How to Use Trend Following within a Portfolio

A question we have been receiving recently is the following:

How should I use trend following within a portfolio?

Generally, the questions are related to our Global Value, Momentum, and Trend Index, which allocates to the (1) Value, (2) Momentum, and (3) Trend factors.

A big difference between the Global Value Momentum Trend (GVMT) portfolio and many other “smart-beta” products is the inclusion of the trend factor–a way to dynamically adjust market exposure based on expected tail risk. We think there are valid empirical and behavioral reasons to use trend-following, but we understand that many financial advisors do not use trend-following within their diversified portfolio. In other words, trend following is a new wrinkle and many advisors are unclear on how to deploy trend following in a portfolio.

An assumption that I am going to make is that most advisors/investors have a standard stock/bond portfolio, such as 60% stocks and 40% bonds.⁽¹⁾ Below we examine where trend following can fit into a portfolio, but first, we examine some trend-following facts.

Some Basic Trend Following Facts

Below I examine trend following on global equities, specifically the U.S. market (SP500) and developed international markets (EAFA).⁽²⁾

The portfolios and their returns are shown first without trend, and then with trend-following.

The returns below are from 1/1/1973-12/31/2017. All asset classes apply a 12-month moving average rule to the total return index, and either invest (1) in the index if a positive trend or (2) in cash (U.S. T-bills) if a negative trend. All portfolios are total returns and gross of any transaction or management fees.

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index.

As is shown above, the overall annualized returns to trend-followed portfolios, over long-time cycles, are similar to B&H but with lower drawdowns.⁽³⁾

To examine the relationship of trend-followed indexes with the B&H indicies, I show the correlations below:

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index.

The correlations above highlight that B&H and trend following are positively correlated.⁽⁴⁾

However, it should be pointed out, that on a relative basis trend-following does not work all the time!

To highlight this fact, I show the returns from 1/1/2010-12/31/2017. All portfolios are gross of any transaction or management fees.

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index.

The results above highlight that over the past 8 years, a trend-followed portfolio has underperformed the passive B&H benchmarks. However, one may still be interested in trend-followed strategies in hopes of lower drawdowns within a portfolio.

How to Use Trend-Following within a Portfolio?

So how does trend following fit within a portfolio?

Most advisors/investors have a standard stock/bond portfolio, such as 60/40 or 80/20. How much one allocates to stocks/bonds generally is determined by an investors risk tolerance. Since bond returns are less volatile than stock returns, investors allocate more to bonds if they prefer to have less “risk”, commonly measured by the standard deviation of the portfolio returns. However, a dynamic trend following strategy also allocates between stocks and bonds (short-duration T-bills), but does so in a dynamic fashion. Similarly, the trend following strategy is focused on the “risk” of the portfolio but does so by focusing on the maximum drawdown. From my analysis here, I find that over the 1973-2017 time period, one would have been invested in the SP500 ~76% of the time and EAFE ~70% of the time (while being in T-bills/cash ~24% of the time for the U.S. and 30% of the time for International).⁽⁵⁾

So building off standard portfolios, how did allocating a portion of the portfolio to trend following perform in the past?

To study this, I show the returns to 4 standard U.S. portfolios below:

1. 40% Sp500, 60% T-Bonds — A monthly rebalanced portfolio investing 40% in the SP500 and 60% in 10-Year Treasuries.
2. 60% Sp500, 40% T-Bonds — A monthly rebalanced portfolio investing 60% in the SP500 and 40% in 10-Year Treasuries.
3. 80% Sp500, 20% T-Bonds — A monthly rebalanced portfolio investing 80% in the SP500 and 20% in 10-Year Treasuries.
4. Sp500 — Investing 100% in the SP500.

Returns are from 1/1/1973-12/31/2017 and all portfolios are gross of any transaction or management fees.

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index.

As one notices, adding bonds into the portfolio (1) lowered the standard deviation and drawdowns, while also (2) lowering the returns.

So how can one add the trend factor to a portfolio? Below I examine two ways one can add the trend factor:

1. Give a 90% weight to the standard portfolio and 10% weight to trend following.
2. Take 10% weight away from the bond complex and allocate towards trend following.

Taking the 60/40 portfolio as an example, if I allocate 90% to this portfolio and 10% to trend following, I come up with the weights being 54% stocks, 36% bonds, and 10% trend. I do the same for the other portfolios.

Below, I generate the returns to the 4 portfolios:

1. 36% Sp500, 54% T-Bonds, 10% Trend (SP500) — A monthly rebalanced portfolio investing 36% in the SP500, 54% in 10-Year Treasuries, and 10% in a trend following portfolio on the SP500.⁽⁶⁾
2. 54% Sp500, 36% T-Bonds, 10% Trend (Sp500) — A monthly rebalanced portfolio investing 54% in the SP500, 36% in 10-Year Treasuries, and 10% in a trend following portfolio on the SP500.
3. 72% Sp500, 18% T-Bonds, 10% Trend (SP500) — A monthly rebalanced portfolio investing 72% in the SP500, 18% in 10-Year Treasuries, and 10% in a trend following portfolio on the SP500.
4. 90% Sp500, 10% Trend (SP500) — A monthly rebalanced portfolio investing 90% in the SP500, and 10% in a trend following portfolio on the SP500.

Returns are from 1/1/1973-12/31/2017 and all portfolios are gross of any transaction or management fees.

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index.

Comparing the results to the “standard” portfolios above, one notices that (1) the standard deviations and drawdowns are slightly lower when adding trend following, while (2) the portfolio returns are slightly higher when trend following. Once again, remember that trend following is dynamically investing in 2 assets (1) Stocks (SP500) or short-duration bonds (T-Bills).

An alternative, but riskier proposition⁽⁷⁾ would be to reduce the bond allocation and transition a portion of that portfolio to trend following. I do this by taking away a 10% allocation to bonds and add a 10% allocation to trend following.

Below, I generate the returns to the 4 portfolios:

1. 40% Sp500, 50% T-Bonds, 10% Trend (SP500) — A monthly rebalanced portfolio investing 40% in the SP500, 50% in 10-Year Treasuries, and 10% in a trend following portfolio on the SP500.⁽⁸⁾
2. 60% Sp500, 30% T-Bonds, 10% Trend (Sp500) — A monthly rebalanced portfolio investing 60% in the SP500, 30% in 10-Year Treasuries, and 10% in a trend following portfolio on the SP500.
3. 80% Sp500, 10% T-Bonds, 10% Trend (SP500) — A monthly rebalanced portfolio investing 80% in the SP500, 10% in 10-Year Treasuries, and 10% in a trend following portfolio on the SP500.
4. SP500 — Investing 100% in the SP500.

Returns are from 1/1/1973-12/31/2017 and all portfolios are gross of any transaction or management fees.

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index.

Comparing the results to the “standard” portfolios above, one notices that (1) the standard deviations and drawdowns are generally higher, while (2) the portfolio returns are higher as well. This makes sense, as we reduced the bonds allocation (which helps to reduce the portfolio standard deviation) in exchange for a higher stock allocation (with trend following).

So now that we have the evidence on the U.S. portfolios, what happens to portfolios that have a mix of U.S. and International stocks?

Here we will use the MSCI World Index to proxy for global (developed) stock returns while keeping the bond complex the U.S. 10-year Bond. Note, that we could have selected other bond portfolios, but here we want to focus on the global stock portfolio.⁽⁹⁾

To study this, I show the returns to 4 portfolios below:

1. 40% World Stocks, 60% T-Bonds — A monthly rebalanced portfolio investing 40% in the MSCI World Index and 60% in 10-Year Treasuries.
2. 60% World Stocks, 40% T-Bonds — A monthly rebalanced portfolio investing 60% in the MSCI World Index and 40% in 10-Year Treasuries.
3. 80% World Stocks, 20% T-Bonds — A monthly rebalanced portfolio investing 80% in the MSCI World Index and 20% in 10-Year Treasuries.
4. World Stocks– Investing 100% in the MSCI World Index.

Returns are from 1/1/1973-12/31/2017 and all portfolios are gross of any transaction or management fees.

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index.

One again notices that the higher the bond allocation, the lower the portfolio standard deviation and the drawdowns. One also notices that compared to a 100% U.S. Stock portfolio (SP500 above), the overall returns to a global (developed) stock allocation were lower over this time period.

To examine the impact of trend following when being added to the portfolio, I will now use a 50/50 allocation to the (1) Sp500 with trend following and (2) EAFE with trend following, as described above.

Since we are examining a global developed stock portfolio, we should add in trend following on both the U.S. and developed markets.

Below, I generate the returns to the 4 portfolios:

1. 36% World Stocks, 54% T-Bonds, 10% Trend (SP500 & EAFE) — A monthly rebalanced portfolio investing 36% in the MSCI World Index, 54% in 10-Year Treasuries, and 10% in a trend following portfolio on the SP500 and EAFE (50/50 allocation).⁽¹⁰⁾
2. 54% World Stocks, 36% T-Bonds, 10% Trend (SP500 & EAFE) — A monthly rebalanced portfolio investing 54% in the MSCI World Index, 36% in 10-Year Treasuries, and 10% in a trend following portfolio on the SP500 and EAFE (50/50 allocation).
3. 72% World Stocks, 18% T-Bonds, 10% Trend (SP500 & EAFE) — A monthly rebalanced portfolio investing 72% in the MSCI World Index, 18% in 10-Year Treasuries, and 10% in a trend following portfolio on the SP500 and EAFE (50/50 allocation).
4. 90% World Stocks, 10% Trend (SP500 & EAFE) — A monthly rebalanced portfolio investing 90% in the MSCI World Index, and 10% in a trend following portfolio on the SP500 and EAFE (50/50 allocation).

Returns are from 1/1/1973-12/31/2017 and all portfolios are gross of any transaction or management fees.

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index.

Similar to the analysis above within the U.S., adding trend following, as a 10% allocation alongside a standard portfolio, generally (1) increased returns, (2) lowered the standard deviation, and (3) lowered the portfolio maximum drawdown.

Now let’s examine what happens when we simply lower the bond allocation of the portfolio.

Below, I generate the returns to the 4 portfolios:

1. 40% World Stocks, 50% T-Bonds, 10% Trend (SP500) — A monthly rebalanced portfolio investing 40% in the MSCI World Index, 50% in 10-Year Treasuries, and 10% in a trend following portfolio on the SP500 and EAFE (50/50 allocation).⁽¹¹⁾
2. 60% World Stocks, 30% T-Bonds, 10% Trend (Sp500) — A monthly rebalanced portfolio investing 60% in the MSCI World Index, 30% in 10-Year Treasuries, and 10% in a trend following portfolio on the SP500 and EAFE (50/50 allocation).
3. 80% World Stocks, 10% T-Bonds, 10% Trend (SP500) — A monthly rebalanced portfolio investing 80% in the MSCI World Index, 10% in 10-Year Treasuries, and 10% in a trend following portfolio on the SP500 and EAFE (50/50 allocation).
4. SP500 and EAFE with trend — A monthly rebalanced portfolio investing 50% in the SP500 with trend following and 50% in the EAFE index with trend following.

Returns are from 1/1/1973-12/31/2017 and all portfolios are gross of any transaction or management fees.

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index.

As with the analysis within the U.S., compared to the standard portfolios (60/40, etc.), reducing the bond allocations and adding trend (1) increases the returns but (2) also increases the standard deviation and maximum drawdowns. To give a reference point, I added the SP500 and EAFE with trend portfolio, so one can view the returns to that portfolio over the same time period, as this is the 10% allocation in the other portfolios. It should be noted, that as shown above, this portfolio can underperform at times.

Conclusion:

The trend factor is definitely different than other factors, such as Value, Momentum, and Quality. This is due to the fact that it can dynamically change the beta of the portfolio from 1 all the way to 0, whereas more “standard” factors will generally have a beta close to 1. So given this fact, how can one add the trend-factor into a standard stock/bond portfolio?

I examined two ways one can add the trend factor:

1. Give a 90% weight to the standard portfolio and 10% weight to trend following.
2. Take 10% weight away from the bond complex and allocate towards trend following.

In the past, both portfolios would have outperformed the standard portfolios, before any fees or transaction costs. However, it should be noted that trend following does not work all the time.

To conclude, while we prefer to allocate 100% to trend following, hopefully, this analysis shows how trend following interacts with a more standard stock/bond portfolio.