- Value strategies have been underperforming long enough for investors to question the validity of the strategy and the existence of a value premium.
- Underwhelming performance of value factors and the negative impact from unintended exposure to other factors were two of the main drivers of the prolonged underperformance.
- It may be time to update how value is measured to reflect changes in the market and economic landscapes, as well as portfolio-construction approaches designed to capture the value premium.
In 2016, we published a blog post called “The value factor marks a decade of disappointment.” Five years on, the title of our post remains all too valid (give or take half a decade).
Today’s post is the first in a series representing MSCI’s latest research into value investing, including a search for the reasons behind its underperformance, potential ways to update definitions of value and ideas for alternative approaches to portfolio construction as investors seek to capture the value-factor premium.
Value Investing Didn’t Offer Much Value Over the Last 10 to 15 Years
The value factor is closing in on the record for continuous underperformance currently held by the size factor.1 And it has not been limited to any specific index or region. Previous MSCI research has shown that value funds in aggregate underperformed the broad market and their respective benchmarks, and those with more value exposure reported greater underperformance.
While there has been a recent uptick in performance, it’s still worth asking whether the common notion among investors that the factor is broken is, in fact, correct.
For starters, we use the MSCI Enhanced Value Indexes as a proxy2 for the performance of value in different regions since 2001, looking at each regional index relative to its market-capitalization-weighted parent index (e.g., the MSCI USA Enhanced Value Index vs. the MSCI USA Index). For developed markets (DM), represented by the MSCI World Index and MSCI USA Enhanced Value Index, the performance drought started around 2006. Emerging markets (EM) fared better, but performance over the last decade was still negative, as illustrated by the MSCI EM and MSCI ACWI Enhanced Value Indexes.3
Value-Index Performance Relative to Market-Cap Index
From ‘Smooth Sailing’ to a ‘Perfect Storm’
To better understand what changed for value indexes and value funds from the first decade of the 2000s to the 2010s, we again use the MSCI World Enhanced Value Index as a proxy, looking at the contributors to its return for the two periods.4 As we see in the exhibit below, from 2001 to 2010, style factors5 were the dominant contributors to index performance, explaining 6.24% of the 6.59% of active return. There were contributions from countries, industries and asset selection or stock-specific returns,6 but they were dwarfed by the impact of style factors.
For the following decade, 2011-2020, the story is very different. Style factors were still significant contributors to performance, but this time they detracted from performance. In addition, stock-specific returns had a relatively large negative impact on performance, contributing 2.29% of the 4.13% underperformance of the index over this period.
Risk/Return Attribution Comparison
Return and risk attribution of MSCI World Enhanced Value Index, using the MSCI Global Equity Model for Long-Term Investors (GEMLT) from Dec. 29, 2000, to Dec. 31, 2010, and Dec. 31, 2010, to Dec. 31, 2020.
From Where Did the Storm Winds Blow?
Given style factors’ large positive and negative contribution to index performance over the two decades, respectively, the natural question is: What changed? To find out, we look at the contribution of each of the 16 style factors in the MSCI Global Equity Model for Long-Term Investors (GEMLT) to the performance of our proxy, the MSCI World Enhanced Value Index. The sharply contrasting pictures of the past two decades help shed light on what changed.
From 2001 to 2010, the index had significant exposure to the two value factors in GEMLT (book-to-price ratio and earnings yield),7 and these exposures resulted in attractive positive returns. In addition, the index’s unintended exposures to non-value factors, whether small or large, mostly had a positive impact on the performance of the index during that period.
The situation almost completely reversed for the 2011-to-2020 period. We still saw significant exposure to the two value factors showcasing the robustness of the index in targeting value-oriented companies, but these exposures didn’t translate into returns of the same magnitude as the previous decade; book-to-price showed a small positive return and earnings yield contribution was negative. And, this time around, exposure to non-value factors was a major source of underperformance. Momentum had the largest negative impact, but most other factors also contributed negatively to returns, a stark contrast to the previous decade.
A Tale of Two Decades of Value Performance
‘Pure’ Value Exposure Was Not a Panacea
Given the large negative contribution of non-value factors to value-index underperformance over the last 10 to 15 years, investors may be tempted to accept that as an explanation. Unfortunately, we found that even if value-factor indexes had managed to neutralize the impact of all other factors, their performance still would not have looked so attractive. As shown in the exhibit below, the performance of the two value factors across various MSCI risk models,8 based on different regions, did show a long-term premium (except for book to price in the U.S. model). However, while slightly different from model to model and across the two factors, one obvious commonality is the lower return over the second half of the 20-year period.
Book-to-Price and Earnings Yield Underperformed from 2011 to 2020
The GEMLT, USE4, EEM AND EULT models apply to global, U.S., EM and European equity markets, respectively.
The clear contrast between the performance of value indexes during the first and second decades of the 2000s can suggest a couple of ideas about why this occurred. It could be that the macroeconomic environment of the post-2008-global-financial-crisis period simply did not favor value investing. But it also could be that the value factor, as defined today, is broken. Do valuation ratios, used for value-factor definitions, need to be refined to suit today’s tech-heavy, asset-light business environment?
And what about portfolio-construction approaches’ impact on an investor’s ability to capture the value-factor premium? While value factors contributed to the performance of value indexes and funds, we saw that unintended exposures to non-value factors and stock-specific returns also contributed to value’s underperformance over the last decade. We’ll discuss these questions in more detail as the series continues.
The authors thank Arihant Jain and Waman Virgaonkar for their contributions to this blog.
1Based on post-1976 data from Kenneth R. French’s faculty page. The size factor (small minus big had an extended period of underperformance (1983-1999). The recent period of underperformance for the value factor (high minus low) started in 2006.
“Kenneth R. French.” Dartmouth.edu.
2Value strategies can differ in many ways (e.g., valuation metrics, portfolio construction methodologies). Throughout this post, we use the MSCI Enhanced Value Indexes as a proxy for value strategies but aim to keep the analysis applicable to a typical value strategy. The MSCI Enhanced Value Indexes are designed to represent the performance of companies that exhibit relatively higher value characteristics within the parent universe of securities. The index includes and reweights stocks from its underlying market-cap-weighted index universe based on three valuation ratios: price to earnings, price to book and enterprise value to cash flow from operations.
3The MSCI ACWI Index is a global index comprising of both DM and EM markets.
4We used MSCI Global Equity Model for Long-Term Investors (GEMLT) to break down the return of the index to its underlying components.
5The aggregate contribution of all 16 style factors in the GEMLT model, including value factors such as earnings yield and book-to-price ratio but also other factors such as momentum and volatility.
6Asset selection or stock-specific return is the part of the return that cannot be explained by the return of country, industry or style factors. When a fundamental factor model is used to analyze the return and risk of a portfolio, the active return of the portfolio is attributed to various factors, based on the exposure of the portfolio over time to any of the underlying factors and the return of that specific factor over the same period. The part of the return that is not explained by the common factors is called asset-selection or stock-specific return.
7Book to price for a company is the ratio of book value per share divided by its stock price. Earnings yield in the GEMLT model is a combination of four fundamental ratios: cash earnings per share divided by the stock price, earnings per share divided by the stock price, predicted earnings per share divided by stock price and earnings before interest and taxes (EBIT) divided by total enterprise value of the company.
8These time series represent the performance of pure factors in respective models. Pure factors are constructed using multi-variate regression are long/short portfolios with unit exposure to the target factors (e.g. book-to-price) and zero exposure to any other factor in the model (style, industry or country).