May 27, 2016

Elm’s strategy in The Journal of Portfolio Management

While past performance is not necessarily indicative of future returns, we draw comfort from knowing that the way we invest has done well for nearly the past one hundred years.1 This is the main conclusion of our research paper, A Case Study for Using Value and Momentum at the Asset Class Level,2 which was just published in The Journal of Portfolio Management. We believe that our simple rules-based approach to dynamic asset allocation has the added benefit that it helps investors avoid feeling they need to make ad hoc changes to their portfolios based on current events, which tends to result in subpar long-term returns.

Here are a few highlights, which you can find in Exhibits 6, 8 and 9 in the paper:

In every decade since 1926, a dynamic asset allocation approach, based on value and momentum applied to asset classes, outperformed a static balanced equity/bond Baseline.
The increase in return delivered by value and momentum since 1975 was 2.5% per year, and also reduced the risk of large losses. The increase in returns in the 1926-1975 period was similar.
Bear markets were the periods when value and momentum did best, resulting in significantly less negative returns than the static Baseline.
Value and momentum worked well together. In the 1975-2013 period, the biggest loss of a portfolio that relied on value combined with momentum was significantly lower than the worst loss suffered by portfolios that relied on value or momentum applied individually.

The paper describes how we apply value and momentum to our Baseline portfolio. As the paper is looking at investment periods wherein many of the asset buckets in which we invest were not available (e.g. emerging market equities prior to the 1980s), the Baseline portfolio that is analyzed differs from the Baseline portfolios we currently use in our various offerings. Also, our base case in the paper had a 50% weight on valuation and a 50% weight on momentum, whereas we currently put a 67% on value and a 33% weight on momentum. In the paper, we do show the sensitivity of past returns to this assumption.
This paper started life as a less formal paper titled, ‘Investing for the Rest of Us,’ which Rich Dewey and I first drafted in 2009.

PREVIOUS

NEXT

X

Baseline Weights: The Baseline defines the strategy's 'average' asset allocation over a long period of time, and serves as the reference point for Long-Term and Trend deviations. The Baseline represents the asset-allocation that a prospective investor would choose were they seeking a static strategy. See here for more information about the Baseline construction methodology.

Target Weights:The target weights used in the periodic portfolio rebalancing. Portfolios may not be rebalanced exactly to target, as Elm's execution engine tries to find a parsimonious set of trades which get as close as possible to target weights while also minimizing transaction costs, tax costs, etc.

Target Weight Breakdown:

Expected Return: The target weight resulting only from the Expected return-based metric. We want target weights to be proportional to the Expected excess returns, while staying within 2/3 above or below the bucket's Baseline Weight.
Risk: The deviation caused by the Risk Metric. We want a positive Deviation in lower-risk environments, and vice versa. The size of Risk-based deviation can be at most 1/3 of the bucket's Baseline Weight.
The Target Weight is equal to the Return + Risk components in the Breakdown.

Expected Return Metric:

Equities: 1 / CAPE - 10y TIPS Real Yield (Excess Expected Return in excess of the long-term risk-free rate)
Fixed Income: 10y TIPS Real Yield
For a small number of asset buckets we use slightly different signals than described above, and a few buckets use secondary metrics as well. Please contact us for details if you're interested.

Risk Metric:

We use a momentum signal as a proxy for the risk-environment: positive momentum indicates a low-risk environment, negative momentum indictse a higher-risk environment.
Balanced: Spot Asset Class Total Return Index / 1y-Avg Total Return Index
(All indexes adjusted for inflation and risk-premia)
All-Equity: [Spot Asset Class Total Return Index / 1y-Avg Asset Class Total Return Index] - [Spot Baseline Total Return Index / 1y-Avg Baseline Total Return Index]
(All indexes adjusted for inflation and risk-premia)

Disclaimer

Elm Partners’ Funds are available only to accredited investors. The content of this website and related materials are not intended to solicit investment, but rather to elicit discussion and exploration of better ways to invest. The contents of this website do not comprise an offering document. Neither historical simulations nor past return performance are necessarily indicative of future returns. Elm Partners Management LLC is an investment advisor registered with the SEC. Here are the most recent copies of our Form ADV Part 1, ADV Part 2A2B and Form CRS.

Fidelity Investments (“Fidelity”) is an independent company, unaffiliated with Elm Partners (“Elm”). Fidelity is a service provider to Elm. There is no form of legal partnership, agency affiliation, or similar relationship between your financial advisor and Fidelity, nor is such a relationship created or implied by the information herein. Fidelity has not been involved with the preparation of the content supplied by Elm and does not guarantee, or assume any responsibility for, its content. Fidelity Investments is a registered trademark of FMR LLC. Fidelity InstitutionalSM provides clearing, custody, and other brokerage services through National Financial Services LLC or Fidelity Brokerage Services LLC, Members NYSE, SIPC. 830247.3.0

Here is the most recent Version of our Privacy Policy

Have a great day!