With the peak of the COVID-19 crisis potentially behind us, investors face a new challenge of predicting where the economy goes from here and, with that, how to best position their equity portfolios. While markets have been in recovery mode since the November 2020 general vaccine announcement, we believe there are significant risks that loom large on the horizon and threaten to derail that recovery: unprecedented fiscal stimulus has driven equity multiples to near record levels and the massive expansion of central bank balance sheets has increased concern about potential inflation and rising interest rates.
In this paper we examine the historical performance of equity factors in various stages of the macroeconomic cycle and under several past yield curve scenarios. We find that not only are equity factors an ideal tool to express views about macroeconomic expectations, but now, sitting at a possible economic cross-roads, may be the perfect time to structure your equity portfolio around factor exposures.
To summarize: Our analysis suggests that the current economic recovery has prompted factors like value1, size2 and dividend yield3 to outperform, which has been consistent with past historical recoveries. While we expect this phase to continue for the immediate future, the potential of re-entering an economic contraction or a COVID resurgence is very real, in which case a defensive exposure to low volatility could be an excellent alternative for investors to consider.
Our research suggests that the interest rate picture tells a very similar story: rising growth and inflation expectations that have buoyed the 10-year Treasury rate, which is also consistent historically with an economic recovery. With the US Federal Reserve (Fed) keeping short interest rates4 low, our research suggests that this “Bear Steepening”5 of the yield curve, has historically been very good for factors like value and size. While we expect the Fed to continue on their current course, should economic growth fail to materialize and long-term rates6 decline, low volatility is expected to outperform.
Using the cyclical trend7 of the Conference Board Leading Economic Index, we studied how well-known equity factors — size, value, momentum8, dividend yield, quality9 and low volatility10 — behaved under various stages of the economic cycle in the U.S.
The four stages are:
Using the definitions above, as you can see in Exhibit 1 below, the U.S. economy experienced four major economic cycles during the analysis period of December 1978 to February 2021 with cycle lengths spanning six to 10 years. In the most recent cycle following the housing crisis in 2009, the U.S. economy from January 2009 to February 2020 went through three expansionary regimes (accelerated or decelerated growth), supported by accommodating monetary and fiscal policies.
This model as seen in Exhibit 1 indicates that, currently, the U.S. economy is in a recovery period of negative but accelerating growth as indicated through a gradual increase in real economic variables such as employment, single-unit housing starts and durable goods orders. It is important to note that because this model uses leading indicators, our opinion is that it is better thought of as a predictor of economic conditions rather than a current barometer. Although this may produce different results than some of the standard literature we have seen on this topic, we find the forward looking nature of the results may provide useful insights not gleaned from other approaches.
EXHIBIT 1: BREAKING DOWN THE BUSINESS CYCLES
We first have to define the business cycles before we evaluate how factors performed. We broke down the cycles from December 1978 to February 2021 using monthly changes in the Conference Board Leading Economic Index (LEI) for the U.S.
EXHIBIT 2: AVERAGE FACTOR EXCESS RETURN ACROSS ECONOMIC REGIMES (DECEMBER 1978-DECEMBER 2020)
Factors are more potent during contractionary periods (contraction and recovery) than expansionary periods (expansion and slowdown) with the notable exception of momentum. Orange cells are highlighted for emphasis.
To further summarize and provide context for these results:
In 2020, factors generally performed in line with the above table, recognizing that there were two separate regimes experienced within that year. The first was a contraction that started about February 19 through March 23 in which the S&P 500 index declined by almost 34%.12 Defensive factors like low volatility and quality performed well during this period while the size factor underperformed them which is consistent with Exhibit 2. However, the market entered a rapid recovery phase after March 23rd with the S&P 500 index gaining about 70% by year end.13 During this time span, defensive factors like low volatility and quality underperformed while size and value posted gains but the momentum factor faltered; again, generally consistent with Exhibit 2.
Interest Rate Risk
Interest rate changes and the shape of the yield curve may provide a lot of information about market conditions that may drive capital allocation decisions throughout the economy and thus influences equity market returns. Interest rates affect stock valuations through two main channels - discount rates that impact the present value of future cash flows and borrowing costs that directly relate to consumer spending. We believe that both influence the performance of equity markets and factors to varying degrees.
In addition to Hunstad (2018), others have explored the relationship between interest rates and factor portfolios and have concluded that unintended interest rate duration risk in factor portfolios is uncompensated. In other words, it contributes to risk but not to return (Choi et al. (2015), Driessen et al. (2017), Golubov and Konstantinidi (2018), Daniel (2018)). See the References section for further sources of study.
To facilitate our analysis, we first place the shape of the yield curve into one of four regimes based on interest rate changes and the shape of the curve:
As in Hunstad (2018), we emphasize that changes in interest rates do not have any causal relationship to pure factor returns. However, we do see interest rate risks creep into factor portfolios in the form of unintentional sector, industry, leverage, regional and country exposures.
Factor Insights during Changing Rate Environments
The following chart summarizes the performance of Fama-French factor portfolios by yield curve regime; portfolios are constructed without controls for unintentional and uncompensated risks.
EXHIBIT 3: FAMA FRENCH FACTOR RETURNS
Our analysis showed that the economic environment played a role in how various factors historically performed.
Size: Smaller companies tend to outperform during bear flattening and steepening environments when the Fed is typically tightening and economic growth prospects are improving.
Value: Value stocks tend to do the best in bear steepening environments when the Fed is typically tightening and the market is pricing in higher economic growth prospects.
Quality: High quality companies benefit from a falling rate environment when the Fed is typically easing, as they are able to deploy capital at cheaper borrowing rates to profitable investments.
Low Volatility: Low volatility stocks tend to do well in a falling rate environment, whereas in a rising rate environment when the Fed is tightening, low volatility stocks tend to underperform.
Momentum: High momentum stocks have tended to do well in most interest rate regimes.
Dividend Yield: Higher yielding stocks tend to do well in a falling rate environment, whereas in a rising rate environment when the Fed is tightening, stocks with higher dividend yield underperform.
We have strong conviction in the long-run efficacy of factors and the benefits of factor- based strategies, but investors that are concerned that the current macroeconomic and interest rate environment may shift should consider the models presented and how adding factor exposure may help achieve their investing goals.
FIND OUT MORE
The FlexShares approach to investing is, first and foremost, investor-centric and goal oriented. We pride ourselves on our commitment to developing products that are designed to meet real-world objectives for both institutional and individual investors. If you would like to discuss the attributes of any of the ETFs discussed in this report in greater depth or find out more about the index methodology behind them please don't hesitate to call us at 1-855-FlexETF (1-855-353-9383).
Hunstad, Michael (2018): Factors not Sectors https://pointofview.northerntrust.com/factors-not-sectors-4d63a2798ffc
What is Quality? (2014). Northern Trust White Paper.
Driessen, Joost and Kuiper, Ivo and Beilo, Robbert, Does Interest Rate Exposure Explain the Low-Volatility Anomaly? (January 3, 2017). Found that about 20% of the “low volatility” premium can be explained by an implicit exposure to interest rates.
Choi, Jaewon, Matthew P Richardson, and Robert F Whitelaw. The Technical Report (2018). Showed that interest rate sensitivity decreases with the priority of instruments in the capital structure.
Daniel, Kent D. and Mota, Lira and Rottke, Simon and Santos, Tano, The Cross-Section of Risk and Return (October 31, 2018). Argue that the duration risk component within factors is not priced and needs to be hedged to achieve mean variance efficient outcomes.
Golubov, Andrey and Konstantinidi, Theodosia, Where Is the Risk in Value? Evidence From a Market-to-Book Decomposition (December 18, 2018). Argues duration risk is correlated mostly with the unpriced/uncompensated (cross-industry) component of the value premium, rather than the priced/compensated (intra-industry) component of the value premium.
1 The value factor distinguishes between value stocks and growth stocks using the ratio of book-value of equity to market capitalization.
2 The size factor differentiates between large cap (greater than $10 billion) and mid to small cap stocks with a market capitalization of between $2 billion and $10 billion, and less than $2 billion respectively.
3 Dividend yield is shown as a percentage and attempts to compute a measure of predicted dividend yield using the past history of dividends and the market price behavior of the stock.
4 Short-term interest rates are the rates at which short-term borrowings are effected between financial institutions or the rate at which short-term government paper is issued or traded in the market. Short-term interest rates are generally averages of daily rates, measured as a percentage.
5 Long-term rates are rising faster than short-term - typically associated with periods when the Fed is loosening monetary conditions and market participants may be anticipating faster economic growth and inflation.
6 Long-term interest rates refer to government bonds maturing in ten years or longer.
7 We used business cycle indicator data from the Conference Board as the input for our economic regime model. The Conference Board publishes three major business cycle indicators: leading, coincident, and lagging indicators, on a monthly basis. We chose the Conference Board Leading Economic Index (LEI) for this study because investors tend to watch LEI more closely to gauge economic outlook and it has shown predictive power on economic turns as documented in various research publications (see Vaccara and Zarnowitz  and Stock and Watson ). Using a signal processing technique known as an HP-filter (see Prescott and Hodrick  for details), we extracted the cyclical trend out of LEI MoM change. The cyclical trend was then classified into four different regimes based on simple rules related to the slope of the trendline.
8 Momentum factor captures sustained relative performance and its effect on risk.
9 Quality factor utilizes the Northern Trust approach that attempts to measure companies that have sustainable competitive advantages and have generated sustainable shareholder value over time. We do this by measuring characteristics including strong profitability, consistent and strong levels of cash flows, and prudent deployment of capital by an efficient management team.
10 Low volatility factor captures relative volatility using measures of both long-term historical volatility and near-term historical volatility.
11 Beta is a concept that measures the expected move in a stock relative to movements in the overall market.
12 Source: Bloomberg. Performance for the S&P500 Index from February 19, 2020 – March 23, 2020.
13 Source: Bloomberg. Performance for the S&P500 Index from March 24, 2020 – December 31, 2020.