowns. Bonds provided negative inflation-adjusted returns from 1926 to 1981, showing that bonds may very well not beat inflation during a rising rate environment, and can underperform inflation for long periods of time.
Looking at their correlations to equity markets, bonds’ correlation with equities has only been negative post-2000, and between 1764 and 2000 – that’s 236 years – the stock/bond correlation was positive. Perhaps as expected due to their positive correlations before 2000, bond returns were negative during the worst equity crashes pre-2000. Historically, higher inflation equated to a positive stock bond correlation and lower inflation to a lower/zero/negative stock bond correlation. Higher levels of inflation may therefore reduce bonds’ ability to hedge equity market crashes.
Turning to cash, cash is guaranteed to lose value over the long run against inflation. An asset which is guaranteed to lose money is going to be difficult to stick with, despite it also being guaranteed to reduce the impact of crashes. It also can be behaviourally difficult to stick with, given how easy it can be to spend your cash balance. Holding cash as a longer-term asset allocation decision requires a mindset shift away from it being spendable money in your account, towards it being an investment decision. Obviously this can be facilitated through the use of fixed deposits or savings accounts with withdrawal penalties, but its accessibility may be a problem for some investors.
Sadly there’s no silver bullet when it comes to crash protection.
I don’t think it’s safe to assume bonds will provide the same level of protection as in the past, so the default option of only holding bonds as a diversifier seems riskier now today than previously.
Luckily I’m in the position of having the willingness and ability to be in a pure equity portfolio, so don’t have to make the decision of how best to insure my equities. However, if I were in that position, I think taking a diversified approach to diversifiers is a sensible option.
By holding a combination of government bonds, inflation linked bonds, cash, and perhaps gold (I’m still undecided on gold – see here for why), I’d be allocating towards assets with varying levels of crash-protection potential, and which are all cheap, liquid, transparent, and easily accessible.
I have no idea which one will prove to be the best crash protection going forwards, but that’s the benefit of diversifying your diversifiers. I’ll never be only invested in the best diversifier, but I’ll never only be invested in the worst, either.
I wrote extensively in this post (Smooth is what we aim for) about how important it is to aim for smooth returns. And that’s exactly what a diversified approach to diversifiers does.
Given the function of crash protection is to provide an emotional hedge against reactive behaviour during stock market volatility, I’m more confident that a basket of diversifiers will help to achieve that than by betting the house on any single choice.