This blog post is available as a PDF download here.

Summary­

  • We exploit the idea that long-only strategies are “long/short portfolios all the way down,” we demonstrate how to isolate the active bets of portfolio managers.
  • Using the example of a momentum / low-volatility barbell portfolio, we construct a simple long/short portfolio using ETFs and S&P 500 futures.
  • Recognizing that not all investors will have access to S&P 500 futures, we argue that the capital efficiency provided by a fund like the PIMCO StocksPLUS Short Fund (“PSTIX”) can potentially allow investors to achieve the same ends.
  • By thinking about our aggregate exposure within a portfolio – and not line-item by line-item – we demonstrate how an investor might alter a standard 60/40 portfolio to introduce a 20% allocation to their own equity long/short portfolio.

In this week’s research note, we will explore a simple idea for how advisors and investors can “roll their own” long/short portfolio even if they do not have the ability to short individual securities.

The idea boils down to two simple concepts:

  1. Active portfolios are long/short portfolios all the way down.
  2. Capital efficiency.

The first concept can be quickly summarized as recognizing that a long-only, active portfolio can be thought of as the sum of two components: (1) a market-cap weighted index and (2) the over- and underweight decisions the active manager makes. The over- and underweight decisions form a dollar-neutral long/short portfolio and the notional size of the bets is equal to the active manager’s active share.  When we overlay the long/short portfolio on top of the market-cap weighted index, we are given the long-only portfolio.

We can, therefore, take this idea in reverse: shorting out the market-cap weighted index from a long-only active strategy leaves us with the long/short portfolio.  This idea allows us to take any actively managed portfolio – whether it is a basket of securities we pick or an actively managed fund – and turn it into a long/short portfolio.

There are a number of ways in which this might be achieved.  For example, an investor could short an S&P 500 ETF or could short futures contracts.  Unfortunately, many advisors and individuals do not have this opportunity.

Another option is to hold a long position in an inverse ETF such as the ProShares Short S&P 500 ETF (“SH”).  The downside here is that precious capital is now tied up in a position that effectively returns Treasury Bills minus the S&P 500.  We could try to use a levered short ETF, such as the ProShares UltraShort Short S&P 500 (“SDS”).  Only half the necessary capital would be tied up with this trade, but we’d be introducing the potential for significant tracking error due to compounding effects.  And that’s all to speak nothing of the cost.

For now, we are going to assume that an investor can simply short S&P 500 futures contracts and we’ll return to how an advisor or investor might go about implementing this idea in a more constrained setting.

As an example of constructing a long/short portfolio, we will use the idea of a momentum / low-volatility barbell which has been proposed by Lawrence Hamtil (and who has written about the idea a number of times).  We will use the iShares Edge MSCI USA Momentum Factor ETF (“MTUM”) and the iShared Edge USA Min Vol Factor ETF (“USMV”) for our long factor positions. Prior to ETF launch, we will use index returns.

We will construct our portfolio as a simple 50-50 barbell, holding 47.5% of the portfolio in each ETF and keeping 5% aside as collateral for a 95% short position in S&P 500 futures.

The results are rather strong, both prior to ETF launch and after.  Over the full period, our long/short portfolio returns 3.6% annualized with a volatility of just 5.4% and a maximum drawdown of 11.5%.

Now we should acknowledge here that constructing our portfolio in such a manner means that the notional size of our long/short is constrained by the active share of the ETFs themselves.  For example, MTUM has an active share of 68.1% and USMV has an active share of 57.2%, giving our barbell an active share of 62.6%.  This implies that when we short out our exposure to S&P 500 futures, the long/short portfolio we are left with is really only using 62.6% of capital.

Though this speaks nothing to the volatility of those active bets, investors should be aware that a methodology like this, when used on a low active share strategy will leave very little residual exposure.

Source: CSI Data; MSCI; Stevens Futures; Calculations by Newfound Research.  Results assume the reinvestment of all distributions. Results are gross of all fees, including, but not limited to, manager fees, transaction costs, and taxes. Past performance is not an indicator of future results. 

 

Unless they are explicitly identified as market neutral, most long/short portfolios retain some equity market beta to benefit from the positive expected long-term equity risk premium.  Therefore, it may make more sense to actually construct this portfolio using a target beta approach.

For example, we might calculate the beta of our 50/50 momentum/low-volatility barbell portfolio to the S&P 500 and then increase or decrease our exposure to our short futures position accordingly.  If our barbell has a beta of 1.0 and we only want a beta of 0.5, we could hold 95% of our portfolio in the barbell, 5% in cash, and short 45% S&P 500 futures exposure to hit our target beta.

Below we plot the equity curves for portfolios targeting a beta of 0 through 0.5.  We estimate beta using the historical realized covariance matrix and use an ensemble approach (varying the lookback window of our covariance matrix calculation) to generate weights.

Source: CSI Data; MSCI; Stevens Futures; Calculations by Newfound Research.  Results assume the reinvestment of all distributions. Results are gross of all fees, including, but not limited to, manager fees, transaction costs, and taxes. Past performance is not an indicator of future results. 

 

Unsurprisingly, given our knowledge of history, the highest beta long/short portfolio had the highest return over the last 20 years.  Isolating different periods, however, provides a more nuanced perspective.  For example, we can see that over the last 12 months the five variations all exhibited nearly identical total returns, while the portfolio that targeted zero beta was able to exhibit positivereturns during the market’s Q4 2018 drawdown.

Here we will pause to note a key difference in the first portfolio we constructed and the target beta portfolios.

In the former case, we used the basic math that a long-only active portfolio is equal to a market-capitalization-weighted portfolio plus a dollar-neutral long/short portfolio.  By shorting a dollar equivalentamount of market beta, we could isolate the implied long/short.

The remaining long/short exposure may or may not have residual beta within in.  For example, USMV has a beta of approximately 0.7, meaning that the long/short portfolio component must have a beta of -0.3.  If we hold USMV and an equivalent dollar amount in short S&P 500 exposure, we would generally expect the residual exposure to be negatively correlated to the market (ignoring idiosyncratic returns for a moment).

If, however, we design our portfolio to explicitly be beta neutral, then we might not hold only $0.7 short in S&P 500 exposure for every $1 we hold in USMV. This results in a portfolio that may more explicitly reflect the idiosyncratic returns of the implied long/short active bets taken by USMV.

Source: CSI Data; MSCI; Stevens Futures; Calculations by Newfound Research.  Results assume the reinvestment of all distributions. Results are gross of all fees, including, but not limited to, manager fees, transaction costs, and taxes. Past performance is not an indicator of future results. 

 

Unfortunately, up until this point, implementation of this idea requires shorting S&P 500 futures.  This is where capital efficiency enters the equation.

The PIMCO StocksPLUS Short Fund (“PSTIX”) provides access to an actively managed portfolio of fixed income securities and overlays that portfolio with a 100% S&P 500 futures short position.  On its own, we would have to hold a near dollar-for-dollar amount in our long equity exposures as PSTIX to hedge out beta, making it inefficient.

If, however, we think from a total portfolio balance sheet perspective, things become much more interesting.

Let’s say, for example, that we currently hold a 60% S&P 500 / 40% Bloomberg Barclay’s Aggregate US Bond portfolio and we would like to sell 10% of our equity exposure and 10% of our bond exposure to create a 20% allocation to a momentum/low-volatility barbell long/short portfolio.

To achieve this exposure, here is how we would construct our portfolio:

  • Hold 50% of our portfolio in the S&P 500
  • Hold 10% of our portfolio in MTUM
  • Hold 10% of our portfolio in USMV
  • Hold 20% of our portfolio in PSTIX
  • Hold 10% of our portfolio in AGG

At first, this appears to be a massive reduction in bonds.  However, when we consider that PSTIX provides us exposure to bonds and short S&P 500 exposure simultaneously, the aggregate picture makes more sense.  Below we plot the exposure from each position as well as the net resulting exposure. We can see that we do indeed create a 50% S&P 500 / 30% Bond / 20% Equity Long/Short portfolio.

Conclusion

In this commentary, we introduce a simple idea for advisors and individuals to construct long/short strategies when they do not have the ability to explicitly short within their portfolios.

The idea is built from two foundational concepts: (1) active strategies can be thought of as a passive benchmark plus a dollar-neutral long/short strategy, and (2) exploiting capital efficiency.  By combining long-only positions with short exposure to the S&P 500 (or other market index), we can isolate the implied active under- and overweight positions as a long/short portfolio.

To achieve our short exposure, we can take advantage of a capital-efficient portfolio such as the PIMCO StocksPLUS Short Fund (“PSTIX”), which provides simultaneous exposure to fixed income and a short S&P 500 position.  By thinking holistically about our portfolio, we can replace existing fixed income with PSTIX and create an implied long/short strategy within our portfolio.

This approach is not without its trade-offs.  In our example, creating a 20% long/short allocation results in PSTIX accounting for 2/3rds of the portfolio’s fixed income exposure.  This means we must not only be incredibly comfortable with the portfolio construction of PSTIX, but we must also be comfortable in the foregone opportunity cost to allocate to other fixed income managers.

This approach may also not be effective for investors who do not currently hold much fixed income.  However, for conservative investors wanting to replace existing fixed income exposure with long/short equity, a combination of long-only exposure with PSTIX may be an effective way to take control of the long/short portfolio construction.


As of the date of this document, both Newfound Research and Corey Hoffstein hold positions MTUM and USMV, and Corey Hoffstein holds positions in PSTIX.  Newfound Research and Corey Hoffstein do not take a position as to whether these securities should be recommended for any particular investor.

 

Corey is co-founder and Chief Investment Officer of Newfound Research, a quantitative asset manager offering a suite of separately managed accounts and mutual funds. At Newfound, Corey is responsible for portfolio management, investment research, strategy development, and communication of the firm's views to clients. Prior to offering asset management services, Newfound licensed research from the quantitative investment models developed by Corey. At peak, this research helped steer the tactical allocation decisions for upwards of $10bn. Corey holds a Master of Science in Computational Finance from Carnegie Mellon University and a Bachelor of Science in Computer Science, cum laude, from Cornell University. You can connect with Corey on LinkedIn or Twitter. Or schedule a time to connect.