A PDF version of this post can be found here.

Summary

  • A well-designed investment plan is an important part of achieving investment objectives, but even the best investment plan is useless if you cannot stick to it.
  • Rolling relative performance can give context to the size of short-term portfolio fluctuations while looking at risk exposures can give context to the timing of these fluctuations.
  • Using our Multi-Asset Income strategy’s holdings as of 9/30 as an example, we shed some light on the expected tracking error and the current and potential evolution of risk exposures from a number of different viewpoints.

Investing reminds us of the fitness industry.  People spend inordinate amounts of time debating the merits of the latest diet or workout program.  Yet, the truth is that there are many approaches to fitness that work.  The key is finding one that you can stick to.

Similarly, there are many approaches to investing that have proven track records of success.  Each has its own strengths and weaknesses and there is no holy grail.  Success requires developing a thoughtful plan that is backed by both theory and data and then having the discipline to stick with that plan.

While few focus on it, the discipline of the investor may ultimately be the most important ingredient. Skill is only part of the picture.

In our experience, setting appropriate expectations about how a strategy may behave relative to certain benchmarks can go a long way towards helping investors stay committed to their portfolios.

How can we go about setting appropriate expectations?

As an example, we will explore Newfound’s Multi-Asset Income portfolio, a strategy designed to provide risk-managed access to global high income asset classes.  As of 9/30, its asset class allocations were as follows:

  • Bank Loans: 14.2%
  • Preferreds: 12.1%
  • Emerging Market Bonds (USD): 10.6%
  • Mortgage REITs: 9.9%
  • High Yield Corporates: 9.4%
  • Investment Grade Corporates: 9.2%
  • Convertibles: 7.5%
  • Emerging Market Bonds (Local Currency): 5.3%
  • Covered Calls: 4.1%
  • International Dividend Stocks: 3.5%
  • S. REITs: 3.4%
  • S. Dividend Stocks: 3.2%
  • S. Treasuries: 2.6%
  • International REITs: 2.6%
  • MLPs: 2.4%

For the rest of this commentary, we will refer to this portfolio as the “MAI” portfolio.  In seeking to set expectations for the current portfolio configuration, we will also assume that the allocations have been static over time (i.e. we ignore any tactical changes that our Multi-Asset Income strategy may have made).  Once again, this portfolio is just a snapshot of our holdings at one point in time and does not reflect current positioning.

Perhaps the simplest approach for setting expectations is to just graph how the MAI portfolio performed relative to its benchmark – a 50/50 ACWI/AGG portfolio – over various time horizons.

For weekly returns, the relative performance of the MAI portfolio vs. the 50/50 ACWI/AGG benchmark has ranged from -1.8% to +1.7% over the last 5+ years - the period over which data exists for all the ETFs in the portfolio.

picture1

Data Source: Yahoo! Finance.  Calculations by Newfound Research.  The benchmark represents a 50/50 portfolio of the iShares MSCI ACWI ETF (ticker: ACWI) and the iShares Barclays Aggregate Bond ETF (ticker: AGG).  Data does not include any fees or expenses other than underlying ETF management fees.  Past performance does not guarantee future results.  

As we expand our window for measuring returns, the range of outcomes also grows.  For monthly returns, the relative performance of the MAI portfolio vs. its benchmark ranges from -3.1% to +2.5%.  For quarterly returns, the range is even larger: -5.1% to +4.7%.

picture2

Data Source: Yahoo! Finance.  Calculations by Newfound Research.  The benchmark represents a 50/50 portfolio of the iShares MSCI ACWI ETF (ticker: ACWI) and the iShares Barclays Aggregate Bond ETF (ticker: AGG).  Data does not include any fees or expenses other than underlying ETF management fees.  Past performance does not guarantee future results.  

picture3

Data Source: Yahoo! Finance.  Calculations by Newfound Research.  The benchmark represents a 50/50 portfolio of the iShares MSCI ACWI ETF (ticker: ACWI) and the iShares Barclays Aggregate Bond ETF (ticker: AGG).  Data does not include any fees or expenses other than underlying ETF management fees.  Past performance does not guarantee future results.  

These simple graphs are helpful for managing emotions and setting expectations.  For example, say the MAI portfolio underperforms by 1.0% in a week.  We can very quickly identify that this is well within our expectations for short-term tracking error and is no reason for alarm.

While it does give us a reasonable view into the distribution of potential tracking error, the downside of viewing relative performance through this lens only is that it gives us no insight as to when outperformance or underperformance may occur.

To gain some insight here, we can scatterplot MAI Returns vs. benchmark returns. We do this below using monthly data.

picture4

Data Source: Yahoo! Finance.  Calculations by Newfound Research.  The benchmark represents a 50/50 portfolio of the iShares MSCI ACWI ETF (ticker: ACWI) and the iShares Barclays Aggregate Bond ETF (ticker: AGG).  Data does not include any fees or expenses other than underlying ETF management fees.  Past performance does not guarantee future results.  

Points above the gray line indicate months when the MAI portfolio outperformed.  Points below the gray line indicate months when the MAI portfolio underperformed.

The dashed orange line is the result of a polynomial regression of High Income returns on benchmark returns.  The equation for this trend line is:

screen-shot-2016-11-18-at-10-52-39-pm

What do each of these coefficients mean?  Let’s tackle them in reverse order.

The third coefficient (0.0004) is the intercept and simply means that the MAI portfolio outperformed the benchmark by 4bps per month over our sample period.

The second coefficient (0.9445) can be thought of as the average beta of the MAI portfolio relative to the benchmark.  On average, the MAI portfolio responds slightly less than one-for-one to moves in the benchmark.

The third coefficient (1.4743) is multiplied by the squared benchmark return.  The fact that this number is positive implies that the MAI portfolio has exhibited positive convexity to the benchmark.  Positive convexity is generally a good thing since it causes upside capture to be greater than downside capture.  Practically, this explains why the MAI portfolio outperformed the benchmark in the extremes (i.e. when the benchmark return is very high or very low).

Another way we can characterize this behavior is by calculating conditional beta.  In this context, conditional beta will represent the beta of the MAI portfolio to the benchmark for a specific subset of the sample data.  We see that when benchmark returns are low (in the bottom 25%), the beta of the MAI portfolio is 0.61.  When benchmark returns are high (in the top 25%), the beta of the MAI portfolio is 1.04. (Note: This regime-varying beta is also commonly seen in tactical strategies, as we discussed here.)

picture5

Data Source: Yahoo! Finance.  Calculations by Newfound Research.  The benchmark represents a 50/50 portfolio of the iShares MSCI ACWI ETF (ticker: ACWI) and the iShares Barclays Aggregate Bond ETF (ticker: AGG).  Data does not include any fees or expenses other than underlying ETF management fees.  Past performance does not guarantee future results.  

We can go even deeper by modeling MAI portfolio returns as a function of select risk factors instead of just as a function of the benchmark return.  For simplicity, we will limit ourselves to the following factors:

  1. Equity risk (S&P 500 return)
  2. Interest rate risk (change in both yield curve level and yield curve slope)
  3. Credit risk (change in credit spreads)
  4. Currency risk (change in trade-weighted dollar index).

Using this approach, we find that the current portfolio is:

  1. Long equity risk
  2. Long interest rate risk (i.e. makes money when rates fall and the yield curve flattens)
  3. Long credit risk (i.e. makes money when credit spreads tighten)
  4. Short the U.S. dollar (i.e. makes money when the dollar depreciates against foreign currencies).

We can also decompose the proportion of risk driven by each factor, assuming that the correlations between risk factors are zero.  We see that 80%+ of the risk is driven by equities, interest rates, and credit with a roughly equal split between the three.

picture6

Data Source: Yahoo! Finance, Federal Reserve Bank of St. Louis.  Calculations by Newfound Research.  Past performance does not guarantee future results.  

This framework can also be used to understand the approximate responses that we may expect the portfolio to have to various market events.  The results from the regression are below.  We can interpret this as follows:

  1. Beta to U.S. Equities is 0.31. A 100bps gain (loss) in stocks implies a 31bps gain (loss) in the MAI portfolio.
  2. Beta to Yield Curve Level is -5.36. A 10bps increase (decrease) in interest rates implies a 54bps loss (gain) in the MAI portfolio.
  3. Beta to Yield Curve Slope is -4.69. A 10bps steepening (flattening) in interest rates implies a 47bps loss (gain) in the MAI portfolio.
  4. Beta to the U.S. Dollar Index is -0.29. A 100bps weakening (strengthening) of the dollar implies a 29bps gain (loss) in the MAI portfolio.
  5. Beta to credit spreads is -5.30. A 10bps increase (decrease) in credit spreads implies a 53bps loss (gain) in the MAI portfolio.

picture7

Data Source: Yahoo! Finance, Federal Reserve Bank of St. Louis.  Calculations by Newfound Research.  Past performance does not guarantee future results.  

Another benefit of this type of risk analysis is that we can start to understand how MAI’s risk exposures can be expected to change should negative momentum start to prevail in the investment universe.  As of last week, the exposures most likely to be removed from the portfolio in the short to medium term should negative momentum take hold, are listed below (in descending order of likelihood, i.e. Treasuries are the most likely to be removed).

  1. S. Treasuries
  2. Emerging Market Bonds (Local)
  3. Emerging Market Bonds (USD)
  4. International REITs
  5. S. REITs
  6. Preferreds
  7. International Dividend Stocks
  8. Investment Grade Corporate Bonds

Below, we plot how the risk exposures of Multi-Asset Income may change should these exposures get removed in the order listed above.

Note on reading the chart: The leftmost bar in each group indicates the risk exposures of the MAI portfolio as of 9/30.  The second bar from the left (orange), indicates the risk exposures of the MAI portfolio with U.S. Treasuries removed.  The third bar from the left (gray) indicates the risk exposures of the MAI  portfolio with U.S. Treasuries and local currency Emerging Market bonds removed.  The rightmost bar indicates the risk exposures of the MAI  portfolio with all eight of the asset classes listed above removed. 

picture8

Conclusion 

There are many endeavors we undertake when appropriate expectations can make it easier to stick to our plans. Investing is a prime example.

Two starting points for setting expectations for an investment strategy are 1) looking at historical relative return ranges and 2) looking at sensitivity to risk factors.

Historical return ranges can give context to current out-/underperformance. If the current return is within the historical ranges, there is probably no cause for concern as long as there are no confounding factors. Assessing the current returns in the appropriate context can reduce the risk of making emotional decisions that would violate the plan.

Looking at sensitivity to risk factors can also set expectations. If you know what underlying market forces have driven the returns of a strategy in the past, that can be a good baseline gauge of what effects changes in these factors can have in the future.

Any cursory plan that is not well thought out will be hard to stick to. Explicitly stating objectives, risks, and expectations are crucial steps in building out a plan that can be followed, which can lead to better long-term portfolio performance and less short-term stress.

We see that our most likely path forward involves reducing interest rate and currency risk in favor of equity risk.

Client Talking Points

  • A well-designed investment plan is an important part of achieving investment objectives., but even the best investment plan will be useless if someone cannot stick to it.
  • The plan not only should outline the investment process but should also include a thorough assessment of risks and expectations for short-term deviations.
  • Having a clear understanding of risks and expectations can reduce the risk of making emotional decisions that lead to failure to achieve long-term investment goals.

 

 

 

 

Justin is a Managing Director and Portfolio Manager at Newfound Research, a quantitative asset manager offering a suite of separately managed accounts and mutual funds. At Newfound, Justin is responsible for portfolio management, investment research, strategy development, and communication of the firm's views to clients.

Justin is a frequent speaker on industry panels and is a contributor to ETF Trends.

Prior to Newfound, Justin worked for J.P. Morgan and Deutsche Bank. At J.P. Morgan, he structured and syndicated ABS transactions while also managing risk on a proprietary ABS portfolio. At Deutsche Bank, Justin spent time on the event‐driven, high‐yield debt, and mortgage derivative trading desks.

Justin holds a Master of Science in Computational Finance and a Master of Business Administration from Carnegie Mellon University as a well as a BBA in Mathematics and Finance from the University of Notre Dame.