Bank of America is forecasting flat returns over the next decade. Yes, you read that correctly — 0% returns to equities is what a team of BofA analysts are anticipating. The research team, led by equity and quant strategist Sativa Subramanian, suggests investors brace themselves for a lost decade ahead. The investment bank’s long-term valuation model indicates a -0.8% annualised return for the S&P 500 over the next ten years. Subramanian says she expects the S&P to rise just 2% over the course of 2022. “This may not end now. But when it ends, it could end badly,” Subramanian warned.
“In shallow waters we will see the real difference between bottom trawling and skilled fishing.”
Manas D. Kumaar, PWE Group CEO
PEAK FINANCIAL CAPITALISM?
The analysts cited supply chain issues, peak globalisation and inflationary pressures squeezing profit margins as cause for concern. Bank of America is not alone in its bearish outlook. Wall Street is relatively neutral on US equities, yet the consensus is as close to a sell signal as it has been since 2007, Subramanian heeded. The investment note suggested that risk premia do not adequately reflect supply chain dangers and other global frictions.
The BofA strategists advised reinvesting dividends to combat reduced capital appreciation. They estimated that a 36% total return could be achieved over the next decade by reinvesting dividends. That represents a paltry 3% annualised total return. In comparison, the S&P 500 returned 360% over the last decade, an annualised return of roughly 16%.
Furthermore, those absolute returns do not account for inflation potentially running hot and eroding real yields much further. Assuming central banks succeed in keeping inflation within their mandated 2% target, BofA’s scenario implies a 1% real dividend yield and no capital gains.
We have become accustomed to easy central bank money, making it relatively effortless to invest passively. A rising tide lifts all boats. Now, most analysts would agree we are headed towards more turbulent waters.
Asset allocation has been the primary determinant of portfolio success in recent years, but security selection is becoming very relevant once again. Direction agnostic hedge fund strategies, which many investors have shunned in favour of passive approaches, will likely play a more prominent role in portfolios going forward. Society will no longer reward simplistic buying and holding. In the long run, challenging markets make for healthier, more efficient markets, where “smart money” is compensated accordingly.
“A smooth sea never made for a skilled sailor. The coming decade will separate the beta from the true alpha. Those who fail to adapt now will not just miss the boat; they will sink.”
Manas D. Kumaar, Group CEO
It will come as no surprise if many equity hedge fund strategies continue to underperform going forward. Numerous authors such as Mitchell and Pulvino (2001), Brooks and Kat (2002), Agarwal and Naik (2004), Capocci et al. (2005), Boyson et al. (2010) and Guesmi et al. (2015) have found strong correlations between the vast majority of equity hedge funds and stock market indices, particularly the Russell 2000. Agarwal and Naik (2004) further found strong correlations between merger arb funds during bear markets. Many funds promise to generate returns during market turmoil, citing downside protection as a justification for their high fees. However, Guesmi et al. (2015) found that most of the industry failed to generate any positive returns during crises such as that of 2008. There is no evidence to suggest the bulk of equity strategies will fare any better in the future, given that markets are expected to become even more challenging.
“Most of the industry failed to generate any positive returns during crises such as that of 2008.”
Manas D. Kumaar, Group CEO
High net worth investors should be equally concerned about the management of their money at the portfolio level. Benchmarking is an integral part of return attribution, but it can incentivise group-think and a herd mentality. Although most portfolio managers are well aware of the equity market outlook, how many of them do you think will dramatically alter their portfolio weighting? Most will probably stick to what they’ve always done and what everyone else is doing, hoping for the best. Hoping the Fed will come to the rescue as usual. But what happens when markets finally realise that the central bank emperors have no clothes left to continue playing strip poker?
FIAT CURRENCY, DIGITAL STRATEGIES
In terms of asset allocation, an overlooked opportunity resides in the currency market. Unlike the cryptocurrency market, the traditional currency market is far more liquid and does not entail imminent regulatory risks, as we discussed previously. The cryptocurrency market may be fashionable, particularly among young investors who have never heard of Tulip mania or the South Sea bubble. However, it suffers the same plights as traditional markets, and then some.
On the other hand, currency hedge fund indices have shown robust diversification benefits, as we will highlight. Fundamental factors drive the global currency market. Econometricians and machine learning models can quantify these factors concretely. The currency market’s vast liquidity and low transaction costs make it well suited to more modern quantitative strategies such as high-frequency trading. HFT and other novel strategies can consistently yield extraordinary returns, with no correlation to broader markets whatsoever. However, for hedge funds, high-frequency trading in the currency markets has been fraught with difficulties. The currency market operates on a sprawling, fragmented over-the-counter broker-dealer network as opposed to centralised exchanges. As a result, conflicts of interest and opaque order flows have made it difficult for most fund managers to navigate.
WHAT IS THE MATRIX?
The below heatmap displays the correlation matrix of various hedge fund strategy indices, ETFs and risk premia proxies, on a monthly basis between June 2013 and October 2021. The green represents a high correlation, while the red indicates a negative correlation.
Perhaps the most striking observation from this heatmap is the performance of currency hedge funds. The HFR FX Value and Eurekahedge FX indices were either uncorrelated or inversely correlated with most other strategies, risk premia and asset classes. The HFR FX Value index was only positively correlated with the VIXY VIX ETF. We selected the VIXY as a proxy for overall market volatility. The association highlights how currency value strategies can perform better in volatile markets while most asset classes suffer.
Similarly, the Eurekahedge FX Hedge Fund index also performed well in volatile markets. Furthermore, it was positively associated with the US Dollar Index, which tends to rise when investors move to cash. The Dollar Index increases as traders unwind Carry positions in times of heightened fear and uncertainty. Interestingly, the Eurekahedge FX Index was also correlated with the Eurekahedge CTA and Trend-Following indices. CTA (commodity trading advisor) funds typically trade based on trend-following and momentum strategies, so it is no surprise that they move in tandem with trend strategies. Both indices have largely lived up to their promise of providing robust diversification benefits, albeit not as much as currency funds. Their association with the FX fund index suggests that many currency strategies employ trend-following rules, which explains the overlap.
“WE ARE ALL INDIVIDUALS”
Another striking feature of the matrix is the degree to which so many allegedly diverse hedge fund strategies are incredibly correlated with one another. For example, the Eurekahedge overall Hedge Fund Index, Macro, Arbitrage, Long-Short Equities, Long-Bias, Multi-Strategy, Relative Value, Event-Driven, Fixed Income, and Distressed Debt, all have either a very high or close-to-perfect correlation with one another.
“The homogeneity of hedge funds should be all the more concerning because correlations tend to increase even more during times of crisis.”
Manas D. Kumaar, Group CEO
Furthermore, Equity Neutral, Structured Credit and Commodities were all relatively correlated with the rest of the pack. Even Artificial Intelligence-based funds displayed some degree of co-movement with the crowd. You can see the specific correlations in the below Fruchterman-Reingold graph by hovering over each security.
The size of each node is based on the absolute value of annualised log returns, ignoring risk. Bitcoin and Cryptocurrency hedge funds have had the highest returns of any asset class over the past decade. Their risk-adjusted returns have also been the highest. However, historical risk-return measures like the Sharpe ratio fail to capture the worrying regulatory risks facing cryptocurrencies going forward. Although crypto funds do have promising diversification benefits, the Eurekahedge Cryptocurrency Index is almost totally exposed to the price of Bitcoin, with 93% of its returns explained by Bitcoin price movements.
“If I wanted a 90% correlation with Bitcoin, I would buy Bitcoin.”
Manas D. Kumaar, Group CEO
A better visual representation of the interplay between each strategy class can be seen in the below Minimum Spanning Tree. The graph reduces the number of edges in the correlation network to a subset that connects all nodes with the lowest possible correlation values. Returns statistics are viewable by hovering over each network node.
“SMART” BETA – AN OXYMORON?
As expected, most returns sources cluster around the S&P 500 ETF (SPY). Zooming in on the cluster (click the + button, then click the cross icon to enable panning), then hovering over the SPY node, we see that the S&P 500 is highly associated with the Fama-French risk premia for quality, size, and value. The Fama French model attempts to describe stock returns using five factors: (1.) Overall market risk, known as beta. (2.) The tendency for small-cap stocks to outperform large-cap stocks. (3.) The tendency for companies with high book values to market values to beat companies with low book values to market values. (4.) Stocks that have high operating profitability tend to outperform. (5.) Companies with high total asset growth tend to underperform.
The ‘Quality’ factor is another risk premium that was not part of the Fama-French model. It’s closely associated with the ‘Value’ factor. Asness and Frazzini (2013) of Yale University define the Quality Factor as “safe, profitable, growing, and well-managed” stocks.
The S&P 500 is commonly used as a proxy for the overall market to calculate systemic beta risk. In addition, the various other risk factors used to dissect returns have given rise to the era of passive investing via Smart-Beta ETF products. However, if the Quality, Size, and Value factors are all over 90% correlated with the S&P 500, it begs the question, ‘how smart is smart-beta?’ Why should any investor pay fees for multiple ETFs, which essentially track the S&P 500, when they could simply buy one S&P 500 ETF? Moreover, should we expect any passively managed investment vehicles to perform well over the next decade?
“If Quality, Size, and Value factors are all over 90% correlated with the S&P 500, it begs the question, ‘How smart is smart-beta?'”
Manas D. Kumaar, Group CEO
On the other hand, the past decade was one of astronomical monetary stimulus from central banks. So naturally, the rising tide lifted many boats in tandem. If we do indeed experience a prolonged depression akin to that of the 1930s, value stocks and quality stocks may become more decoupled from broader markets.
Additionally, there is one factor in particular that may be misleadingly portrayed within the above Spanning Tree. The momentum factor, represented here by the iShares MSCI USA Momentum Factor ETF (MTUM), exhibited a correlation of 0.91 with the S&P 500 since 2013. However, by its very definition, momentum investing seeks to buy stocks that have been rising and sell stocks that have been falling. Therefore, momentum strategies will have a higher beta during bull markets and become inversely correlated with stocks during a bear market.
On that note, looking at the Spanning Tree, we can see that CTA and Trend-Following funds form a branch away from the central cluster. As previously discussed, their diversification benefits have been formidable. Their returns profile is all the more impressive, considering the period under observation was primarily a bull market. Surprisingly, neither index was all that correlated with the momentum factor. This differentiation may reflect that funds now employ more complex strategies than the traditional cross-sectional and time-series momentum rules suggested by Jegadeesh & Titman (1993) and Carhart (1997). For example, the iShares MTUM smart-beta ETF in the above graph simply screens for 6-month and 12-month risk-adjusted price momentum. The lack of substantial correlation between MTUM and the CTA and trend-following hedge fund indices underscores the difference between active management and passive products which seek to mimic their returns. Although the smart-beta ETF significantly beat those funds over the past decade, we might expect rigid trend-following and momentum strategies to fare less smoothly in turbulent markets.
THE ALTERNATIVE ALTERNATIVE
Currency carry strategies performed poorly, as we discussed in our previous Currency Market Strategy article. The spanning-tree shows their exposure to emerging markets. When developed markets sneeze, emerging markets catch a cold. Furthermore, considering global interest rate spreads are tightening, there is no reason to expect the trend will reverse any time soon.
Finally, zooming back out on our spanning tree (click the home button), perhaps the most striking depiction is the outliers. The only hedge fund indices with either no correlation or negative correlations with the rest are currency fund indices — the Eurekahedge FX Index and the HFR FX Value index. Despite the extraordinary diversification benefits accruing to those funds, their returns have been relatively lacklustre. That may not be an issue for many sophisticated investors who utilise leverage and prioritise risk-adjusted returns, but it is nevertheless sub-optimal.
What if there was a direction-agnostic currency market strategy that delivered consistently exceptional returns?
Although high-frequency trading has been around for quite some time, we did not include any HFT indices in this analysis. That is because there are no high-frequency trading hedge fund indices. HFT strategies are not replicable by any smart-beta ETFs, and there are no publicly available HFT index products. In August 2021, the Bank for International Settlements published a paper titled “Quantifying the high-frequency trading ‘arms race’.” While we do not necessarily agree with the implications of the BIS paper, it highlights the breadth of opportunities for high-frequency trading in the currency markets. The BIS estimates that latency arbitrage generates $5 billion per annum in equity markets alone. HFT strategies are not a homogenous group. They may incorporate arbitrage, momentum, trend-following, or other components, but at speed. HFT strategies typically utilise machine learning because humans physically cannot react at sub-second intervals. However, rather than entirely relying on a “black-box” approach, there are various mechanisms and methodologies by which we can “merge man with machine” to augment split-second decisions successfully.
PWE Capital is a fintech firm specialising in the development of technologies to facilitate high-frequency trading in the currency, commodities and derivatives markets. We created our own trading platform to rectify some of the common pitfalls associated with high-frequency currency trading. The typical issues all have one common root — the broker and counterparties. We successfully remediated the broker issue by removing them from the equation entirely. As a result, we can rely on our algorithms to perform in the conditions they were trained in — a transparent trading environment.
For more information about our current offering, feel free to learn more here.