If you’ve been reading our blog for a number of years you’re 1) probably a finance geek, and you’re 2) probably tired of us discussing the following themes:
- Value investing: buy cheap stocks (see our research category on value investing)
- Momentum investing: buy stocks with strong relative strength (see our research category on momentum investing)
- Trend-following: invest with trends (see our research category on trend following)
We get it. Discussing permutations on the same three themes can get boring, but one thing is clear from years of reading/writing/discussing research: value, momentum, and trend are arguably the undisputed factor kings. Few phenomena have the empirical breadth and depth to match these subjects.
We have developed our own algorithms to capture these three exposures and we explain our research via books and our index methodology via detailed blogs. Are the Indexes we create the infallible holy grail of investing? Definitely not: there are many great ideas out there and we don’t stake a claim to being perfect. Moreover, our Indexes are designed for a unique segment of investors that are long-term oriented, willing to invest the time to understand our processes and have the courage to sometimes be vastly different than the crowd. In short, we’re not for everyone.
Our Value Momentum Trend Indexes simply represent a transparent and systematic approach to combining focused value, momentum, and trend exposures into one unified process. We hope this piece helps investors better understand our Value Momentum Trend Indexes. Please leave questions related to these Indexes in the comments so we can address them in a public forum.
Note: For those who want to dive right into the specifics of the Indexes, information is available below (for compliance purposes, we require a quick registration to access):
An Introduction to the Value Momentum Trend Indexes
The mission of the Value Momentum Trend Indexes is the following:
- We seek to capture global value and momentum equity exposures while simultaneously minimizing the chance of extreme losses.
There are two versions of the Value Momentum Trend Indexes:
- Global Value Momentum Trend Index (GVMT_INDEX): unlevered equity exposure with a dynamic market exposure ranging from 0% to 100%.
- Alpha Architect Long/Short Index (AA L/S_INDEX): levered equity exposure with a dynamic market exposure ranging from 0% to 100%.
Here is an approximate breakdown of the model exposures:
- GVMT Index
- AA L/S Index(1)
The Indexes seek to deliver a diversified portfolio exposure, while simultaneously solving some of the known issues associated with many of the strategies available in the alternatives space. The chart below outlines some of the costs and benefits of the Indexes:
Performance figures contained herein are hypothetical, unaudited and prepared by Alpha Architect, LLC; hypothetical results are intended for illustrative purposes only. Past performance is not indicative of future results, which may vary. Index returns are for illustrative purposes only and do not represent actual fund performance. Index performance returns do not reflect any management fees, transaction costs, or expenses, which would reduce returns. Indexes are unmanaged and one cannot invest directly in an index. Concentration increases the index exposure to individual stock volatility.
Understanding the Value Momentum Trend Index Processes
The index process can be broken into three distinct aspects, which we explore in-depth so readers have a clear understanding of the approach. Here are the three elements:
- Global value and momentum equity
- Trend-following model
- Putting it all together
Here is a visual depiction of the process:
Element #1: Global Value and Momentum Equity
Generic Value and Momentum Summary Statistics
Historically, there have been strong return premiums associated with value and momentum equity investments. Systematic value strategies simply buy cheap stocks based on a price to fundamental ratio (e.g., P/E). Systematic momentum strategies buy stocks that have performed the best relative to other stocks (e.g., relative strength measured over 12 months). For example, in the chart below we show the invested growth of a classic cheap stock portfolio using data from Ken French’s website. We map the growth of a $100 invested in the top decile of cheap book-to-market firms (annually rebalanced) and the S&P 500 Total Return Index.(2)
Systematic value has done well over time, but as we’ve discussed in the past, the short-term relative performance pain is often extremely difficult to digest.
Momentum, like value, has also done well over time:(3)
Just like value, momentum “works,” but these processes require that an investor endure horrific bouts of relative underperformance.
Bottom line: systematic value and momentum investing have worked, but they are long-term commitments.
We Deploy Differentiated Value and Momentum
As was previously highlighted, systematic value and momentum strategies have historically shown promise. We spent many years identifying ways in which we could improve these strategies and have developed our own approaches to exploit value and momentum premiums. We describe these processes in great detail (value here, momentum here), but the basics are outlined below:
Quantitative Value Index: We seek to buy the cheapest highest quality value stocks
- Identify Investable Universe: We typically generate 900 names in this step of the process.
- Forensic Accounting Screens: We usually eliminate 100 names, bringing the total to 800 stocks.
- Valuation Screens: Here we screen the cheapest 10% of the universe, or 80 stocks.
- Quality Screens: We calculate a composite quality score and eliminate the bottom half, leaving 40 stocks.
- Invest with Conviction: We invest in our basket of 40 stocks that are the cheapest, highest quality value stocks.
Portfolio construction characteristics:
- ~40 stocks
- Quarterly rebalanced (international is semi-annually rebalanced)
- 25% sector/industry constraint
- No financials
- Pre-trade liquidity requirements
Quantitative Momentum Index: We seek to buy stocks with the highest quality momentum
- Identify Universe: We typically generate ~ 1,000 names in this step of the process.
- Core Momentum Screen: Select the top decile of firms on their past momentum, or 100 stocks.
- Momentum Quality Screen: Select high-momentum firms with the smoothest momentum, 50 stocks or 50%.
- Seasonality Screen: Rebalance the portfolio near the beginning of quarter-end months.
- Invest with Conviction: We invest in our basket of 50 stocks with the highest quality momentum.
Portfolio construction characteristics:
- ~50 stocks
- Equal-weight construction
- Quarterly rebalanced
- 25% sector/industry constraint
- Pre-trade liquidity requirements
We leverage our heavily researched quantitative value and momentum index methodologies to capture focused value and momentum premiums that we believe have the potential to generate excess returns in the future.
How We Determine our Quantitative Value and Momentum Exposures?
We have 2 Quantitative Value Indexes: US and International. We also have 2 Quantitative Momentum Indexes: US and International. Each index contains 40-50 stocks, so when combined into a global value and momentum portfolio there are generally 180-200 stocks in the overall portfolio. But how do we allocate across the 4 Indexes? The easiest solution is to equal-weight across all four. This is a fine approach, but we decided to go with a volatility-weighted approach because we want to equalize our risk exposure (as opposed to dollar exposure) across the 4 Indexes.(4) Our process for volatility weighting is often deemed “simple risk parity.”(5) The downside of volatility-weighting the positions is an added layer of complexity, but our approach is 1) simple, and 2) annually rebalanced to minimize the noise/frictional costs of complex allocation strategies. In general, our strategy can expect to see a small dollar-weighted tilt towards value. For example, the value exposure might be 55% and the momentum exposure might be 45%, whereas an equal-weight approach would sit at 50/50.
Element #2: Trend-Following Model
The Problem with Large Tail Risks
Timing the market is very difficult. Everyone wants to get high returns with no risk, but this is impossible to achieve. However, while achieving zero risk with equity-like positive returns is a pipe dream, there is some evidence that extreme risk (e.g., massive drawdowns) can be avoided via trend-following methodologies. Why might the avoidance of large tail risks be desirable?
The chart below explains why:
The chart on the left highlights the problem with permanent loss of capital events. A 50% loss of capital needs 11 years of compounding at 7% to beak even. Some investors won’t be around in 11 years, let alone have the ability to sit tight for a decade just to break even. Clearly, huge losses need to be avoided. But as we mentioned, there is no such thing as getting returns and avoiding all risk. But “risk,” if it is defined as volatility, may not be a huge issue. The chart on the right makes this a bit more clear. In this scenario, the investor eats a 15% loss. Definitely not fun, but not devastating. The portfolio only needs 3 years to dig out of its hole and get back to breakeven. 3 years of sitting on your hands is painful — no doubt — but a lot better than 11 years.
So what’s the lesson?
Well, we must assume that in order to get any returns we have to take on volatility — this is reality. But the intensity of the volatility matters. The ideal system would not eliminate volatility — this is impossible — but one thing is required for a good system and that is tail risk management…try and avoid the 50% losses!
One way of achieving risk management is via diversification. Diversification is a no-brainer for all investors — we obviously need to be diversified. But as 2008 highlighted for many investors, diversification won’t necessarily eliminate tail risk. Can anything solve this problem? We’ve looked at just about everything one can imagine, from the most complex ideas to the simplest ideas. Based on our collective take on the independent external research and extensive internal research, trend-following seems to be the most promising way to minimize extreme tail-risks on a portfolio, while simultaneously not hurting long-term expected returns. Trend-following won’t eliminate volatility — it may increase volatility — but if tail risks are your concern, this class of models seem reasonable.(6)
Details for Our Trend-Following Model
Our Trend-Following Model is further described in our empirical analysis, but here are the high-level details:
The mechanics of the rules are outlined below:
- Absolute Performance Rule: Time Series Momentum Rule (TMOM)
- Excess return = total return over past 12 months less return of T-Bills
- If Excess return >0, go long risky assets. Otherwise, hedge the position by shorting the passive index (S&P 500 and/or EAFE).
- A concept made popular by Gary Antonacci
- Trending Performance Rule: Simple Moving Average Rule (MA)
- Moving Average (12) = average of 12 monthly prices
- If Current Price – Moving Average (12) > 0, go long risky assets. Otherwise, hedge the position by shorting the passive index (S&P 500 and/or EAFE).
- A concept made popular by Meb Faber
- Our Trend-Following System: Combination of TMOM and MA (ROBUST model)
- 50% TMOM, 50% MA
Below we show a chart of the 10 largest US drawdowns over a 200+ year period in domestic equity. We examine the 10 worst buy-and-hold drawdowns periods and show the corresponding performance of the equity drawdown when our Trend-Following Model in place.(7)
The trend system was effective at cutting down the pain (although not eliminating) for some of the massive drawdowns of the domestic market over the last 200+ years. The system doesn’t eliminate risk, but these systems do seem to show effectiveness is minimizing the chance of riding an intense pain train.
Element #3: Putting it All Together
We know how our global value and momentum equity book operates: we focus on our sleeves of value (Domestic and International Quantitative Value Indexes) and momentum (Domestic and International Quantitative Momentum Indexes) and allocate to these focused exposures in such a way that we equalize the risk across the 4 portfolios.
We also know how the Trend-Following Model works: Look at the 2 trend-following rules, collect their signals, and assign a 50% weight to each.
But how do we combine this into an integrated system?
There are two variations on the Value Momentum Trend Index methodology: the GVMT Index version and the AA L/S Index version.(8)
Here is an illustration to highlight the mechanics of how we put the system together:(9)
The Index always has 100% of the capital deployed into the global value and momentum equity portfolio and our trend-system manages the overall level of market exposure. Let’s understand how the Trend-Following System is applied:
There are essentially three states of the world:
- No hedge (“Long”)
- Partial hedge (“Hedged”)
- Full hedge (“Market Neutral”)
If no rules are triggered we operate as a long-only exposure that is 100% long global Quantitative Value and Quantitative Momentum stocks.
If one rule is triggered, but the other is not, we are partially hedged and operate similarly to a long/short hedge fund. We are 100% long global Quantitative Value and Quantitative Momentum stocks and 50% short the passive index. In effect, we are 50% exposed to the general market and running the global Quantitative Value and Quantitative Momentum stocks versus passive spread bet for 50% of the portfolio.
If both rules are triggered, we go to a fully hedged stance and essentially operate as a market neutral hedge fund that is 100% long global Quantitative Value and Quantitative Momentum stocks and 100% short a passive index (e.g., S&P 500 and/or EAFE). In other words, we have no market exposure (on a dollar basis) and we are spread betting between the global Quantitative Value and Quantitative Momentum stocks and the underlying passive indexes.
AA L/S Index
The AA L/S version of our Value Momentum Trend process is similar to the GVMT Index, but with a twist. Here is an illustration to highlight the mechanics of how AA L/S functions at a high-level:
Unlike the GVMT Index, the AA L/S Index always has 150% of the capital deployed into the global Quantitative Value and Quantitative Momentum portfolio and has a 50% short position in global passive equity. This is achieved via leverage, which increases the risk and volatility of the index program. A simple way to think about AA L/S versus the GVMT Index is that AA L/S is essentially the GVMT Index, but with a 50% allocation to a market neutral “hedge fund” bet between global Quantitative Value and Quantitative Momentum stocks and global passive equity. Similar to the GVMT Index, we leverage our trend-system to manage the overall level of market exposure.
Here’s how the AA L/S trend-following overlay works:
- 150% long, 50% short. No hedge (“Long”)
- 150% long, 100% short. Partial hedge (“Hedged”)
- 150% long, 150% short. Full hedge (“Market Neutral”)
If no rules are triggered we operate as a long-only exposure that is 150% long global Quantitative Value and Quantitative Momentum stocks and 50% short passive global equity. The net market exposure is 100%, but there is an additional 50% bet between the global Quantitative Value and Quantitative Momentum stocks and passive global stocks.
If one rule is triggered, but the other is not, we are partially hedged and operate similarly to a long/short hedge fund. We are 150% long global Quantitative Value and Quantitative Momentum stocks and 100% short the passive index. In effect, we are 50% exposed to the general market and running a leveraged bet between global Quantitative Value and Quantitative Momentum stocks and the global passive portfolio for 100% of the portfolio.
If both rules are triggered, we go to a fully hedged stance and essentially operate as a market neutral hedge fund that is 150% long global value and momentum and 150% short a passive index (e.g., S&P 500 and/or EAFE). In other words, we have no market exposure (on a dollar basis) and we have a leveraged spread bet between the global Quantitative Value and Quantitative Momentum stocks and the underlying passive indexes.
Understanding the Mechanics of the Hedge
The Trend-Following Model outlined above is a simplified explanation of how the process is applied in the Indexes. In practice, there is a small wrinkle that deserves further explanation: we hedge domestic exposures and the international exposures, separately. And while the two markets are correlated, there are periods when the Index can be fully hedged on the domestic exposure, but long-only on the international exposure (and vice versa). Mechanically, the hedge signals are calculated as follows:
- Domestic Equity Hedge: Each month end, calculate the TMOM and MA rule (described above) on the S&P 500 Total Return Index.
- This signal applies to the domestic exposure within the Index.
- International Equity Hedge: Each month end, calculate the TMOM and MA rule (described above) on the MSCI EAFE Total Return Index.
- This signal applies to the international exposure within the Index.
A hypothetical example might be illustrative of how the Index hedging would work.
- It is at the close on March 31, 2018 and we are determining the hedge exposures for the Indexes that will be in effect April 1, 2018
- 55% in domestic equity (25% Quantitative Momentum, 30% Quantitative value)
- 45% in international equity (20% International Quantitative Momentum and 25% International Quantitative Value)
- 12-month cumulative total return on treasury bills are 2%, S&P 500 Total Return Index is 5%, and MSCI EAFE Total Return Index is -1%.
- S&P 500 TR Index is above the 12-month moving average, MSCI EAFE Total Return Index is above the 12-month moving average
Let’s first look at the domestic equity trend-following rules:
- Domestic TMOM: The S&P 500 TR Index has earned 5% relative to the treasury bill return of 2%. No rule trigger.
- Domestic MA: The S&P 500 TR Index is above the 12-month moving average. No rule trigger.
—> Domestic equity exposure is in a “no hedge” status and we are 100% long for that component of the Index (55% of the portfolio is fully exposed based on the example)
Let’s first look at the international equity trend-following rules:
- International TMOM: The MSCI EAFE TR Index has earned -1% relative to the treasury bill return of 2%. Rule triggered.
- International MA: The MSCI EAFE TR Index is above the 12-month moving average. No rule trigger.
—> International equity exposure is in a “partial hedge” status and we are 50% long for the international components of the Index (45% long the underlying, but short 22.5% via MSCI EAFE TR Index).
In summary, the hypothetical portfolio will look as follows:
GVMT Index version
- 55% domestic equity; 25% Quantitative Momentum; 30% Quantitative Value
- No domestic exposure hedging
- 45% international equity; 20% International Quantitative Momentum; 25% International Quantitative Value
- 22.5% short position in MSCI EAFE TR Index tracking vehicle (ETF or future)
- Summary: 100% global value and momentum equity exposure; 22.5% short international passive equity
AA L/S Index version
- 82.5% domestic equity (1.5*55%); 37.5% Quantitative Momentum (1.5*25%); 45% Quantitative Value (1.5*30%)
- Permanent 27.5% (55%/2) short position in passive domestic equity
- No additional domestic exposure hedging
- 67.5% international equity (1.5*45%); 30% International Quantitative Momentum (1.5*20%); 37.5% International Quantitative Value (1.5*25%)
- Permanent 22.5% (45%/2) short position in passive international equity
- An additional 22.5% short position in MSCI EAFE TR Index tracking vehicle (.5*45%)
- 45% (22.5%+22.5%) short passive international hedge (in total)
- Summary: 150% global value and momentum equity exposure; 27.5% short domestic passive equity; 45% short international passive equity
Historically, here is the Index exposure between Domestic and International from 1995 to 2017 (AA L/S is approximately 1.5x the percentages in the chart below):(10)
Here is a breakdown of domestic buy-and-hold versus domestic trend-followed exposures over time:
Here is a breakdown of international buy-and-hold versus international trend-followed exposures over time:
In addition to the analysis above, we have some details on the cost and benefits of our hedging approach in the appendix.(11)
The Alternative Nature of Our Value Momentum Trend Indexes
Our Value Momentum Trend Indexes try to capture the three factors that we believe are the most effective: Value, Momentum, and Trend. Because of the focused exposures on these factors, these indices will not follow broad passive equity exposures that closely. To this point, here are the correlations between the Indexes and the passive domestic and international benchmarks from 1995 to 2017:
Performance figures contained herein are hypothetical, unaudited and prepared by Alpha Architect, LLC; hypothetical results are intended for illustrative purposes only. Past performance is not indicative of future results, which may vary. Index returns are for illustrative purposes only and do not represent actual fund performance. Index performance returns do not reflect any management fees, transaction costs, or expenses, which would reduce returns. Indexes are unmanaged and one cannot invest directly in an index.
In general, the indexes are roughly 50% correlated with long-only passive exposures.(12)
There are many periods where the Indexes outperform the long-only global passive index and there are also many periods where the Indexes underperform the generic global index. So while our Value Momentum Trend Indexes may be misperceived as “equity” because they are sometimes long-only, we want to emphasize that these exposures can act very differently than long-only equity. Investors following these Indexes should keep this in mind when attempting to benchmark these systems against long-only benchmarks (arguably, a 50% treasury bill + 50% global passive equity benchmark is most appropriate).
Conclusions Regarding the Value Momentum Trend Index Process
We’ve spent a lot of time developing systematic processes that seek to capture the value premium (i.e., Quantitative Value Indexes), the momentum premium (i.e., Quantitative Momentum Indexes), and seek to prevent large drawdowns via our systematic trend-following process. Each of our processes can be used in specialized cases and to solve specialized problems. However, our Value Momentum Trend Indexes seek to package our best ideas into one approach. For some investors, this may be a more appropriate approach, as opposed to utilizing the individual value, momentum, and trend components. We leave this up to the investor and/or their advisors to determine.
In the end, the Value Momentum Trend Index processes seek to deliver focused exposure in what we believe are the three most robust factors in investing: value, momentum, and trend.
Buy ’em cheap; buy ’em strong; and hold ’em long…but only when the trend is your friend
— Wesley R. Gray and Jack R. Vogel
Information on our Value Momentum Trend Indexes is available here.
Here are some specific research/educational materials:
- The views and opinions expressed herein are those of the author and do not necessarily reflect the views of Alpha Architect, its affiliates or its employees. Our full disclosures are available here. Definitions of common statistics used in our analysis are available here (towards the bottom).
- Join thousands of other readers and subscribe to our blog.
- This site provides NO information on our value ETFs or our momentum ETFs. Please refer to this site.
References [ + ]
|1.||↑||For a $100 investment, the strategy uses leverage to purchase $150 worth of value and momentum securities, while holding a constant $50 broad-market short position in the portfolio. So $100 is value and momentum beta and $50 is value and momentum alpha, as there is a bet that Value and Momentum securities will outperform the broad market. A more detailed explanation is below in the text.|
|4.||↑||This is especially important because our Quantitative Momentum Indexes are substantially more volatile than our Quantitative Value Indexes|
|5.||↑||see risk parity for dummies|
|6.||↑||Here is a post where we examined domestic equity returns from 1/1/1801 until 9/30/15 using our trend system.|
|7.||↑||Details on this study are found here.|
|8.||↑||Note that T = # of trend following rule triggers. Rule #1: Time-series trend; Rule #2: Moving Average trend|
|9.||↑||Note that we apply the trend rules to the U.S. allocation and International allocation separately, as described below in the text.|
|10.||↑||These are the weights (U.S. or International) to the long book of the risk-parity weighted portfolio|
Why Do We Hedge As Opposed to Going to Cash?
Let’s say a trend rule triggers and we now need a short position to hedge our long portfolio. For illustration purposes, let’s also assume we need to fully hedge the portfolio (“Full Hedge”).
What happens next?
As we discussed above, the GVMT INDEX is always long a portfolio of Value and Momentum stocks. (Note the hedging for the AA L/S Index is similar, but not the same. For example, we may use futures for the short position, which can be more tax efficient over time, in the AA L/S Index.) Let’s say the Index value is $100. If the Index wants to short $100 of a passive index to maintain a fully hedged stance (Note one can also use futures), typically done via an ETF vehicle such as the SPY or EFA, the Index can rely on the fully paid securities ($100) as collateral to open a short index ETF position.
The Index could also sell all the Value and Momentum stocks to minimize market risk on the portfolio and invest the proceeds in cash or Treasury Bills.
So why did we choose to have our Index always own the underlying Value and Momentum stocks and hedge via a passive instrument, as opposed to selling down the equity?
Aside from the greatly enhanced frictional and tax costs of trading in and out of the holdings, there is a potential investing benefit of maintaining the underlying stock positions during a hedging event. Please see our detailed education materials for this analysis.
Operational Costs of Hedging
To open a short position, a prime broker will assess the collateral and decide the margin requirement. Typically professional investors access what is referred to as portfolio margin (as opposed to Reg-T margin), which accounts for portfolio characteristics when assessing margin requirements. This is important in our context because the margin requirements to enter a short position, which is negatively correlated with our long positions, means that our margin requirements will often be much less than 50% (Reg-T requirement). Let’s say the prime broker determines that the margin requirement to hold the short position is 20%. In this case, the portfolio would need to borrow $20 ($100*20%) from the prime broker, and will pay an interest rate on this borrowed amount. The interest rate is typically, Borrow Rate = Fed Funds + markup, where the markup ranges across brokers.
Once the prime broker determines there is the necessary collateral in place, the portfolio can initiate the short ETF position, and will receive $100 in cash from the short sell. The portfolio will pay a cost to borrow the position (often called the short rebate) to maintain the position (SPY is around 30-35bps/year), but the portfolio will also receive interest on the $100 in short proceeds. The interest rate received on the short position is typically the Fed Funds rate minus some haircut.
So in total, the carry cost to the portfolio to initiate the short position is given below:
Cost = (Margin Requirement %)*($100)*(Fed Fund Rate + Markup) + ($100)(Short Rebate Cost) – ($100)*(Fed Funds Rate – haircut).
Let’s say the margin requirement is 20%, Fed Funds is 1%, the markup/haircut are 50bps, and the rebate is 30bps. The overall carry cost of the short position will be as follows:
Annual Costs = 20%*$100*(1.5%) + $100*0.30% – $100*(0.50%) ~ 0.10% or 10bps.
In other words, there is a 10bps negative carry for holding the short position. In general, with higher interest rates the carry becomes more positive and with lower interest rates, holding a short position will have a negative carry.
Last, it should be pointed out that the portfolio, like any other portfolio that engages in short-selling or other derivative transactions, does take on some counterparty risk by employing such a strategy.
|12.||↑||In more detailed analysis we examine the beta of our Indexes against a variety of factor models (CAPM, FF 3-factor, Carhart 4-Factor, FF 5-factor, and AQR 6-factor) and confirm that the beta is typically around .5.|