Heading into September, dividend strategies were one of the year’s poorer performers, trailing the returns of the S&P 500® Index by 8.5% year-to-date and 2.1% over the past year.1 This underperformance occurred despite the fact that the top two sector weights in high yielding dividend strategies-Utilities and Real Estate-outperformed the market this year and over the past year.2

Then, in September, everything changed. 

Not what we thought 

The performance trends leading up to September run counter to the traditional mindset of investors typically seeking out bond-proxy dividend equities to source income in an environment where bond yields had plumbed to 2016 lows, and where 60% of US equities were then yielding more than the 10 Year US Treasury Bond at the end of August.3 This mindset, even though performance was lackluster, is depicted in investor fund flow patterns, however.  Over the past year dividend oriented ETF strategies listed in the US have amassed more than $25 billion4 of assets, equating to nearly 40% of all Smart Beta related fund flows – based on our proprietary classification system.  Investors, despite the performance, sought the yield associated with these equities even as the total return dynamics underperformed the broader market.  

But what drove this unexpected lackluster performance in what would appear to be a supportive market environment? To answer that, we performed a Brinson attribution. We partitioned returns by sectors, seeking to understand if the sector effect-sector weighting differences relative to the broader market-were the main culprit of poor performance or if it was the individual security selection (e.g. owning Campbell Soup at a higher weight than it its market cap weighted exposure). We analyzed the returns for the S&P 500 High Dividend Yield Index (“the high dividend strategy”) and the S&P 500(“the market”) over the past year and year-to-date through August.  

As shown below, we found that in each time period, the sector biases for the high dividend yield strategy were a net positive contributor. The stock weights, however, were net detractors. In almost every sector for both periods, security selection was negative.  

Stocks with high dividends, therefore, could be said to have been poor performers, but not high dividend sectors. The latter point explains why two bond-proxy sectors (Utilities and Real Estate) can be top performers but bond-proxy dividend paying equities cannot.  

But we wanted to take our analysis a step deeper. The below chart shows the same Brinson attribution analysis but with securities bucketed into the more granular industry groups. The extra granularity didn’t reveal anything different. 

please insert chart here

Source: Bloomberg Finance L.P. as of 08/31/2019.  Past performance is not a guarantee of future results. Returns are unmanaged and do not reflect the deduction of any fees or expenses. 

Going a click deeper 

But did this Brinson analysis really uncover what was really happening beneath the sector surface? To answer that we ran a factor based attribution. 

With factor based attribution, returns are decomposed based upon quantifying historical sensitives to industry (e.g. Banks, Retail, and Industrials) and style factor returns such as value, growth, momentum, size, and volatility. The residual return after accounting for both industry and style factor effects is defined as the idiosyncratic stock specific return. This is a more granular analysis than Brinson, as it shows how much of the return can be attributed to just being a “dividend” style stock and how much is based on actual stock specific matters such as an unexpected earnings surprise or M&A announcements.  

What did we find? Factor based attribution contradicted the assertion that the underperformance of a high dividend strategy was due to the fact that the stocks were bad picks. In fact, after accounting for style effects, the stock specific return attribution was positive! The real-driver was style factors associated with this basket of stocks, as shown below for both a trailing one-year and year-to-date-through-August time frames.

please insert chart here

Source: Bloomberg Finance L.P. as of 08/31/2019.  Bloomberg US Fundamental Factor Model used for Factor Based Attribution. Past performance is not a guarantee of future results. 

But what factors? To answer that, we plotted the factor exposure and the associated return contribution for the past year below. The year-to-date through August results were the same, so for simplicity the longer trailing one-year returns are shown. The driving force from a factor exposure perspective was size, value, profitability, and growth. Dividend yield was actually a positive contributor. This confirms the notion that the dividend style should produce positive returns when bond rates are falling, and it refutes the misperception that “dividends” just didn’t work as rates fell.  

The chart also shows the exposure to each one of these factors, as represented by the circles. The positive exposure figure for “value” indicates the portfolio is overweight inexpensive stocks. This value exposure is not a surprise given that the value and dividend factor have a historical 73% correlation of excess returns.5 Now, the negative exposure number on size indicates that the dividend strategy is underweight some of the larger-cap names versus the market portfolio. Because the return for “size” was negative- meaning large mega caps did better than smaller large capitalization firms-it was the increased exposure to smaller large capitalization firms within the S&P 500 that led to negative returns.  

Overall, the high dividend strategy was overweight value stocks (positive exposure) and underweight profitable, growth, and momentum oriented stocks (negative exposure). Each of these factor exposures were not beneficial over the past year, as being a lower size, value, and anti-growth/momentum stock drove 97% of the overall negative style return impact.

please insert chart here

Source: Bloomberg Finance L.P. as of 08/31/2019.  Bloomberg US Fundamental Factor Model used for Factor Based Attribution. Past performance is not a guarantee of future results. 

Remember, remember the factor reversal of September 

In a sudden reversal of performance, the S&P 500 High Dividend Index rallied 7.5% compared to the market’s 2.5% return through the first sixteen days of September.6 The rally wasn’t spurred because falling  bond yields or the more than 4% yield on large-cap equities became insatiably attractive. Bond yields actually reversed course, with the US 10 year yield posting its biggest weekly move since the presidential election in 2016. 

The reason for dividend’s pop was a noticeable value/size versus growth/momentum performance reversal, as shown below. The month-to-date return differential between value and momentum, and value versus growth were the largest month-over-month moves since the election as well, with both reversals registering as a greater than one standard deviation event.7 The moves have been attributed to short covering and de-risking of crowded trades by hedge funds and fundamental investors as well as an unwinding of bearish positions given the reprieve in tariffs. Given this, a basket of the most shorted stocks was also included in the chart, capturing this relationship.

please insert chart here

Source: Bloomberg Finance L.P. as of 09/16/2019.  Past performance is not a guarantee of future results. Returns are unmanaged and do not reflect the deduction of any fees or expenses.

Why is this important? Because the high dividend yield strategy discussed throughout this blog has a hefty value/anti-growth and momentum bias. Below are the return trend contributions for those specific style factors within the high dividend yield strategy relative to the S&P 500. Yes, value return contribution increased. But what really drove returns as of late was what the strategy did not have a lot of: growth and momentum. Dividend yield factor fell slightly, once again confirming the interest rate dependent viewpoint discussed earlier as rates rose.

please insert chart here

Source: Bloomberg Finance L.P. as of 09/16/2019.  Bloomberg US Fundamental Factor Model used for Factor Based Attribution. Past performance is not a guarantee of future results. 

Know what you own

Understanding investment holdings is an important part of due diligence. The strength and validity of the recent dividend rally is still unknown. Given the technical forces of de-risking crowded trades, the likelihood of continued value exuberance is low. Underlying fundamentals still portend to a slowing of economic sentiment and corporate profits. 

Based upon historical analysis, value stocks typically do not perform well in such a slowdown.8 A relaxation of trade tariffs and a steepening of the yield curve may be a catalyst for these stocks. However, timing those two market variables is akin to trying to hug a tiger-I’m sure once or twice it will work but not every time, and on one occasion don’t be surprised if you get your arm ripped off.  

Dividend strategies are a close cousin of value. When investors use these strategies as a source of income in a low rate world but are also seeking more stability, it may be more optimal to focus on dividend growers. Unlike high dividend yield strategies, where the emphasis is only on the yield today, there is more attention paid to the sustainability of dividend growth strategies. Just look at the payout ratio, which shows the median dividend growth stock has a payout ratio of 48.6 versus 65.4 for high dividend yield strategies.9 As a result, relative to high dividend yield strategies, dividend growth strategies have a positive factor exposure toward profitability, growth, and with low earnings variability10 – but they still have a portfolio yield in excess of the broader market (2.88 versus 1.94%).11

Knowing what you own is important, and ever more so when deviating from the market cap weighted paradigm when seeking to pinpoint the drivers of a portfolio’s return. 

To learn more about smart beta due diligence and broader factor trends, check out our smart beta page and smart beta blog posts. For more market insights also continuing follow my work on SPDR Blog

1Returns based upon the S&P 500 High Yield Dividend Index and the S&P 500 Index as of 08/31/2019 per Bloomberg Finance L.P.
2S&P 500 Utilities and Real Estate Sector returned 17.2% and 16.3%, respectively from 08/2018 to 08/2019 per Bloomberg Finance L.P. as of 08/31/2019 versus 0.9% for the S&P 500 Index. Returns year to date through August were 17.6%, 26.0%, and 16.7% for the Utilities, Real Estate and S&P 500, respectively. These are the top two weighted sectors represented in the S&P 500 High Yield Dividend Index
3Bloomberg Finance L.P. as of 08/31/2019
4Bloomberg Finance L.P. as of 08/31/2019, calculations by SPDR Americas Research
5Based on monthly price returns for the S&P 500 Pure Value Index and the S&P 500 High Dividend Yield Index versus the S&P 500 Index from 06/1995 to 08/2019
6Bloomberg Finance L.P. as of 09/16/2019
7Based upon monthly index returns for the S&P 500 Pure Value, Pure Growth, and Pure Momentum Indexes from 6/1995 through 09/16/2019.  Data sourced from Bloomberg Finance L.P., calculations by SPDR Americas Research.
8Bloomberg Finance L.P., as of 08/31/2019. Calculations by SPDR Americas Research. Value underperformed the market by 4% during a slowdown, based on the average cumulative return for during an entire cycle since 1990. Slowdowns are defined by slower changes to the LEI Index year-over-year growth rate. Quality, Minimum  Volatility, Value, and Size are represented by the MSCI USA Quality, MSCI USA Minimum Volatility, Russell 1000 Value, and Russell 2000 Indices. The market return is represented by the MSCI USA Index. Past performance is not a guarantee of future results. Index returns are unmanaged and do not reflect the deduction of any fees or expenses.
9Bloomberg Finance L.P., as of 09/16/2019.
10Bloomberg Finance L.P., as of 09/16/2019. Bloomberg US Fundamental Risk Model used for Factor Based Attribution. Past performance is not a guarantee of future results.
11Bloomberg Finance L.P., as of 09/16/2019. Based upon the index yields for the S&P High Yield Dividend Aristocrats Index and the S&P 500 Index

Definitions

Payout Ratio 
The payout ratio shows the proportion of earnings paid out as dividends to shareholders, typically expressed as a percentage of the company's earnings. The payout ratio can also be expressed as dividends paid out as a proportion of cash flow.

S&P High Yield Dividend Index 
The Index is designed to measure the performance of the top 80 high dividend-yielding companies within the S&P 500 Index

S&P High Yield Dividend Aristocrats index 
The Index screens for companies that have consistently increased their dividend for at least 20 consecutive years, and weights the stocks by yield

Brinson attribution
Refers to performance attribution based on active weights relative to a benchmark. There are different variations, but the effects usually include allocation, security selection, currency, and potentially others. 

Risk-based, or factor, performance attribution
Decomposes excess return to active risk factor exposures.

The Bloomberg US Equity Fundamental Factor Model
The model employs a multiple factor modeling approach, which allows a responsive yet stable assessment of major risk factors affecting US Equity assets and portfolios.

Momentum
Separates stocks that have outperformed over the past year and those that have underperformed.

Value
Is a composite metricthat differentiates “rich” and “cheap” stocks.

Dividend Yield
Is another dimension of value, but distinct enough to be a standalone factor measured by the most recently announced net dividend (annualized) divided by the current market price

Size
Is a composite metric distinguishing between large and small stocks.

Trading Activity
Is a turnover based measure. Bloomberg focuses on turnover which is trading volume normalized by shares outstanding. This indirectly controls for the Size effect.

EarningsVariability
Gauges how consistent earnings, cash flows, and sales have been in recent years.

Profitability
Studies firms' profit margins to differentiate between money makers andmoney losers.

Volatility
Differentiates more volatile stocks and less volatile ones by quantifying “volatile” from several different angles.

Growth 
Aims to capture the difference between high and low growers by using historical fundamental and forward-looking analyst data.

Leverage 
Is a composite metric to gauge a firm’s level of leverage.