Value investing is based on the principle of buying companies at a discounted price from their perceived intrinsic values.

Similar to how a bargain hunter searches multiple department stores to find the lowest price on a high-quality cashmere label, value investors look to buy high-quality companies at discounted prices.

Supported by decades of academic and practitioner research,1 value investing has historically offered a premium over the market. We work on behalf of value investors to capture the premium cost-effectively and efficiently using value index exchange-traded funds (ETFs).

There is some risk; some value indexes skew toward limited sectors, which can result in less diversification. But our continual research helps us to regularly improve our strategy and ensure we’re considering factors that could help reduce that risk.

A Brief History of Value Investing

Historically, the frugal investing mentality that underpins a value investing strategy has paid off.

Value investing was first codified in 1934 in Benjamin Graham and David Dodd’s “Securities Analysis,” in which they advocated for purchasing stocks below their intrinsic values.

Since then, value investing has had many champions, most notably Warren Buffett.

In fact since 1927, a value portfolio composed of the cheapest 30% of stocks by book-to-market value has outperformed the middle 40% of the market by 2.2%, and the most expensive 30% of the broad market2 by 3.3% on an annualized basis.3

This is not to say that there haven’t been periods of underperformance—the last 10 years have notably seen a value portfolio beaten by a broad market portfolio.4

Value investing went on to receive a rigorous treatment during the 1990s, when Nobel Prize-winning economist Eugene Fama and Dartmouth professor Kenneth French identified three factors to explain stock performance.

These factors became known as the “Fama-French Three-Factor Model”:

  • Market sensitivity (beta)
  • Company size (market capitalization)
  • Company valuation (book-to-market ratio)

Fama and French found that investors earned a premium in the form of excess returns over the broad market by investing in companies with attractive book-to-market valuations.

This outperformance of value stocks versus stocks in the rest of market is known as the “value premium.”

Since their original paper was published in 1992, further research has found value premia across global markets and asset classes.5

Accessing the Value Premium

Value indexes systematically select stocks that rank highly on traditional valuation ratios (such as book-to-market or earnings-to-price) to capture the excess return over the broad market that value stocks have exhibited historically.

This makes them transparent, and the ETFs that track them generally low-cost.

Value Portfolio vs. Rest of the Market: 1927-2016


A value portfolio has outperformed the rest of the market over almost 90 years of data.

The chart below shows this growth by tracking $1 invested in January 1927 through May 2016.

You can see a value portfolio of the 30% highest ranking book-to-market (cheapest) stocks in the United States, versus a portfolio of the middle 40% and the lowest 30% (most expensive) of stocks in the market, ranked by book-to-market.

Traditional value index ETFs, such as Russell 1000 Value (IWD) and S&P 500 Value (IVE), are common choices for investors who wish to access value premium cheaply and easily.

The historic returns to a value strategy are compelling, but in the process of investing in stocks based on valuation ratios, value index ETFs end up tilting toward sectors which offer little if any additional return while adding to the fund’s overall risk. But why does this happen?

Here, we explore the non-explicit sector risks that value index ETFs take when investing in value stocks.

Value Funds in a Three-Factor World

Using the Fama-French Three-Factor Model, we can measure exactly how much value premium a value fund is capturing, as well as the two other factors in explaining returns: market sensitivity and size.

To establish a baseline, we look at common large-cap broad market index ETFs.

For example, and perhaps not surprisingly, each of the three broad market funds studied captures the full premium associated with market sensitivity. Also note that broad market ETFs, by holding an equal amount of value and non-value stocks, have little-to-no aggregate exposure to the value premium.

Factor Exposures: Broad Market ETFs

The table below shows sensitivities to Fama and French’s three factors that explain stock returns for these broad market ETFs: iShares Russell 1000 (IWB), SPDR S&P 500 (SPY), and Vanguard Large-cap (VV), from January 2006 to December 2015.

A higher market sensitivity percentage means more of a fund’s returns are driven by changes in the broad market. At 100%, this means indexes will move closely with the broad market.

The higher the large size percentage, the more a fund behaves like a large-cap stock. Percentages around 10% indicate that funds behave similarly to large cap stocks, which makes sense because they are composed of the largest stocks in the United States.

A value exposure of 100% means that the fund has captured the full return premium associated with that factor, while a value exposure of 0% means that the fund’s performance is unrelated to that factor and has experienced none of the returns (or risks) associated with it.

In the table, you’ll notice that value exposure is close to 0% for each of these broad market ETFs.


*** Denotes statistical significance at the 99.9% level; ** at 99% level; * at the 95% level. (These sensitivities are estimated using returns data. Like all estimates there is some amount of error involved in their measurement. Statistical significance tells us that, even given the amount of measurement error, the sensitivity is meaningfully different from 0.)
Data sources: Xignite, Kenneth French Data Library (Tuck School, Dartmouth)

Now that we have established a baseline with broad market indexes, consider the table below which shows sensitivities to the three Fama and French factors for three value ETFs.

You’ll notice that value exposure is much higher compared to broad market ETFs.

Factor Exposures: Value ETFs

The next table shows sensitivities to Fama and French’s three factors that explain stock returns for three popular value index ETFs: iShares Russell 1000 Value (IWD), iShares S&P 500 Value (IVE), and Vanguard Value (VTV), from January 2006 to December 2015.

You’ll notice that value exposures are significantly higher (30%) compared to those of the broad index ETFs (0%).


*** Denotes statistical significance at the 99.9% level; ** at 99% level; * at the 95% level. (These sensitivities are estimated using returns data. Like all estimates there is some amount of error involved in their measurement. Statistical significance tells us that, even given the amount of measurement error, the sensitivity is meaningfully different from 0.)
Data sources: Xignite, Kenneth French Data Library (Tuck School, Dartmouth)

It’s clear that value ETFs have significantly more exposure to the value premium, approximately 30%, while exposure to the other two factors (market sensitivity and large size) are similar to broad market ETFs.

This 30% value exposure is what investors hope will result in above-market returns.

Inherent Sector Concentration

If we dig a little deeper and explore where value exposure is actually coming from, we find that much of it comes from sector concentration.

Value index ETFs have few, if any, constraints on risk exposures. They simply select stocks based on valuation ratios. This can lead to significant sector exposures that aren’t explicitly part of the fund design.

Lack of Diversification

Stocks in the same sector tend to move together and are often influenced by the same economic factors. Accounting and general business practices may also make some sectors cheaper than others.

As a result, many value indexes are skewed toward certain sectors, which means the diversification in value exposure from an index of hundreds of stocks will likely be less than it appears.

To measure the impact of sectors on value exposure, the next step in our analysis is to break up market sensitivity into 10 individual sector sensitivities.

To the extent that sectors explain value ETF returns better than the Fama and French value factor, we should see a decrease in value exposure once sectors are included, because fund returns are better explained by sector than value.

Cheap Sectors

When we add sectors to our analysis, the exposure to value diminishes substantially.

Value exposures drop by 60% or more after including sectors to explain returns. This means value funds are taking on sector risk as a byproduct of their value-based stock selection and that sector risk accounts for more than half of their perceived value exposure.

Factor Exposures When Controlling for Sectors: Value ETFs

This table shows Fama and French exposures for the same three value index ETFs above (IWD, IVE, and VTV) after allowing for sectors to explain part of fund returns.

You’ll notice much lower value exposures (10%) than in the previous analysis, which didn’t include sectors (30%). This shows that sectors contribute significantly to these funds’ value exposures.


*** Denotes statistical significance at the 99.9% level; ** at 99% level; * at the 95% level. (These sensitivities are estimated using returns data. Like all estimates there is some amount of error involved in their measurement. Statistical significance tells us that, even given the amount of measurement error, the sensitivity is meaningfully different from 0.)
Data sources: Xignite, Kenneth French Data Library (Tuck School, Dartmouth)

Put another way, sorting by valuation ratios (such as book-to-market ratios) loads up on cheap sectors. When we introduce sector returns into the model, it captures this effect which was previously attributed to the value factor.

We can see that value stocks tend to cluster in certain sectors rather than in a broad set of diversified stocks, causing value ETFs to overweight those particular sectors. Naturally, it’s worth investigating whether those sector concentrations help investors.

Controlling Sectors May Reduce Risk

Historically, the sectors that value indexes have overweighted and underweighted have not been beneficial to returns.

In a study of portfolios constructed using book-to-market ratios, almost all of the value premium came from individual stock selection, while sector bets provided no significant improvement in returns.6

While sector bets do little to help returns, they often have an adverse effect on volatility. Portfolios constructed without sector bets were half as volatile as those that allowed for sector overweights and underweights.

Distribution of Monthly Value Premium Returns: 1964 to 2012


The chart below shows the distribution of monthly value returns for a portfolio without sector constraints, which overweights and underweights sectors as a byproduct of value-ranking stocks, and a portfolio that removes sector biases from its value ranking.

You’ll notice that the value strategy with sector overweights and underweights (blue line) has a much greater dispersion in returns than the value strategy with no sector bets (green line). This shows, graphically, the additional risk that sectors introduce while average returns for the two strategies are roughly the same and statistically indistinguishable.

There are a couple of intuitive reasons for this.

First, sector risk is not something that investors are generally compensated for. While taking more risk to certain sectors will likely increase your overall risk, you should not expect any increase in returns.

Second, because of fundamental differences in industries, both in accounting practice and business operation, some sectors will almost always look cheap while others will look expensive. As a result value indexes will have these inherent biases.

Finally, we would rather have value exposure coming from individual and uncorrelated securities than highly correlated securities within a sector. Exposure through stock selection allows for many small, less-correlated opportunities to capture the value premium.

Value exposure through sectors, on the other hand, provides fewer opportunities to benefit from the positive premium. One would expect that, with fewer opportunities, volatility in returns would be greater, which is exactly what we see.

Enter Purer Factor Funds?

At Betterment, we use value index ETFs in our portfolio because they are generally an efficient, cost-effective way to pursue a value investing strategy and capture excess returns associated with the value premium.

The straightforward composition of value indexes based on valuation ratios, however, may increase exposure to uncompensated risk in sectors.

Until recently, a passive investor could do little to avoid incidental and uncompensated sector exposures.

Now, some so-called value factor funds, such as iShares Edge MSCI USA Value Factor ETF (VLUE) and Fidelity Value Factor ETF (FVAL), have addressed potentially unwanted sector biases by sector-neutralizing their value score before ranking stocks.

But there are caveats. These ETFs have limited live data history, small fund assets under management, thin liquidity, higher turnover and higher expense ratios compared to their traditional value ETF counterparts.

These barriers are too high for Betterment to include the ETFs in our current portfolio; the cost to hold and trade these funds still outweighs the potential benefit of sector risk reduction. But as they gain more traction and liquidity deepens, the case for including them could strengthen. Until then, we will continue researching the risks and returns associated with value investing, as well as ways to practically and cost-effectively implement this strategy for our customers.

1Asness, Moskowitz, Pedersen, 2013; “Value and Momentum Everywhere”
Fama, French, 1992; “The Cross-Section of Expected Stocks Returns”
Fama, French, 1996; “Multifactor Explanations of Asset Pricing Anomalies”

2“Broad market” consists of all equities in the New York Stock Exchange (NYSE), NASDAQ, and American Stock Exchange (AMEX)
3U.S. Research Returns, Data Library, Kenneth R. French
4Stop Lying to Yourself About Value Investing,” Nir Kaissar, Bloomberg Gadfly
5Asness, Moskowitz, Pedersen, 2013; “Value and Momentum Everywhere”
6Novy-Marx, 2009; “Competition, Productivity, Organization and the Cross Section of Expected Returns”

This article intended for educational purposes only and not as investment advice.

This article originally appeared on

More from Betterment: