Regime-Based Investing – MAN

Today’s post looks at a recent paper from MAN on regime-based investing.

Regime-Based Investing

The new paper (from February 2024) asks:

1. whether regimes can be identified in real-time (rather than just in hindsight, and more importantly,
2. if (and how) we can profit from correctly identifying them

It’s a follow-up to a 2021 MAN paper called “The Best Strategies for Inflationary Times”.

• I covered this as part of a series of articles I published that year on whether the traditional 60-40 stock/bond portfolio needed updating in the light of looming inflation.

So before we look at the new paper, I’ll recap what we learned in that series.

Recap

First, the basics:

1. Stocks are riskier (more volatile in price) than bonds but have higher long-term returns.
2. The logic of blending two assets together comes from the idea that their prices will move in different directions.
   • When stocks go down, you hope that bonds will move up in price.
3. The riskiest portfolio normally recommended by financial advisors to customers with differing levels of risk appetite would be 80-20 stocks/bonds – and the safest would be 40-60
   • So the 60-40 portfolio can be thought of as the middle option of three.
4. Bonds are a pretty bad hedge for stocks in any case, unless you either hold a lot of them (bad for overall returns) or use a lot of leverage (potentially risky, and not easy for private investors).
   • However, the 60:40 portfolio is easy to understand and has proved pretty easy to market to investors.
   • It’s also done pretty well over the decades, so many investors would miss it.
   • It was a decent choice following the 2008 crisis until the wheels came off in 2022.

The impact of inflation

1. If inflation rises for an extended period, this could be bad for both stocks and bonds.
   • In quant speak, the bond-equity correlation might turn positive (it did).
2. This is actually the normal relationship, but in recent years, investors have become used to bonds rising when stocks fall.
   • In particular, recession-driven stock market crashes lead to a demand for safe-haven assets, which are traditionally government bonds.
   • So the price goes up, and the yield goes down.
   • When the economy recovers, so do stocks, but bond prices fall back.
3. But as bond yields approach zero during the good times, they don’t have far to fall in a crash.
   • And so they have a limited ability to protect you.
4. And with bond yields close to rates on savings accounts, there’s little incentive for private investors to hold bonds at all.

Alternative assets

More diversified portfolios are more stable under inflation, so the obvious fix is to replace some or all of your bonds with alternative assets:

1. The obvious candidates are gold and cash.
   • Index-linked bonds don’t protect against inflation (it’s in the price) but rather against higher-than-expected inflation (which is harder to predict).
2. More adventurous investors might consider:
   • other precious metals and commodities in general (particularly oil)
   • private equity and venture capital
   • hedge funds (global macro and long/short equity – the latter is not readily available to DIY investors)
   • multi-asset investment trusts from the the AIC flexible investment sector – Ruffer, Caledonia, Personal Assets, Third Point, Capital Gearing, RIT and Hansa are familiar names here
   • infrastructure funds and royalty companies
   • and even a small allocation to crypto (particularly in combination with gold)
3. Even more esoteric alternatives include:
   • managed futures/trend-following (or CTA funds in the US)
   • factor funds (smart beta) and particularly multi-factor market-neutral funds (providing exposure to value, size, momentum, low volatility, and quality)
   • patience and diversification are required since no single factor works all the time and many factors have underperformed for years (eg. value right now)
   • long volatility -often a simple combination of the JPY/AUD FX rate and gold (both safe-haven instruments).
   • tactical asset allocation (TAA)
   • options strategies
4. Real assets – like property, art and farmland – are another way to go.
   • but most investors in the UK have too much property exposure through their home (since prices are high)
   • the other two assets are difficult and expensive to access.
5. Older investors, in particular, might also have DB pensions, which act as an inescapable bond portfolio without you choosing to own bonds.

Points of interest

Over my four posts, we looked at seventeen articles on fixing 60-40 / dealing with inflation. Interesting points included:

1. Stock returns are inversely related to starting valuations (eg. PEs), which are currently high
   • Stocks prefer falling inflation to rising inflation
   • Inflation higher than 5% pa is bad for stocks, though returns don’t turn negative overall (but you need to ride out some drawdowns)
   • Small size does particularly badly with inflation
   • The worst-affected sectors (particularly under very high inflation) are those that deal directly with consumers (consumer goods, autos, retail) – they are free to increase their prices but struggle to pass the impact of inflation on to their customers.
   • The best-performing sectors are energy and materials, though traditional energy might become less important to the global economy in the future.
2. Bond returns are driven by the coupon (yield) – these were low at the time, but are now closer to normal
   • More esoteric bonds offer higher yields, but also higher correlations with equities
   • Bonds can be diversifying even with (modest) rising inflation, but as their yields rise, they will lower portfolio returns, so the portfolio is not protected
3. Commodities and gold have close to zero correlation to stocks, but correlations are slightly positive when inflation is rising
   • They also have decent inflationary returns, particularly in rising inflation.
   • Energy stocks also do well, though not as well as the commodities themselves.
   • Note that commodities trend flat over the long-term
   • “A bet on commodities is like a bet against human progress” – Nicolas Rabener
4. Real estate, PE and infrastructure are not true diversifiers – they depend on positive economic growth (like stocks).
   • The diversification effects they show on paper are flattered by smoothed and lagged valuations.
5. Managed futures and long vol generate similar returns to bonds (with higher volatility)
   • Multi-factor market-neutral generates a lower return than bonds with lower volatility.
   • Managed futures and multi-factor have zero correlation with stocks
6. Long vol has a negative correlation
   • Replacing bonds in a 60/40 with long vol works well – the new portfolio has a higher return, lower volatility and lower maximum drawdown.
   • Replacing bonds with long vol, trend and multi-factor also beats the traditional 60/40, though only slightly.
7. TAA works
   • Gold is the best alternative to stocks, but this leads to larger drawdowns
   • Shorting the S&P 500, and switching to 10-year Treasuries also work, but switching to energy stocks doesn’t protect against inflation (though it works generally)

Original MAN inflation paper

The Best Strategies for Inflationary Times said:

Unexpected inflation is bad news for traditional assets, such as bonds and equities, with local inflation having the greatest effect. Commodities have positive returns during inflation surges but there is considerable variation within the commodity complex.

Trend-following provides the most reliable protection during important inflation shocks. Active equity factor strategies also provide some degree of hedging ability.

The best factor strategies were quality and momentum. Verdad commented on these findings:

Harvey et al. find that trend and quality have been top performing strategies in inflationary regimes historically, as trend would have pushed investors out of stocks at the start of their drawdowns while high quality companies tend to be more resilient in bad times.

Man also looked at crypto, concluding that bitcoin is a speculative asset with a positive beta to the US stock market – which makes bitcoin a poor inflation hedge.

The new paper

The new paper is written by Henry Neville, one of the authors of the original MAN paper.

• It looks at seven regime dimensions and tracks how six assets and strategies (Equities, Bonds, Commodities, Long-Short Value, L/S Momentum and Trend) perform through the states within these dimensions.

Inflation

The original paper looked at:

Historic episodes of general price growth acceleration as those where year-on-year (YoY) inflation rose through 2%, then 5%, then peaked (a state we labelled FIRE). 

The team has since added three other regimes:

DISINFLATION (high and falling inflation), ICE (low and falling) and REFLATION (low and rising).

The top half of the table shows real annualised return (strength of state signal), the bottom shows hit rate (persistence of signal – red or green cells denote those where performance is consistently positive or negative, defined as in more than 75% of episodes).

Real GDP Growth

We take the National Bureau of Economic Research (NBER) defined recessions as a baseline. We use the same percentage thresholds as we did for inflation to define the other three episodes. BOOM is high and rising growth, SLUMP is high and falling, RECOVERY is low and rising.

Yield curve

We split the history of the 10Y-3M curve according to a 2×2 matrix: whether the curve in general is rising or falling (bear and bull, respectively), and whether its gradient is steepening or flattening. Given the interest in the inverted curve, we add this as a fifth regime state.

Stock market

Interest rates

Commodity curve

AQR has defined whether the futures curve in aggregate is in backwardation or contango. We’ve taken this data [and] subsumed any episode where the contango or backwardation was shorter than six months within the prior state.

Value vs Growth

[The chart] shows the long-term performance of the Fama-French HML factor (as a proxy for a strategy that buys cheap and sells expensive stocks), with time split into Value (yellow) and Growth (blue) states.

Conclusions

Henry makes eleven points (though he claims there are only nine):

1. Inflation is probably the most predictable of the regime frameworks.
   • If you are only allowed to use one economic datapoint to guide your decisions, US headline CPI should be it.
2. There has not been a real growth boom since the 1990s. Fingers crossed for the AI productivity miracle.
3. Value is consistently one of the hardest investment strategies to fit into a regime framework.
   • It is a myth, for example, that it consistently benefits from high and rising inflation.
   • Don’t touch it when the yield curve is bull flattening.
4. Equity bull markets are 80% of history.
5. Bonds are positive (in real terms) in little more than half of equity bear markets.
   • Historically there have been better defensive diversifiers.
6. Equities have ALMOST ALWAYS worked when real growth is booming and when monetary policy is being loosened.
7. Trend has ALMOST ALWAYS worked when Value is in the ascendant.
8. Trend and Commodities have ALWAYS worked in states of high and rising inflation.
9. Bonds have ALWAYS worked in deflationary environments, as well as when monetary policy is being loosened.
10. Momentum has historically done 10% real in states where Value is underperforming, versus 1% in the opposite.
11. In almost all frameworks, Momentum and Trend are positive more than half the time in all states.
    • If you can hold through initial reversals, patterns often reassert.

That’s it for today.

• It’s been a nice refresher for me on defensive portfolio construction and has encouraged me to further investigate market timing (regime) approaches.

Henry says:

This is somewhat experimental, scratching an itch that has been nagging me a while. These are my first thoughts, hopefully a conversation starter, do get in touch if you’d like to continue it.

Let’s hope that people do.

• Until next time.