Riding the Waves of Liquidity and Momentum – Introducing the BT Momentum and BT Wave Runner Portfolio Strategies

Our last post got us thinking about how difficult it must be for individual investors to stick to a portfolio strategy and drown out the noise given the recent news about inflation, interest rates, and Russia/Ukraine war and calls for a potential recession or even stagflation. 

So we set out to create a systematic trend following portfolio strategy based on academic and practitioner investment research that would work well in any market or economic regime (based on simulated historical analysis) and would allow investors to potentially make money in any environment by using both fundamental and sentiment indicators to help position portfolios to ride the waves of liquidity and momentum.

These portfolio strategies build upon the posts we have previously published that show how our opinions and views may be executed to provide tangible value over time.    

We will cover the following topics in today’s post:  

Topics Covered:

1. Introduction to Momentum/Trend-following Investing
2. Introducing BT Momentum Strategy – Systematic Portfolio Management
3. Introducing BT Wave Runner Strategy – Systematic Portfolio Management
4. BT Wave Runner Risk Composite Indicator – the components
   • Fundamental
     • BT Global Liquidity Index
     • OECD Composite Leading Indicator
     • U.S. Housing Starts
   • Market Sentiment  
     • High Yield vs US Treasuries
     • High Beta vs Low Beta Currency
     • Copper-to-Gold Ratio

Summary:

• We believe that trend-following/tactical asset allocation portfolio strategies provide great risk-return trade-offs and offer a significant improvement in terms of risk-adjusted returns over static asset allocation portfolio strategies. We evaluate 2 portfolios (BT Momentum and BT Wave Runner) based on simulated asset returns. To be included in the investment universe, the asset classes require a separate ETF to exist and are investable by individual investors.  

• The combination of fundamental and market sentiment data provides risk control when incorporated with momentum that has improved the risk-adjusted returns of portfolio strategy by riding the liquidity and momentum waves.

1) Introduction to Momentum/Trend-following Investing:

What is Momentum?

Momentum is the phenomenon that securities that have performed well relative to peers (winners) on average continue to outperform, and securities that have performed relatively poorly (losers) tend to continue to underperform.

Momentum is an investment style that has been studied extensively by academics for many years. AQR Capital Management has extensively researched momentum as an investment style and shown evidence for momentum is pervasive and supported by almost two decades of academic and practitioner research.

Studies have documented momentum in the U.S. as far back as the Victorian age. The evidence also shows that momentum works broadly across asset classes, including foreign stocks, bonds, commodities, currencies, index futures, and global country index selection. These studies are largely based on price returns and some fundamental considerations (i.e. earnings momentum). 

So Why Does Momentum Work?

Momentum is driven by investor behavior: slow reaction to new information, asymmetric responses to winning losing investments, and the bandwagon effect/overreaction as short-term traders may use recent performance as a signal to buy or sell.

Momentum tends to persist for some time (6-12 months) before leading to reversals as too many investors pile in the same investments as prices become detached from fundamentals, causing volatility and drawdowns in asset prices.   

2) Introducing BT Momentum Strategy – Systematic Portfolio Management:

Based on this excellent research, we wanted to see if we may apply it in a systematic way that may reduce investor behavioral biases (i.e. selling at the bottom of the market cycle or buying at the top). As we have noted in our research, investor risk appetite/behavior is a very significant factor in the results as measured through Household Balance Sheet Allocation to Equities (based on Flow of Funds position Accounts) and Margin Debt.

We have taken this research and applied it to ETF investing on a global basis looking at 64 ETFs/asset classes spanning equities, currencies, bonds, commodities, and Real Estate Investment Trust (REITs).

The idea is if you can maximize the investment universe, an investor may maximize their returns by catching a momentum wave in a particular asset class at a particular time (i.e. when the trend is your friend).   

We have called this the BT Momentum strategy (a long-only strategy), which is a simple strategy of investing in the top 3 ETFs (weighting of 40%/40%/20%) each month based on prior month data using a momentum score called Fast Momentum (13612W) which we will discuss later on, starting with monthly asset class returns in 1985.

Each month the investor would sell the previous month’s investments and buy the 3 strongest asset classes based on the 13612W fast momentum filter for each, based on buying at the beginning of the month based on the prior month’s momentum signals.   

This would lead to systematic buying and selling every month based on momentum signals. We have examined the portfolio statistics over the last ~40 years and split them into 4 different periods (roughly 10 years) to determine if there is persistence in returns as the research states. Given that trading is largely commission-free, implementing this strategy on a go-forward basis, we believe transaction costs should be minimal.        

In Table 1, we show the results of the simulated portfolio starting with $100 (no further contributions) in 1985, which has multiplied to $28 million by end of 2021 based on compound annual returns of 40% per year, far outperforming the Wilshire 5000 by ~30% per year and 60% of the time on a monthly basis – which are very strong results.

The overall correlation of monthly results of the BT Momentum portfolio to US equity markets is low (0.32) over the period, thus providing some diversification to standard long-only heavily weighted US equity portfolios. These results are impressive indeed and appear that results are persistent across time.

Table 1 – Summary Results – BT Momentum Portfolio Strategy vs Wiltshire 5000 – 1985 to 2021 (monthly results)

However, the annual volatility of 40% of the BT Momentum strategy is far too much for most investors to stick with this strategy. Also, the BT Momentum portfolio strategy provides little drawdown protection relative to a buy-hold strategy of the Wilshire 5000 index (see Table 2). 

Table 2 – Summary Results – BT Momentum Portfolio Strategy vs Wiltshire 5000 – 1985 to 2021 (monthly results)

We set out to try to reduce the annual volatility/max drawdown by using the BT Market Risk Indicator composition which identifies periods of economic and liquidity expansion, which we will cover later on in our Wave Runner Portfolio Strategy section.

Defining the Momentum Score

Typical momentum strategies use a trailing 12-month return or 10-month Simple Moving Average (SMA) to determine which asset classes may exhibit higher relative scores of momentum. 

We do not use the traditional 10-month SMA filter in our analysis, but something slightly faster, the weighted variant of the well-known average return over the last 1, 3, 6, and 12 months, called Fast Momentum (13612W filter) which has been popularized by Wouter J. Keller on his website TrendXplorer. 

Table 3 below shows how different momentum filters pick up monthly returns at a different paces. We note that the 13612W filter picks up about 70% weighting on the previous 3 months (bottom right corner chart). So the 13612W momentum factor picks up reversals in trend faster than other commonly used methods. We have used the 13612W momentum score within both the BT Momentum and BT Wave Runner portfolio strategies as well as Risk Composite Indicator.

Table 3 – Momentum Filter Comparisons

Investment Universe

For inclusion in the investment universe, all asset classes require a separate ETF to exist and are investable. We have attempted to maximize the investment universe to all global equity markets (including sector ETFs), select commodity markets that currently have ETFs in the market, Bond Markets (largely $US Sovereign Debt), and FX ETFs (based on US dollar). In all, we have selected 64 asset classes – see Table 4 for select asset classes.

As the longest-serving ETFs only go back to 1994, and we wanted to have at least over 30 years of monthly price data in our test, we have simulated historical results based on index returns using information from MSCI, and the US Federal Reserve FRED database. 

We have chosen to not include crypto assets such as Bitcoin and Ether in these back test as we have previously modeled crypto prices and noted that they tend to trade similar to a 5x leveraged NASDAQ index. In future back tests, we may include crypto assets as more data becomes available.

Table 4 – Investment Universe – 1985 to 2021

Asset Allocations BT Momentum – 1985 to 2021

As we mentioned earlier, capital is allocated every month to the top 3 ETFs based on momentum scores in a weighting of 40%/40%/20% to the risk assets.

Below in Table 5, we have noted that the top portfolio allocations over time have been largely allocated to commodities and energy (priced in US dollars), supply of US dollar liquidity, and consistent with a growing global economy over time.

The momentum effect is well known in commodity markets and trend-following strategies are used by commodity trading advisor (CTA) firms, so it is not surprising to see many in the Top 10. Also, countries that benefitted from high economic growth and further development during the period showed higher momentum (i.e. Mexico, China, India, and Russia).

We also note that US REITs have a high allocation over the period, but very little has been allocated to US equities. As we have noted in the past, Real Estate is the largest asset class in the world, and US Real Estate through increasing collateral values and increasing credit creation/higher liquidity is particularly important to the “Wealth Effect” and driving consumption in the US, which helps support the global growth cycle through further consumption. Over each period, the BT Momentum Strategy provides returns at least as good as US equity market returns.   

Table 5: BT Momentum Strategy – Average Asset Allocations Over Time

We note that BT Momentum provides some risk diversification to US equity portfolios based on low diversification over the measurement period (below 0.50 for most of the period based on a 3-year rolling correlation) based on Table 6.

Table 6: Rolling Correlation of BT Momentum vs US Equity Markets

3) Introducing BT Wave Runner Strategy – Systematic Portfolio Management:

The BT Wave Runner Strategy builds upon the BT Momentum strategy by adding a Risk Composite Indicator based on Fundamental data and the BT Market Risk Sentiment Indicator as a risk control to reduce the annual volatility, as the indicator identifies periods of economic and liquidity expansion.

The Market Risk Sentiment indicator is the same one introduced in our first post which helps out monitor the global growth cycle daily and has been highly correlated with the OECD Composite Leading Indicator on a monthly basis.

The idea of this portfolio strategy is simple, to avoid investing in risk assets in volatile periods that tend to be more susceptible to drawdowns (i.e. investor risk appetite tends to change quickly and frequently, which tends to create volatility in momentum portfolios), invest when the trend of liquidity in markets and industrial production/general economic outlook (as measured by the OECD Composite Leading Indicator) is better than the 3-year moving average trend. Or in other words, the trend is your friend.

In periods of more volatile markets, the portfolio is invested in a Safety portfolio based on the 13612W momentum filter of the safety investment universe.

In Tables 6 and 7, we show the results of the simulated portfolio starting with $100 in 1985 (with no future contributions), which has multiplied to $265,000 by end of 2021 based on compound annual returns of almost 25% per year, far outperforming the Wilshire 5000 by 2x and outperformed the Wilshire 50% of the time on a monthly basis. The largest drawdown is about 50% of equity markets as well and offers similar drawdown protection as a 50% risk-asset/50% safety asset portfolio.

The overall correlation of monthly results of the BT Wave Runner portfolio to US equity markets is low (0.20) over the period, thus providing some diversification to standard long-only more heavily weighted equity portfolios. These results are impressive as volatility is roughly that of US equity markets. The BT Wave Runner portfolio strategy offers about 50% of return/volatility of the BT Momentum Portfolio. It offers 42 times the cumulative value of a long-only equity market portfolio by sidestepping large drawdowns and riding the combined waves of momentum and liquidity.

Table 6 – Summary Results – BT Wave Runner Portfolio Strategy vs Wiltshire 5000 – 1985 to 2021 (monthly results)

Table 7 – Summary Results – BT Wave Runner Portfolio Strategy vs Wiltshire 5000 – 1985 to 2021 (monthly results)

BT Wave Runner Decision Rules:

In Table 8, the flow chart reviews the decision rules in the Wave Runner model.  

If the Risk Composite Indicator is above trend, we expect that global investor risk appetite should be above trend based on the 13612W momentum filter. As a result, the portfolio is invested in the top 3 asset classes as measured by 13612W.  If Risk Composite Indicator is below trend, the portfolio is invested in a safe portfolio allocated to 6 assets that tend to be very safe in periods of extreme market volatility (flight to quality to reserve assets).

Table 8: Decision Rules BT Wave Runner Strategy

The Safety Portfolio:

The safety portfolio is based on the idea that the US dollar and US Treasuries are the world’s reserve assets underpinning the world’s financial system. In our previous post, we have discussed the dominance of the US dollar, US Treasuries, and Gold held by Central Banks reserves as the US dollar holds significant weight in the Financial System. There tends to be a flight to quality (US dollar reserve assets and Gold) when markets are volatile. 

As a result, we have restricted our safety allocation to largely US dollar reserve assets. When the Risk Composite is below trend, the portfolio is invested in the top 3 assets based on the 13612W momentum filter based on a weighting of 40% for 1^st and 2^nd and 20% for 3^rd highest momentum.

• iShares 20+ Year Treasury Bond ETF (TLT)
• iShares 7-10 Year Treasury Bond ETF (IEF)
• iShares 1-3 Year Treasury Bond ETF (SHY)
• iShares GNMA Bond ETF – Mortgage-backed Securities (GNMA)
• Invesco DB US Dollar Index Bullish Fund (UUP)
• SPDR Gold Shares (GLD or PHYS)

4. Risk Composite Indicator

The Risk composite indicator is a Z-Score measure based on the following measures below. The weighting of each component of the composite is based on the inverse of the volatility of the Z-Scores for each component.

52% of the time the portfolio is in the safety portfolio (see Table 9 for historical Risk Composite Indicator) and 48% of the time in the BT Momentum Strategy (Risk-on). We should note that we have not fitted this indicator via regression to equity markets to optimize the weighting between the various indicators (this could significantly improve the backtest metrics).   

Fundamentals – 50% of the Risk Composite: 

• BT Global Liquidity Index, 6-mths advanced  
• OECD Composite Leading Indicator, 6-mths advanced
• US Housing Starts

Market Sentiment – 50% of the Risk Composite:

• High Yield Credit vs US Treasury
• High Beta Currency (AUS)  vs. Low Beta (JPY)
• Copper vs Gold Ratio 

Table 9: Risk Composite Indicator

A. Fundamentals (50% of Risk Composite Index)

1. Why is Liquidity Important for Asset Returns?  

As we discussed in our previous post, ‘The ‘Wealth Effect’ and Debt – Two Sides of the Same Coin – Part 1’ liquidity and the credit cycle are important to the functioning of our global financial system.  

The credit cycle refers to the self-reinforcing interactions between perceptions of value and risk, risk-taking, and financing constraints. Typically, rapid increases in credit drive up property and asset prices, which in turn increase collateral values and thus the amount of credit the private sector can obtain until, at some point, risk appetite reduces due to change in conditions where the debt can no longer be serviced or bankers are no longer willing to supply credit (i.e. increasing unemployment/bankruptcies, pandemics which reduces the labor supply, natural disasters, interest rates are raised beyond the neutral rate of interest, war/tensions, etc.).  

Each time a bank creates a loan, a corresponding deposit is created through the banking system, which adds to the money supply. Money is used to purchase goods and services and accumulate and save for a property. As the money supply grows, cash is devalued against assets (financial assets, real estate, commodities, art, etc.), causing nominal prices to rise.

Famed investor Stanley Druckenmiller has said before: “Earnings don’t move the overall market; it’s the Federal Reserve Board… focus on the central banks, and focus on the movement of liquidity… most people in the market are looking for earnings and conventional measures. It’s liquidity that moves markets”.

We can see this in action in Table 10 as BT Global Liquidity Index leads S&P 500 growth by about 12 months. We use information from National Statistical Authorities OECD, IMF, Bank of International Settlements, and US Federal Reserve to calculate the BT Global Liquidity Index.   

Table 10: Global Liquidity vs S&P 500 EPS Growth – 1993 to 2022

Financial liquidity explicitly drives investors’ risk appetite and asset allocation. Rising collateral values then positively feedback to underpin new liquidity/credit creation. We have seen that more liquidity has reduced the risk premium on risk assets, as well as cutting defaults, allowing investors to invest in riskier investments/longer duration assets such as technology stocks that may not turn a profit until many years in the future. Higher investment in technology also improves economic productivity, which offsets older demographics that plague most countries globally. We can see below in Table 11, that Global Liquidity, advanced 6-months tends to lead NASDAQ returns (Z-Scores) (a proxy for returns on technology investment).  

Table 11: BT Global Liquidity Index vs NASDAQ Returns – 2000 to 2022

We have used a similar methodology that Cross-Border Capital uses to determine Global Liquidity as defined in Michael Howell’s book Capital Wars – “The Rise of Global Liquidity”.

Howell defines Global Liquidity, “as a source of funding that measures gross flows of credit and international capital feeding through the world’s banking system and collateral-based wholesale funding markets. It determines the balance sheet capacity of all credit providers and the private sector’s access to cash through savings and credit”.

• Private Sector Liquidity: Includes credit flows from the Banking sector, and Shadow Banking sector (consisting of entities such as asset-backed commercial paper (ABCP) conduits, credit hedge funds, finance companies, government-sponsored enterprises (GSEs), money market mutual funds (MMMFs), securities lenders, insurance/pension  and structured investment vehicles (SIVs))
• Central Bank Liquidity: Foreign and Gold Reserves and balance sheet allocation to Quantitative Easing programs.     
• Cross Border Funding: Foreign investors and lenders through cross-border flows. These flows have historically provided a preview of the pending economic crisis and tend to be marginal capital allocated which drives markets.    

BT Global Liquidity Index includes flows from the following countries: China, United States, Europe, UK, Japan, Canada, Australia, South Korea, Russia, India, Brazil Mexico, South Africa, Switzerland, and Singapore. Table 12 shows US Liquidity by sector.

Table 12: US Liquidity by Sector

2) Is the OECD Composite Leading Indicator helpful in forecasting the real economy? 

The OECD system of composite leading indicators (CLI), first developed in the 1980s, is designed to give early signals of turning points in economic activity.

The CLIs aim to anticipate fluctuations in economic activity over the next six to nine months based on a range of forward-looking indicators such as order books, confidence indicators, building permits, long-term interest rates, new car registrations, etc., as well as a monthly index of industrial production (IIP) as a proxy measure for economic activity. OECD CLIs aim to predict turning points in this business cycle estimate (signal a turning point in the business cycle in 6-9 months).

The CLI is timely and published every month (a lag of about 45 days after month-end) and is more frequently updated than GDP which is usually updated every quarter with a 2-month lag. High-frequency monthly production data such as Plastic production and Electricity consumption appear significant in China and Mexico (countries that rely more on production for economic growth on a relative basis). The OECD CLI is a significant input in our country’s Machine Learning models which we have discussed in prior posts.    

The OECD CLI is used within the BT Wave Runner portfolio strategy to aid in assessing the growth/inflation (business) cycle and has been able to identify turning points in advance so that investors may reduce risk exposure in anticipation of drawdowns/market volatility in risk assets.

In Table 13, we observe that GDP growth is strongly correlated with OECD CLI. Based on a Granger causality test, the OECD CLI leads the real GDP growth by about 2-quarters (F=3.7247, statistically significant at 5% level).

We believe the OECD CLI would be useful in measuring the real activity in the economy and is complementary to the BT Global Liquidity Index as we believe the flows of liquidity via the ‘Wealth Effect’ through the financial cycle drive real economic growth.

Table 13: OECD Real GDP growth vs OECD CLI

3. Why Do U.S. Housing Starts Matter?

Housing/Real Estate — and all of the ancillary spending associated with the purchase of a home, including renovation and remodeling costs and utilities — is a huge part of the global economy.

Real Estate is the largest asset class globally. Every recession since 1960 has been preceded by a double-digit decline in housing starts. It can take several months for home builders to construct a new property. And homebuilders are reluctant to break ground on new projects if they fear the economy may slump later in the year.

In our previous post, we have covered the importance of US Real Estate and how it contributes to consumption in the US which drives U.S. GDP growth, and further drives the wealth of the Top 10% within the country. Continuing global growth depends on the consumption of the American consumer and the continuance of the Wealth Effect.

Table 14 – U.S. Housing Starts – Good Recession Indicator?

B. BT Market Sentiment Indicators (50% of the Risk Composite Index):

In our first post, we discussed a daily risk indicator that we have used in our Wave Runner portfolio. To manage and monitor our risk on a more frequent basis, we have attempted to create a customized index that may be tracked daily that aligns well every month with the OECD CLI which tracks the business cycle.  

This custom index which we have called the BT Market Risk Indicator includes the following measures:

• Copper-to-Gold Ratio
• High Beta (AUD/USD)-to Low Beta (JPY/USD) Currency Pairing 
• U.S. Discretionary Consumer Spending-to-Consumer Staples Ratio*
• U.S. High Yield Corporate Credit to Intermediate Gov’t Bond Ratio
• U.S. TIPS Bonds to Intermediate Gov’t Bond  Ratio*

* Not included in current 1985 to 2021 backtest given data limitations.

The OECD CLI is a significant indicator based on our research to track going forward, however, there is a lag of about 45 days to report results for a given month. This makes our approach susceptible to quick market drawdowns, such as the one we experienced in February/March 2020. However, looking back during that time the OECD CLI was already in negative territory and would have indicated an extremely cautious risk position in our investment portfolio.

In Table 15, we show that the BT Market Risk Sentiment indicator and OECD CLI tend to move in the same direction over time. We have discussed in our prior posts, the importance of the Copper-to-Gold Ratio and how commodity currencies (Australia dollar, Canadian dollar) tend to follow a cyclical upswing, and generally, during a crisis, there is a flight to safety (U.S. Treasury bonds, gold, and the Japanese Yen).  We have also discussed how credit spreads tend to widen in the late-cycle period as market interest rates start to rise higher than the neutral rate of interest, reducing investor risk appetite.

Table 15 – BT Market Risk Sentiment – A Timely Proxy for OECD CLI?

Asset Allocations – BT Wave Runner

The Wave Runner portfolio shows a similar allocation to BT Momentum, however, was able to side steps large drawdowns, and reduce the annual volatility by about 50% (see Table 16).

Table 16 – BT Wave Runner Average Allocation over time (Risk Assets Only)

The Wave Runner Portfolio is in the Safety Portfolio for 229 months out of 443 months (52% of the time). This has significantly reduced the annual volatility, but also lowered the return as well. The Safety Portfolio (when in Risk-off) provides strong returns with limited volatility and drawdown, providing an annual return of about 7% as noted in Table 17. 

Capital is allocated to the Safety portfolio based on a similar momentum scoring system (13612W) with about 20% allocated to each Intermediate Treasury, Long-term Treasuries, and Gold (Table 18).

Table 17 – BT Wave Runner Safety Portfolio Summary

Table 18 – BT Wave Runner Average Allocation (Safety Portfolio)

Though we use US equities as a benchmark, we note that based on the holdings throughout BT Wave Runner (50% risk assets and 50% safety assets) based on the movement of the BT Risk Composite Indicator, the more appropriate benchmark should be measured on similar allocation (50% risk assets and 50% safety), which the BT Wave Runner provides a significant improvement over as cumulative value throughout 1985 to 2021 is 121 times.

In Table 19, we note similar max drawdowns were exhibited by both Wave Runner and Appropriate Benchmark with static allocations at around 25% of capital (both portfolios provide similar improvements relative to risk asset drawdowns). 

Table 19 – BT Wave Runner vs Risk Appropriate Benchmark

The overall returns of the Wave Runner portfolio are also significantly non-correlated with US equity markets (see Table 20), more so than the BT Momentum portfolio, given the low-to-negative correlation of the Safety Portfolio to equity markets.

Table 20 – Rolling Correlations – BT Wave Runner vs US Equities

The combination of fundamental and market risk sentiment data as risk control, incorporated with momentum has improved the risk-adjusted returns of the portfolio by riding the liquidity and momentum waves. Who said riding waves isn’t fun? 

If you are interested in this portfolio strategy or had any questions, feel free to contact us at beowulftreasury@gmail.com.

BT