New Dividends Appreciation data series - backtesting

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
siamond
Posts: 3323
Joined: Mon May 28, 2012 5:50 am

New Dividends Appreciation data series - backtesting

Post by siamond » Wed Aug 09, 2017 7:09 pm

In the past few days, I received two independent requests to assemble a Dividends Appreciation data series, and run some backtesting about portfolios including such asset class. Personally, I always think in terms of total returns, and never paid much attention to such funds trying to select companies with high dividends, as it makes little sense to me as an investing strategy. Still, such dividends-oriented strategy is often raised and discussed on the forum, and I think it would be useful to have a good set of facts available. So I am thinking to add such a data series to the Simba backtesting spreadsheet.

After diving a bit in the topic, I realized that such funds (well, mostly the indices they track) have quite an impressive track record over their entire history (call me surprised!), although such real-life funds didn't seem to make much of a difference in the past 10 years (call me unsurprised!). Please check this Morningstar comparison chart. This compares two dividend appreciation funds and the regular S&P 500 Vanguard fund (VFINX). Hardly convincing!

One such fund is the Vanguard Dividend Appreciation ETF (VIG), the corresponding mutual fund investor being VDAIX. Vanguard tracks the Nasdaq Dividend Achievers index. Unfortunately, both the fund and the index started in 2006, that's only 10 years of history.

The iShares Select Dividend ETF (DVY) addresses the same category, and it tracks Dow Jones US Select Dividend index, which has a solid history (back to 1992). It displays amazing past performance though 912% CAGR since 1992), and while digging a bit more, it appears to be much more in the Mid-Cap-Value realm, while Vanguard and its index are more in the Large-Cap-Blend realm. Ok, that's a different animal.

I also looked for indices providing more history. I found the amusingly named S&P 500 Dividend Aristocrats index, allowing to go back to 1990. This one seems more compatible with the Vanguard large-cap optic, although it uses an equal-weighted technique while Vanguard/Nasdaq are cap-weighted. An example of a fund tracking this S&P index is ProShares S&P 500 Dividend Aristocrats (with a clever ticker: NOBL).

What troubles me is that the S&P Aristocrats index displays a big 11% return since 1992, while the S&P 500 only returned 9%. That is a little hard to believe, unless somehow the S&P Aristocrats filtering procedure ended up indirectly creating a strong factor tilt (well, maybe it did?).

Anyhoo, I ended assembling the following data series (adjusting with the Vanguard ER 0.17% for the index years) as a candidate for the Simba spreadsheet:
- 1985-1989: S&P 500 (by lack of a better source)
- 1990-2006: S&P 500 Dividend Aristocrats
- 2007-2016: Vanguard VDAIX

Any thoughts? Any better choice you can think of?

User avatar
siamond
Posts: 3323
Joined: Mon May 28, 2012 5:50 am

Re: New Dividends Appreciation data series - backtesting

Post by siamond » Wed Aug 09, 2017 7:24 pm

Hmpf. Just realized that the plot is more complicated than I thought... There is also:
- Vanguard Dividend Growth Fund (VDIGX), created mid-1992, and ALSO tracking the NASDAQ US Dividend Achievers index. Note that it's an active fund, currently closed to new investors.
- Vanguard High Dividend Yield Index Fund Investor Shares (VHDYX), created in 2006, and this one tracks the FTSE High Dividend Yield Index, oh joy.

If somebody has some context (including historical) about this hodgepodge from Vanguard, please share!

PS. while we're at it, there is also the Schwab U.S. Dividend Equity ETF (SCHD), which tracks Dow Jones U.S. Dividend 100 Index (available since 1999).

User avatar
nedsaid
Posts: 8504
Joined: Fri Nov 23, 2012 12:33 pm

Re: New Dividends Appreciation data series - backtesting

Post by nedsaid » Wed Aug 09, 2017 9:08 pm

Siamond, it doesn't surprise me that dividend growth has been outperforming the S&P 500. Dividend growth is associated with earnings growth which in turn is associated with Profitability and Quality. In a slow growth economic environment, a premium is placed on consistent earnings growth. Value just does not do as well. My understanding is that Value does better in a strong economy. I suspect there might be a bit of yield chasing in there too.

I have posted about this many times, comparing the S&P 500 to Dividend Growth, High Dividend, and Large Value. High Dividend is associated with Value. It is interesting that High Dividend leaves Large Value in the dust and almost matches the S&P 500. Dividend Growth beats the S&P 500. It seems that Large Growth and Dividend Growth take turns outperforming each other depending on when you look. I see this just eyeballing the "growth of $10,000" charts at Morningstar for the prior 10 year period. So it appears, at least to me, that yield chasing was a factor for Dividend Growth.

The thing is, there is a lot of overlap between High Dividend and Low Volatility. It seems that Low Volatility have slow but dependable growth and higher dividends. The pure Value stocks probably have more leverage on their balance sheets than the Low Volatility Stocks. A few weeks ago, I did analysis on this with my opinions on the relative valuations of Growth, Dividend Growth, S&P 500, Low Volatility, Large Value, and High Dividend.

So I would like you to expand your analysis a bit. My thesis is that there is a lot of overlap between Value, Low Volatility, and High Dividend. There also should be overlap between Growth and Dividend Growth. Actually, the good old S&P 500 should be fairly growthy and because of market cap weighting, it should tilt a bit to Quality. (The most successful stocks have the biggest market weightings).

Pretty much to see if my assertions and suspicions really are true. Let's see if I can repost my analysis and see if it might be any use to you.
Last edited by nedsaid on Wed Aug 09, 2017 9:32 pm, edited 1 time in total.
A fool and his money are good for business.

User avatar
nedsaid
Posts: 8504
Joined: Fri Nov 23, 2012 12:33 pm

Re: New Dividends Appreciation data series - backtesting

Post by nedsaid » Wed Aug 09, 2017 9:27 pm

Siamond, below is a repost from July 3, 2017. I have added comments.
There is overlap between Large Value, High Dividend, and Low Volatility. What I did was look at Vanguard Value Index, Vanguard High Dividend, Vanguard Consumer Staples, and at Low Volatility. Growth of $10,000 for 10 years shows $17,419 for Vanguard Value Index, $19,394 for Vanguard High Dividend Index, $21,526 for Vanguard Dividend Growth, $22,993 for Vanguard Growth Index and $19,796 for the S&P 500. The pattern that I have seen for the last two years or so has held up. By the way, Consumer Staples over 10 years had a growth of 10,000 of $26,469.

You can see that since the financial crisis of 2008-2009 that we have been in a Large Growth
market. Large Growth has done the best over 10 years, then Dividend Growth, then the S&P 500, then High Dividend Index, and you can see that plain old boring Large Value got left in the dust. Dividend Growth is associated with Quality and Profitability (essentially Large Growth) and High Dividend is associated with Value. Dividend Growth has trailed the Growth Index but still beat the S&P 500.
Edit: It seemed that before that Dividend Growth had been consistently beating Large Growth but recently, I have noticed Large Growth beating Dividend Growth. I think this might be a function of rising short term interest rates, so perhaps on the Growth side, the dividend chasing has abated. On the Value side, the relationship between High Dividend and Large Value was unchanged. So the dividend chasing seems to be continuing on the Value side of the market.
You can also see that High Dividend Index was the best of the Value strategies. It handily beat the Value Index and barely trailed the S&P 500. A few years ago, Larry Swedroe said to avoid High Dividend but he underestimated the extent of yield chasing and it actually turned out to be the best value strategy. As I recall, he said it was a bad Value strategy and he was wrong, at least over the last 10 years. My thesis is that rising interest rates will hurt High Dividend the most but as you can see, that hasn't happened yet.

Let's look at forward P/E ratios as calculated by Morningstar. I also show Price/Book and Price/Sales. For fun, I threw in Vanguard Consumer Staples Index as Consumer Staples are popular defensive stocks.

Vanguard Value Index Admiral P/E 16.95 P/B 2.09 P/S 1.55 2.45% Yield
Vanguard High Dividend Index Investor P/E 17.73 P/B 2.64 P/S 1.95 3.03% Yield
Vanguard S&P 500 Index Admiral P/E 19.87 P/B 2.80 P/S 2.09 1.88% Yield
Vanguard Dividend Growth Investors P/E 20.40 P/B 4.15 P/S 1.37 2.00% Yield
Vanguard Growth Index Admiral P/E 24.66 P/B 4.40 P/S 3.30 1.28% Yield
iShares Edge MSCI Min Vol USA ETF P/E 22.61 P/B 3.24 P/S 2.51 2.09% Yield
Vanguard Consumer Staples Index Adm P/E 21.14 P/B 4.21 P/S 1.31 2.62% Yield

You can see that Low Volatility is more expensive than the S&P 500 and I will use this as a proxy
for the market itself. Larry Swedroe has warned to stay away as Low Vol appears to be overgrazed. Consumer Staples looks similarly overpriced.

Vanguard Value Index is the cheapest of all the funds shown though Dividend Growth and Consumer Staples have slightly lower Price/Sales ratios. From a valuation standpoint, plain old boring Large Value looks pretty good right now. Pretty much, I am saying buy Large Value but don't chase for dividends and don't chase low volatility.
Edit: What I am trying to say is that yield chasing seemed to affect performance on both the Growth and the Value sides of the market. High Dividend seemed to be the best Value strategy, at least in the Large Cap space. For a long time, it seemed that Dividend Growth was the best of the Growth strategies until recently. Looking at both the Value and Growth sides would give you a more comprehensive look at how dividends, or maybe good old fashioned yield chasing, affected market performance. It also of course leads to a factor discussion.

So I don't know if this helps you at all. It might just be Nedsaid riding his hobby horse.
A fool and his money are good for business.

User avatar
nedsaid
Posts: 8504
Joined: Fri Nov 23, 2012 12:33 pm

Re: New Dividends Appreciation data series - backtesting

Post by nedsaid » Wed Aug 09, 2017 9:38 pm

Siamond, look at how "growthy" that Consumer Staples looked. These tend to be slower growth but consistent growth and tend to pay a bit higher dividends. Low Volatility looked pretty growthy to me too. I would expect Consumer Staples and Low Vol to be more like Value stocks or at least much less growthy, but it shows you how hard the low volatility train was ridden. Another point was how the factor and yield chasing distorted normal market relationships.
A fool and his money are good for business.

stlutz
Posts: 3968
Joined: Fri Jan 02, 2009 1:08 am

Re: New Dividends Appreciation data series - backtesting

Post by stlutz » Wed Aug 09, 2017 9:43 pm

Important note on the Vanguard Dividend Growth Fund (VDIGX): "Effective December 6, 2002, the fund changed its investment objective and concentration policy. Prior to making these modifications, the fund was called Vanguard Utilities Income Fund, reflecting its former policy of investing in income-producing stocks of utilities companies. The performance prior to December 6, 2002, reflects the fund’s performance under its former investment objective and concentration policy."

jbolden1517
Posts: 868
Joined: Sun Jul 09, 2017 2:53 pm

Re: New Dividends Appreciation data series - backtesting

Post by jbolden1517 » Wed Aug 09, 2017 9:54 pm

[quote="siamond"
I also looked for indices providing more history. I found the amusingly named S&P 500 Dividend Aristocrats index, allowing to go back to 1990. This one seems more compatible with the Vanguard large-cap optic, although it uses an equal-weighted technique while Vanguard/Nasdaq are cap-weighted. An example of a fund tracking this S&P index is ProShares S&P 500 Dividend Aristocrats (with a clever ticker: NOBL).

What troubles me is that the S&P Aristocrats index displays a big 11% return since 1992, while the S&P 500 only returned 9%. That is a little hard to believe, unless somehow the S&P Aristocrats filtering procedure ended up indirectly creating a strong factor tilt (well, maybe it did?). [/quote]

Equal weighting for the SP500 makes a big difference.

Image
The reason it makes such a big difference is margin compression. The cap weighted index represents industries that have much better than average margins:
Image

High margin industries are likely to have much more capital investment leading to margin pressure than low margin industries. Telecommunications is a good example. 2017 is a bull market year but it was dreadful for telco as all the capital investment has caused many of these companies earning to implode. Healthcare is possibly the main exception here where it hasn't experienced any margin pressure even with tremendous investment, that I can't explain. But for other sectors I think that's what the Aristocrats are capturing: reliable dividend stocks, low volatility and non-market weighting for likely future margin pressures.

User avatar
siamond
Posts: 3323
Joined: Mon May 28, 2012 5:50 am

Re: New Dividends Appreciation data series - backtesting

Post by siamond » Wed Aug 09, 2017 10:30 pm

All right, let's step back a bit and come back to basics. Apologies, I never paid attention to such dividends-oriented stocks before.

So, from the little I understand now (this article helped), there are (at least) two possible strategies:
- one seeking stocks with recently growing dividends (that's a "dividends growth" strategy, like Vanguard VDAIX)
- one seeking stocks with established high dividends (that's a "high dividends" strategy, like Vanguard VHDYX)

And since those two strategies are fundamentally different, we might actually seek to construct not one, but TWO new data series. Correct?

As to the S&P 500 Aristocrats index, from what I see in its description, it is Dividends Growth at face value, but heck, after 25 years in a row, the dividends must be pretty high! Cf. this quote:
S&P 500® Dividend Aristocrats® measure the performance of S&P 500 companies that have increased dividends every year for the last 25 consecutive years.

As to iShares DVY, it appears that this is more a "high dividends" strategy, cf. this quote:
The iShares Select Dividend ETF seeks to track the investment results of an index composed of relatively high dividend paying U.S. equities.

And finally, as to the Vanguard hodgepodge, Vanguard VDIGX is an old active fund, issued in 1992, it pursued a strategy centered on Utilities to begin with, then moved more clearly into the Dividends Growth category in 2002, while remaining active. In 2006, Vanguard decided to create a purely passive corresponding fund (VDAIX). And also decided in 2006 to enter the High Dividends market with VHDYX. Did I get that right?

AlohaJoe
Posts: 2413
Joined: Mon Nov 26, 2007 2:00 pm
Location: Saigon, Vietnam

Re: New Dividends Appreciation data series - backtesting

Post by AlohaJoe » Wed Aug 09, 2017 10:36 pm

siamond wrote:What troubles me is that the S&P Aristocrats index displays a big 11% return since 1992, while the S&P 500 only returned 9%. That is a little hard to believe, unless somehow the S&P Aristocrats filtering procedure ended up indirectly creating a strong factor tilt (well, maybe it did?).
According to Vanguard in their paper "An analysis of dividend-oriented equity strategies"

Image

They provide the following factor regression for the S&P 500 Dividend Aristocrats index.

Image

Note that "U.S. dividend growth equities are represented by the Standard & Poor’s 500 Dividend Aristocrats Index"

User avatar
siamond
Posts: 3323
Joined: Mon May 28, 2012 5:50 am

Re: New Dividends Appreciation data series - backtesting

Post by siamond » Wed Aug 09, 2017 11:12 pm

AlohaJoe wrote:According to Vanguard in their paper "An analysis of dividend-oriented equity strategies"
Ah, thank you, thank you, this is a great paper! Here is the corresponding link. The explanations are very clear, I found multiple very useful tidbits in there, this is a MUST read.

First the mapping to the usual factors (explaining the over-performance):
the performance of these strategies has been time-period dependent and largely explained by their exposure to a handful of equity factors:
- value and lower volatility for high-dividend-yielding equities and
- lower volatility and quality for dividend growth equities.


Next, I had missed one key index in my preliminary research, the MSCI High Dividend Yield index. Which exists in 'MSCI World' and 'MSCI USA' versions (and many more), and it starts at the end of 1975, which will prove very handy for the corresponding data series, spliced with Vanguard VHDYX.

Finally, Vanguard confirmed that they view the S&P 500 Aristocrats as the historic 'Dividends Growth' index (equal-weighted), while using the more modern Nasdaq Dividend Achievers for more recent years (cap-weighted). So the data series I suggested in my first post (spliced with Vanguard VDAIX) seems to hold water.

snarlyjack
Posts: 426
Joined: Fri Aug 28, 2015 12:44 pm
Location: Montana

Re: New Dividends Appreciation data series - backtesting

Post by snarlyjack » Wed Aug 09, 2017 11:13 pm

Siamond,

This is great news.
(since we have been discussing (dividends) for months, it would be
great to have research in your data base with charts & figures)

Larry Swedroe has talked a lot about "factor investing", with scv.
But you can also get factor's by investing in:
1). Dividends.
2). Value.
3). Quality.
4). Lower Volatility.
5). Momentum.

Big blue chip stocks & medium chip stocks (that pay dividends) are some great companies.
If you look at the stocks in the High Dividend Yield Index Fund & Dividend
Appreciation Index Fund what I see is quality, quality, quality...A person could
sit on these funds for years + get a income.

I definitely think this is needed research, so we can all look at it &
input our pro's & con's and make very informed choices.

Thank you very much Siamond.

User avatar
siamond
Posts: 3323
Joined: Mon May 28, 2012 5:50 am

Re: New Dividends Appreciation data series - backtesting

Post by siamond » Wed Aug 09, 2017 11:24 pm

snarlyjack wrote:I definitely think this is needed research, so we can all look at it & input our pro's & con's and make very informed choices.

Thank you very much Siamond.
You're welcome. I think I got the hang of it now (thank you for opening my eyes to those categories, which I had admittedly totally ignored so far!), and can assemble the two data series you suggested. Will work on this tomorrow.

jbolden1517
Posts: 868
Joined: Sun Jul 09, 2017 2:53 pm

Re: New Dividends Appreciation data series - backtesting

Post by jbolden1517 » Wed Aug 09, 2017 11:31 pm

siamond wrote:All right, let's step back a bit and come back to basics. Apologies, I never paid attention to such dividends-oriented stocks before.

So, from the little I understand now (this article helped), there are (at least) two possible strategies:
- one seeking stocks with recently growing dividends (that's a "dividends growth" strategy, like Vanguard VDAIX)
- one seeking stocks with established high dividends (that's a "high dividends" strategy, like Vanguard VHDYX)

And since those two strategies are fundamentally different, we might actually seek to construct not one, but TWO new data series. Correct?
Correct. Dividend growth is a more conservative GARP strategy. The goal is growth. Aristocrats is an equal weighted GARP index.

High dividend funds often buy companies with inflation adjusted negative growth, but still carrying debt. I think a good analogy is what junk bonds are to high quality corporates, high dividend funds are to TIPS. These are investors looking for high current inflation adjusted income. VGIDX which you cited is a good analogy. Utilities once established have reliable profits and often no easy expansion opportunity. They pay down debt, pay dividends and just run the company. Because of the high debt loads they were terrific inflation hedges and they had current high income. Many of the really good equity income funds were at least 1/2 utility funds. I think 1992 Vanguard was a little late the party and the shift to a more modern design makes sense. But in terms of the 1992 version that's what the investors would have wanted.

Many of these funds get extremely aggressive in their hunt for yield picking companies that are quite troubled but yield better than bonds. The higher risk stuff is going to be heavy in utilities, REITS, MLP, deep value... they are all active and quirky. CTL (Century Link) the highest yielding stock in the SP500 is a good example of the kind of stock they would aim to hold. As earnings have halved current payout ratio is 107%. CTL is going to pick up more debt and restructuring costs with the Level3 acquisition. CTL just landed a big federal contracts so they aren't going broke. And where else do you get a 10% inflation adjusted (if not more than inflation adjusted) yield?

For your purposes though these funds all suck. Fund investors are definitely buying manager alpha so not great for a backtest. The ETF space is getting interested in this market so maybe they'll create strategies which aren't so one off. For someone middle of the pack Columbia Dividend Opportunity Fund I They aim for 1.5 times the SP500's yield. They won't take on high risk stocks but they quite deliberately sacrifice earnings growth for current yield. Vanguard's VHDYX because of the low ER is quite good. It probably represents the most conservative extreme in terms of stock selection.

jbolden1517
Posts: 868
Joined: Sun Jul 09, 2017 2:53 pm

Re: New Dividends Appreciation data series - backtesting

Post by jbolden1517 » Wed Aug 09, 2017 11:34 pm

siamond wrote: Next, I had missed one key index in my preliminary research, the MSCI High Dividend Yield index. Which exists in 'MSCI World' and 'MSCI USA' versions (and many more), and it starts at the end of 1975, which will prove very handy for the corresponding data series, spliced with Vanguard VHDYX.
That seems perfect for Simba. Agree.

snarlyjack
Posts: 426
Joined: Fri Aug 28, 2015 12:44 pm
Location: Montana

Re: New Dividends Appreciation data series - backtesting

Post by snarlyjack » Thu Aug 10, 2017 12:17 am

Let's take a quick look at these 2 funds as of 8-9-2017.

The High Dividend Yield Index Fund (top 10 stocks):
Microsoft
Johnson & Johnson
Exxon
J.P. Morgan
Wells Fargo
G.E.
AT&T
Procter & Gamble
Pfizer
Chevron
405 stocks
30.6% Net Largest Holdings
2.97% Dividend
----------------------

Dividend Appreciation Index Fund (top 10 stocks):
Johnson & Johnson
Microsoft
Pepsi
3M Co.
Medtronic
United Technologies
Union Pacific
Walgreens
Abbott
CVS
187 stocks
31.6% Net Largest Holdings
1.92% Dividend

Both funds are index funds with a low ER that pay a quarterly dividend.
Managed by Vanguard. Index funds are self cleansing. Blended dividend
rate would be 2.445% as of today.

This is a good "basic" place to start the analysis.

lazyday
Posts: 2994
Joined: Wed Mar 14, 2007 10:27 pm

Re: New Dividends Appreciation data series - backtesting

Post by lazyday » Thu Aug 10, 2017 7:03 am

jbolden1517 wrote:Correct. Dividend growth is a more conservative GARP strategy. The goal is growth. Aristocrats is an equal weighted GARP index.
Maybe I don't remember correctly, but I thought to be an Aristocrat all you had to do is raise your dividend in nominal terms, by any amount no matter how small, every year, for enough years to be in the index.

I expect the Aristocrats to be more of a quality/profitability/low investment/low volatility/low beta index. GARP should be more like a momentum and value compromise.

If you can avoid dividend cuts for many years, it might mean that you have steady earnings and/or low debt. Both of which were in Grantham's definition of quality, I believe before the Novy Marx paper. Also as I recall S&P used steady or growing dividends to mean "quality" long before Novy Marx.

jbolden1517
Posts: 868
Joined: Sun Jul 09, 2017 2:53 pm

Re: New Dividends Appreciation data series - backtesting

Post by jbolden1517 » Thu Aug 10, 2017 7:15 am

lazyday wrote:
jbolden1517 wrote:Correct. Dividend growth is a more conservative GARP strategy. The goal is growth. Aristocrats is an equal weighted GARP index.
Maybe I don't remember correctly, but I thought to be an Aristocrat all you had to do is raise your dividend in nominal terms, by any amount no matter how small, every year, for enough years to be in the index.
Correct. You also have to qualify for the SP500.
lazyday wrote: I expect the Aristocrats to be more of a quality/profitability/low investment/low volatility/low beta index. GARP should be more like a momentum and value compromise.
I don't think GARP has any momentum factor. If anything I'd say it leans anti-momentum since it sells overvalued and buys undervalued.
lazyday wrote: If you can avoid dividend cuts for many years, it might mean that you have steady earnings and/or low debt. Both of which were in Grantham's definition of quality, I believe before the Novy Marx paper. Also as I recall S&P used steady or growing dividends to mean "quality" long before Novy Marx.
For many years you likely have growing earnings. These may be because you are gradually paying down debts but ...
SP500 dividend stocks with a long term record of rising dividends are generally quality. They could be going through a rough patch though that disqualifies them for quality. For example consider Verizon whom many consider their earnings quality to be in jeopardy because of over expansion. Yet they continue to raise their dividend (though slowly). They would make a dividend growth screen but not a quality screen.

lazyday
Posts: 2994
Joined: Wed Mar 14, 2007 10:27 pm

Re: New Dividends Appreciation data series - backtesting

Post by lazyday » Thu Aug 10, 2017 7:29 am

jbolden1517 wrote:I don't think GARP has any momentum factor. If anything I'd say it leans anti-momentum since it sells overvalued and buys undervalued.
20 years ago, I think some people lumped together price momentum, earnings momentum, sales momentum. In the last 10 years I've only heard about price momentum. Until this month: http://www.etf.com/sections/index-inves ... nopaging=1

I would expect price momentum and earnings momentum to be related, but don't know. There may be some real differences, the article above reports that "fundamental momentum" (earnings and other measures) does not have reversals like price momentum.

jbolden1517
Posts: 868
Joined: Sun Jul 09, 2017 2:53 pm

Re: New Dividends Appreciation data series - backtesting

Post by jbolden1517 » Thu Aug 10, 2017 7:53 am

lazyday wrote:
jbolden1517 wrote:I don't think GARP has any momentum factor. If anything I'd say it leans anti-momentum since it sells overvalued and buys undervalued.
20 years ago, I think some people lumped together price momentum, earnings momentum, sales momentum.
That's a good point. When the market was dominated by people trading the 6-18 mo earnings outlook the correlation was probably really strong. There would have been exceptions but in broad strokes they might have been the same thing mostly. I suspect the most common place where they would have been a difference (for companies with sales and earnings) would be margin contraction (or more rarely expansion). Under pressure, which is common in the USA, sales momentum and earnings momentum can diverge. Also of course tech back then would have been an exception where sales growth was often good while earnings were quite negative. Grow the sales, grow the earnings then grow the dividends. Today there is more of a check to make sure sales growth has the possibility for earnings growth.

I don't think the 6-18 months earnings traders have much control of stock prices today. Today they are just not the same thing at all. There are far too many float traders who are indifferent to earnings either realized or expected. Regardless though... IMHO GARP is a value strategy. A GARP investor is going to reject a company with good earnings growth and a stock trading "too high". That really IMHO is the big difference between GARP and growth investing, how much you care about valuation.
lazyday wrote: I would expect price momentum and earnings momentum to be related, but don't know. There may be some real differences, the article above reports that "fundamental momentum" (earnings and other measures) does not have reversals like price momentum.
Which makes sense. Earnings growth is going to be a lot stickier. That's why growth investing beats value investing in most 3 year time periods.

lazyday
Posts: 2994
Joined: Wed Mar 14, 2007 10:27 pm

Re: New Dividends Appreciation data series - backtesting

Post by lazyday » Thu Aug 10, 2017 8:11 am

jbolden1517 wrote:IMHO GARP is a value strategy. A GARP investor is going to reject a company with good earnings growth and a stock trading "too high". That really IMHO is the big difference between GARP and growth investing, how much you care about valuation.
I guess if you're a GARP investor and care a lot about valuation, you'll probably be buying mostly companies with moderate growth. So I see how you could view the Aristocrats as GARP.

User avatar
nedsaid
Posts: 8504
Joined: Fri Nov 23, 2012 12:33 pm

Re: New Dividends Appreciation data series - backtesting

Post by nedsaid » Thu Aug 10, 2017 8:19 am

snarlyjack wrote:
Larry Swedroe has talked a lot about "factor investing", with scv.
But you can also get factor's by investing in:
1). Dividends.
2). Value.
3). Quality.
4). Lower Volatility.
5). Momentum.
In light of the pie fights over dividends in many threads, and in light of Larry Swedroe's anti-dividend
jihad here on Bogleheads; the suggestion that dividends in themselves are a factor is the equivalent of fighting words. If anything would get Swedroe out of his slumber here, that would be it. The equivalent of throwing down the gauntlet. :wink:

I have liked dividends for many years. Silly me didn't understand that dividend stocks often did better because of the underlying factors and not the dividends themselves.

So here we don't body shame. But it seems like there has been a fair amount of "dividend shaming" here when people here admit that they like dividends and are interested in the income generated from their investments. I mean, don't us silly people realize we can sell shares and create our own dividend? And of course the lectures about the superiority of total return over an income approach.

These pie fights are a lot of fun. I don't know how many dividend threads I have participated in. I don't know, I just can't help it. Sort of like the bug zapper and moths. Somehow I can't resist the light. Larry doesn't post here anymore, so those of us who like dividends feel more free to post. When the cat is away, the mice will play! So pretty much, if you are bored and need entertainment, just break out a six pack of beer and read the dividend threads.

I suppose there will be a lot of back and forth. I further suppose that Siamond will discover that the better returns that we often see in dividend stocks are due to the underlying factors and not the dividends themselves. The pie fight will be a lot of fun but ultimately Siamond will probably prove Larry right. Somewhere out there in cyberspace, Larry will be smiling.
A fool and his money are good for business.

snarlyjack
Posts: 426
Joined: Fri Aug 28, 2015 12:44 pm
Location: Montana

Re: New Dividends Appreciation data series - backtesting

Post by snarlyjack » Thu Aug 10, 2017 8:57 am

Nedsaid,

Yes, I have body burns from being "tarred & feathered" a few times.

That is why I' am so excited that Siamond has taken on this project.
It would be very helpful to the Bogleheads to have analysis & charts
that we can all look at to help us make informed choices. Plus it
gives us more choices. Maybe I would like to "tilt" this way or that
way. I think with this analysis a person should be able to make
informed choices knowing all the pro's & con's without all the emotions.

I learned a lot from Larry Swedroe & I really like him. I wish he
was still posting but I do have some body burns from tar & feathers...

jbolden1517
Posts: 868
Joined: Sun Jul 09, 2017 2:53 pm

Re: New Dividends Appreciation data series - backtesting

Post by jbolden1517 » Thu Aug 10, 2017 9:10 am

nedsaid wrote: I suppose there will be a lot of back and forth. I further suppose that Siamond will discover that the better returns that we often see in dividend stocks are due to the underlying factors and not the dividends themselves. The pie fight will be a lot of fun but ultimately Siamond will probably prove Larry right. Somewhere out there in cyberspace, Larry will be smiling.
Every factor can always be expressed as a combination of other factors. The factors are just combinations of one another. The options guys are right in the end volatility is the only asset class.

Stryker
Posts: 370
Joined: Tue Nov 24, 2009 6:13 am
Location: Canada

Re: New Dividends Appreciation data series - backtesting

Post by Stryker » Thu Aug 10, 2017 9:34 am

I still remember David Dreman's study back in 2004.

"Why Dividends Matter"

https://www.forbes.com/forbes/2004/0419/106.html

Da5id
Posts: 1401
Joined: Fri Feb 26, 2016 8:20 am

Re: New Dividends Appreciation data series - backtesting

Post by Da5id » Thu Aug 10, 2017 9:45 am

nedsaid wrote: I suppose there will be a lot of back and forth. I further suppose that Siamond will discover that the better returns that we often see in dividend stocks are due to the underlying factors and not the dividends themselves. The pie fight will be a lot of fun but ultimately Siamond will probably prove Larry right. Somewhere out there in cyberspace, Larry will be smiling.
I'm a moderate dividend skeptic myself - indifferent in tax advantaged, intellectually prefer not to get dividends in taxable but don't takes steps to avoid it. That said, color me skeptical that dividends per se are going to be a way to select winning stocks. If it were the case that large and mid cap stocks with whatever automated screening for dividend rule provide better risk adjusted returns, seems like that is really easy for the hedge funds and normal investors alike to take that advantage and arbitrage it away. The stocks in question are large cap and liquid, and cheap to trade. Seems like barriers to arbitrage don't exist. I believe in factors more than dividends, but still just go 3 fund myself...

But adding the data series sounds awesome. More data is good.

User avatar
siamond
Posts: 3323
Joined: Mon May 28, 2012 5:50 am

Re: New Dividends Appreciation data series - backtesting

Post by siamond » Thu Aug 10, 2017 1:04 pm

nedsaid wrote:I suppose there will be a lot of back and forth. I further suppose that Siamond will discover that the better returns that we often see in dividend stocks are due to the underlying factors and not the dividends themselves. The pie fight will be a lot of fun but ultimately Siamond will probably prove Larry right. Somewhere out there in cyberspace, Larry will be smiling.
Well, the Simba spreadsheet is not quite intended (nor designed) for an analysis of underlying factors. I can certainly show some correlation metrics with other series (e.g. value, momentum), but only a few, and using annual numbers. I could do a separate analysis of monthly numbers (correlation metrics would be more accurate) using a separate spreadsheet of mine, but I would suggest that Vanguard's analysis seems excellent in this respect and I have no reason to doubt their findings. I just wish they had extended the analysis further in the past though. Anyhoo, I'll get to it this afternoon, I'm glad to see that the topic is of interest to many people.
jbolden1517 wrote:Every factor can always be expressed as a combination of other factors. The factors are just combinations of one another. The options guys are right in the end volatility is the only asset class.
Totally agreed with the first point. Momentum or quality or whatever are correlated with dividend stocks returns, but we could say it the reverse way around. As usual, correlation isn't causation. It does show overlap though, and there would probably be little point betting on dividends stocks while making separate bets on correlated factors. jbolden, I don't understand your last sentence though (the point about volatility?).

User avatar
nedsaid
Posts: 8504
Joined: Fri Nov 23, 2012 12:33 pm

Re: New Dividends Appreciation data series - backtesting

Post by nedsaid » Thu Aug 10, 2017 1:12 pm

snarlyjack wrote:Nedsaid,

Yes, I have body burns from being "tarred & feathered" a few times.

That is why I' am so excited that Siamond has taken on this project.
It would be very helpful to the Bogleheads to have analysis & charts
that we can all look at to help us make informed choices. Plus it
gives us more choices. Maybe I would like to "tilt" this way or that
way. I think with this analysis a person should be able to make
informed choices knowing all the pro's & con's without all the emotions.

I learned a lot from Larry Swedroe & I really like him. I wish he
was still posting but I do have some body burns from tar & feathers...
I still do own individual stocks. I have a preference for dividends but really what I am looking for is
shares in a business that I want to own for a long, long time. I would never buy a stock just for its dividend. Indeed, that is a secondary consideration. I have owned non-dividend paying stocks too. Right now, only one of my stocks does not pay a dividend.

What Larry was trying to do was to keep us up to date on academic research and to help us to invest our monies more efficiently.

As far as Larry, he is the typical New Yorker, tough on the outside but actually a very nice guy. He still takes a lot of time answering questions by e-mail and personal message. He is amazingly unselfish with his time. To my great surprise, he saw what I had posted regarding one of his articles and sent me a personal message. Pretty much, he is ever the teacher.

I do like to engage people on the forum and I have to admit debating with him here. Like Snoopy playing the World I ace on top of his doghouse, it just seemed like the "Red Baron" always got the best of it in a dogfight. My doghouse would fly back to base riddled with bullet holes and smoke coming out the back. Curses, Red Baron!

The Value threads are also interesting and I have posted a lot in those. I also got static when I admitted that I had a very relaxed attitude towards rebalancing. The reactions I got was like I didn't practice basic hygiene or like I never called my Mom on Mother's Day. Rebalancing, like cleanliness, is next to Godliness here! Seeing the evils of my ways, I have been in a program of mild rebalancing for over four years.

Mostly the "hot button" threads are in good fun and as you read through them, the usual suspects show up over and over again.

Siamond will do a very good job. His threads are always very good.
A fool and his money are good for business.

User avatar
nedsaid
Posts: 8504
Joined: Fri Nov 23, 2012 12:33 pm

Re: New Dividends Appreciation data series - backtesting

Post by nedsaid » Thu Aug 10, 2017 1:21 pm

siamond wrote:
nedsaid wrote:I suppose there will be a lot of back and forth. I further suppose that Siamond will discover that the better returns that we often see in dividend stocks are due to the underlying factors and not the dividends themselves. The pie fight will be a lot of fun but ultimately Siamond will probably prove Larry right. Somewhere out there in cyberspace, Larry will be smiling.
Well, the Simba spreadsheet is not quite intended (nor designed) for an analysis of underlying factors. I can certainly show some correlation metrics with other series (e.g. value, momentum), but only a few, and using annual numbers. I could do a separate analysis of monthly numbers (correlation metrics would be more accurate) using a separate spreadsheet of mine, but I would suggest that Vanguard's analysis seems excellent in this respect and I have no reason to doubt their findings. I just wish they had extended the analysis further in the past though. Anyhoo, I'll get to it this afternoon, I'm glad to see that the topic is of interest to many people.
I learned a lot from Nisiprius. He would check out assertions by doing comparisons with actual Vanguard Index funds using the growth of $10,000 charts at Morningstar. He also posted the graphs. It was a great exercise in looking at real life investments
and real life results.

What I am looking for is evidence whether assertions I have made are actually true. Well presented data and graphs are very illuminating. I will look at the Vanguard white paper. Thanks for your excellent work.
A fool and his money are good for business.

User avatar
siamond
Posts: 3323
Joined: Mon May 28, 2012 5:50 am

Re: New Dividends Appreciation data series - backtesting

Post by siamond » Thu Aug 10, 2017 1:59 pm

Ok, so here is the plan for those two new data series. If you have a better idea, speak up! I'll start collecting numbers in the mean time.

Dividend Growth (aka Dividend Appreciation)
S&P 500: 1985-1989 (by lack of a better source)
S&P 500 Dividend Aristocrats: 1990-2006
Vanguard Dividend Appreciation Index Fund (VDAIX): 2007+

High Dividend Yield
MSCI USA High Dividend Yield: 1975-2006
Vanguard High Dividend Yield Index Fund (VHDYX): 2007+

jbolden1517
Posts: 868
Joined: Sun Jul 09, 2017 2:53 pm

Re: New Dividends Appreciation data series - backtesting

Post by jbolden1517 » Thu Aug 10, 2017 2:05 pm

siamond wrote:
jbolden1517 wrote:Every factor can always be expressed as a combination of other factors. The factors are just combinations of one another. The options guys are right in the end volatility is the only asset class.
Totally agreed with the first point. Momentum or quality or whatever are correlated with dividend stocks returns, but we could say it the reverse way around. As usual, correlation isn't causation. It does show overlap though, and there would probably be little point betting on dividends stocks while making separate bets on correlated factors. jbolden, I don't understand your last sentence though (the point about volatility?).
Essentially the way options look at assets is you get a risk free rate and a volatility rate.
In any given year the long bond pays (again assuming you are holding it leveraged)
long bond pays = duration premium + realized change in capital gains due to changes in interest rate
where the duration premium (the "long term interest rate")
duration premium = volatility risk premium - risk free rate
That is you get paid a higher interest rate for accepting the risk of volatility in interest rates. The spread is the premium.

stock pays = dividend - money market rate + expected capital gains + realized volatility
expected capital gain is highly positive that's the premium you get for carrying the risk of the stock volatility.

A good analogy a friend of mine came up with is the wave / particle view in physics.
In the particle view: stocks, bonds are assets. And volatility futures are just another assets
In the wave view: types of volatility are the only assets and stocks and bonds are just made up volatilities.

From their point of view (which I tend to think has a lot of validity to it) all the assets classes are really just volatility premiums. Volatility is the only asset class.

lazyday
Posts: 2994
Joined: Wed Mar 14, 2007 10:27 pm

Re: New Dividends Appreciation data series - backtesting

Post by lazyday » Thu Aug 10, 2017 2:19 pm

siamond wrote:Dividend Growth (aka Dividend Appreciation)
S&P 500: 1985-1989 (by lack of a better source)
If your data sources include S&P Quality ranking, you could limit these to stocks of, for example, A- or better.

I don't know where to get that data. From a quick google, it's been calculated since 1956.

User avatar
siamond
Posts: 3323
Joined: Mon May 28, 2012 5:50 am

Re: New Dividends Appreciation data series - backtesting

Post by siamond » Thu Aug 10, 2017 3:29 pm

lazyday wrote:
siamond wrote:Dividend Growth (aka Dividend Appreciation)
S&P 500: 1985-1989 (by lack of a better source)
If your data sources include S&P Quality ranking, you could limit these to stocks of, for example, A- or better.

I don't know where to get that data. From a quick google, it's been calculated since 1956.
That is an interesting idea. You're right, plenty of quotes on the Internet along those lines:

The S&P Quality Rankings System measures growth and stability of earnings and recorded dividends within a single rank. Scores have been calculated on common stocks since 1956. S&P 500® High Quality Rankings are designed for exposure to constituents of the S&P 500 identified as high quality stocks – stocks with Quality Rankings of A- and above.

Here is what I found on Morningstar:
- S&P 500 Quality TR USD: starts in 2016 (I swear!)
- Fidelity US Quality Income TR USD: starts in 2006
- Russell 1000 Quality Factor Index: starts in 2001
- MSCI USA Sector Neutral Quality GR USD: starts in 1998 (and iShares tracks it with an ETF dubbed QUAL since 2013!)
- MSCI USA Quality GR USD: starts in 1975

I would be quite reluctant to use the MSCI data series for more than a few years as a mapping for Dividends Growth, but this could be a good candidate to fill the 1985-1989 gap, this shouldn't be too consequential overall.

Let's dive a little more, comparing the 1990-2016 numbers from S&P 500 Aristocrats index with the regular S&P 500 numbers and the MSCI USA Quality numbers:
- the CAGR is undoubtedly a better match (Aristocrats: 11.8%, S&P 500: 9.4%, MSCI Quality: 11.0%)
- the correlation numbers with Aristocrats aren't quite different (S&P 500: 0.84, MSCI Quality: 0.81)
- the tracking error numbers aren't quite different either, and quite poor (S&P 500: 6.22%, MSCI Quality: 6.53%)

I'd be happy to amend the data series if only for the better CAGR match. So here is the new proposal:

Dividend Growth (aka Dividend Appreciation)
MSCI USA Quality: 1985-1989 (by lack of a better source)
S&P 500 Dividend Aristocrats: 1990-2006
Vanguard Dividend Appreciation Index Fund (VDAIX): 2007+

High Dividend Yield
MSCI USA High Dividend Yield: 1975-2006
Vanguard High Dividend Yield Index Fund (VHDYX): 2007+

PS. I will also take the opportunity to add a Quality Factor data series with the MSCI index combined with iShares QUAL while we're at it (we already do something very similar with iShares MTUM). I don't believe there is a corresponding Vanguard 'Quality' fund?

lazyday
Posts: 2994
Joined: Wed Mar 14, 2007 10:27 pm

Re: New Dividends Appreciation data series - backtesting

Post by lazyday » Thu Aug 10, 2017 4:07 pm

I mentioned the letter grade ranking version of S&P Quality because it's partly based on dividend stability, so maybe similar to Aristocrats. Though any version of quality might be useful.
PS. and maybe I should also take the opportunity to add a Quality Factor data series with the MSCI index combined with iShares QUAL while we're at it. I don't believe there is a corresponding Vanguard fund?
I think the closest to US Quality are the Dividend Appreciation (index) and Dividend Growth (active) funds.

User avatar
siamond
Posts: 3323
Joined: Mon May 28, 2012 5:50 am

Re: New Dividends Appreciation data series - backtesting

Post by siamond » Thu Aug 10, 2017 4:31 pm

lazyday wrote:I mentioned the letter grade ranking version of S&P Quality because it's partly based on dividend stability, so maybe similar to Aristocrats. Though any version of quality might be useful.
Understood, but I couldn't find proper S&P Quality Rankings index numbers with my usual research tricks... Maybe somebody will have a better pointer. I proceeded with MSCI Quality in the mean time. Almost done.

lazyday
Posts: 2994
Joined: Wed Mar 14, 2007 10:27 pm

Re: New Dividends Appreciation data series - backtesting

Post by lazyday » Thu Aug 10, 2017 5:12 pm

I didn't expect the S&P data, just trying to be explain what I meant.

It is impressive you got the MSCI data so far back. :)

User avatar
siamond
Posts: 3323
Joined: Mon May 28, 2012 5:50 am

Re: New Dividends Appreciation data series - backtesting

Post by siamond » Thu Aug 10, 2017 6:07 pm

Here we are, I retrieved the necessary numbers from Morningstar, did some quick sanity checks (if you spot anything weird, do speak up!), and created the new data series in my working version of the next Simba update. I'll keep my working file in sync with this Google Drive location, so that some of you can run more tests as you see fit:
https://drive.google.com/open?id=0B0svR ... GVIYlBZR2s

Then I tested a few portfolios, while looking at the 1985-2016 time frame:
- 100% Dividend Growth
- 100% High Dividend Yield
- 100% TSM
- 100% Mid Cap Value (just to make a point! :wink:)
- the 'Snarly Jack' portfolio (40% Dividend Growth; 50% High Dividend Yield; 10% TSM)

Here is the usual risk/return chart (risk defined as volatility):

Image

Here are a bunch of metrics (download the Simba spreadsheet itself and check the glossary if a term isn't clear):

Image

And here is a telltale chart, using TSM as benchmark (check this Wiki page if you're not familiar with telltale charts):

Image

I will let more qualified people comment on the outcome, but I have to say that Dividend Growth has a pretty sweet track record (except in 1999!), while Dividend High Yield is disappointing. Although I wonder why Dividend Growth faltered in the past few years? Side-effect of stock buybacks? The tough reality of real-life funds? Else?

Next step: I will add a couple more Factor series to the spreadsheet - we'll be able to check correlation numbers with the dividend strategies (or anything else).

lazyday
Posts: 2994
Joined: Wed Mar 14, 2007 10:27 pm

Re: New Dividends Appreciation data series - backtesting

Post by lazyday » Thu Aug 10, 2017 6:50 pm

Maybe the Aristocrats tend to do slightly worse than the market when it's booming (or just rising) and much better than the market when it's crashing.

User avatar
nedsaid
Posts: 8504
Joined: Fri Nov 23, 2012 12:33 pm

Re: New Dividends Appreciation data series - backtesting

Post by nedsaid » Thu Aug 10, 2017 6:54 pm

Dividend Appreciation Index should be related to the Quality factor. I believe that Quality is related to Growth. So how does Dividend Growth compare to the Vanguard Growth index. I did a Morningstar Growth of $10,000 chart from 04/27/2006 to 8/9/2017. What I did was do a comparison from inception date of Dividend Appreciation Index. The Vanguard Funds are all investor class. Here is what I found:

Dividend Appreciation Index $23,298.
Growth Index $26,727.
S&P 500 $24,102
Dividend Growth $26,511

Hmmm. Dividend Appreciation Index didn't even beat the S&P 500 and was left in the dust. In these comparisons, I use Vanguard Dividend Growth and it did much better than the Dividend Appreciation Index. So the Dividend Growth fund, which is active, nearly matched the Growth Index. The Dividend Appreciation Index, what can I say? It did a really crummy job. Here active did beat passive. So my bit of eyeballing is inconclusive, the active fund tracked the Growth Index where the passive fund just was awful. My comparisons have always used Dividend Growth.

I will trust Siamond's data as it is over a longer period of time and he used the very best data sources. His work shows that Dividend Growth is a very good strategy. It had very good returns compared to the risk taken. Mid-Cap Value had better returns but also a higher standard deviation.

I would be interested to see how Dividend Growth did against Growth Index to test my thesis that Dividend Growth is a good Growth strategy. To me, the answer is yes, particularly if you use Vanguard's active version instead of the passive version. My guess is that Siamond's work would uphold my thesis. My further guess is that Dividend Growth would over long periods of time, just about track the Growth Index, showing that dividends in themselves are not a performance factor. And again, I am equating Quality to Growth even though in real life they are probably not 100% the same but close enough.

At some point, I will do a similar comparison of High Dividend Index to Value Index.
A fool and his money are good for business.

snarlyjack
Posts: 426
Joined: Fri Aug 28, 2015 12:44 pm
Location: Montana

Re: New Dividends Appreciation data series - backtesting

Post by snarlyjack » Thu Aug 10, 2017 8:03 pm

Thank you Siamond.

(Let the analysis begin)...Just a few thoughts before we begin.

1). I think we need to realize what we are comparing the portfolio's to (TSM/SP500).
2). What are the investors goals? Total return or income investing?
a). Total return is using the 4% (trinity study) for withdrawals.
b). Income investing is using dividends to support your income (not using the 4% withdrawal method).
3). Keeping the portfolio in tact or slowly liquidating it.

My first glance is the High Dividend Yield Fund & Dividend Appreciation Fund matched up
very well with the Total Stock Market Fund.

The (SJ portfolio) (snarlyjack) is:
50% High Dividend Yield Index Fund.
40% Dividend Appreciation Index Fund.
10% Total Stock Market Index Fund.
What I' am trying to achieve is "maximize" the portfolio for years to come.

A big Thank you Siamond.

Theoretical
Posts: 1063
Joined: Tue Aug 19, 2014 10:09 pm

Re: New Dividends Appreciation data series - backtesting

Post by Theoretical » Thu Aug 10, 2017 8:28 pm

This article may be of some interest.

https://www.researchaffiliates.com/en_u ... rowth.html

A lot of the high loaders on both profitability and investment (per portfolio visualizer data) are dividend growth funds and WisdomTree's dividend output funds.

User avatar
siamond
Posts: 3323
Joined: Mon May 28, 2012 5:50 am

Re: New Dividends Appreciation data series - backtesting

Post by siamond » Thu Aug 10, 2017 9:25 pm

Let me add a few more charts for the same 1985-2016 time periods. When looking at the metrics, it struck me that Dividend Growth really improved on the max drawdowns compared to others, including TSM. Here is a more detailed chart of the drawdowns that occurred in this period, first comparing Dividend Growth to TSM, then the 'Snarly Jack' portfolio to TSM. I also tried High Dividend Yield, and this one really doesn't display much value-added in this respect compared to TSM.

Image

Image

Sure enough, we shouldn't jump to conclusions, every major crisis displayed a completely different pattern than its predecessors. Still, between this pattern of contained drawdowns for total returns and the psychological effects of dividend streams, I am pretty sure that Dividend Growth investors would have breathed a bit more easily than many others.

This being said, I struggle a bit to accept that the future should rhyme with the past in the case of dividends strategies. The shift away from dividends and towards stock buybacks seems a rather fundamental change in dynamics. I am also puzzled by the fact that neither dividend strategy did particularly well in the recent years, while I would have expected quite some inflows (hence price inflation) towards dividends stocks from investors getting disgruntled with bonds and chasing yield.

snarlyjack
Posts: 426
Joined: Fri Aug 28, 2015 12:44 pm
Location: Montana

Re: New Dividends Appreciation data series - backtesting

Post by snarlyjack » Thu Aug 10, 2017 9:48 pm

Here is a article from Professor Jeremy Siegel
that you might find interesting concerning dividend stocks.
Dated 5-25-2016

Enjoy...

http://www.newsmax.com/Finance/StreetTa ... id/730621/

Theoretical
Posts: 1063
Joined: Tue Aug 19, 2014 10:09 pm

Re: New Dividends Appreciation data series - backtesting

Post by Theoretical » Thu Aug 10, 2017 10:15 pm

One thing about buybacks is that a lot of companies are doing both buybacks and new issuance at the same time, so short of a Net Buyback approach that the Russell RAFI funds use, I wonder if the buybacks are a bit mirage-y.

User avatar
JoMoney
Posts: 4313
Joined: Tue Jul 23, 2013 5:31 am

Re: New Dividends Appreciation data series - backtesting

Post by JoMoney » Thu Aug 10, 2017 11:54 pm

Despite all the hype of how well a dividend based portfolio strategy would have performed, I'll point out that there are real world funds like Vanguard (and FIdelity) Equity Income Funds, and Vanguard's Dividend Growth fund, which haven't had such spectacular results and have decades of history to look at.

Vanguard Equity Income vs Vanguard 500 since 1988

Over the past decade, we have the iShares DVY index, and the Wisdom Tree divided 'factor' funds hyped by Prof. Siegel... over the period of their existence they haven't performed anything like what the back-tested portfolios used to sell them did.
Wisdom Tree Total Dividend / iShares DJ Div Select / Total Stock Market
"To achieve satisfactory investment results is easier than most people realize; to achieve superior results is harder than it looks." - Benjamin Graham

lazyday
Posts: 2994
Joined: Wed Mar 14, 2007 10:27 pm

Re: New Dividends Appreciation data series - backtesting

Post by lazyday » Fri Aug 11, 2017 7:11 am

siamond wrote:Sure enough, we shouldn't jump to conclusions, every major crisis displayed a completely different pattern than its predecessors.
I think there was an old S&P paper that might have looked at 1956 to something like 2000 or 2005, and found that high quality (their old version of quality, based on earnings and dividends) did well in bad times. I can't find it now, or a couple years ago when I looked. I probably downloaded it as a pdf from an S&P site, maybe 10 years ago, and it wasn't brand new when I did.

And of course, we would expect quality stocks to do well in most big downturns. A company that, for example, hasn't had to cut dividends in 25 years is probably a stable and strong company with less economic risk.
This being said, I struggle a bit to accept that the future should rhyme with the past in the case of dividends strategies. The shift away from dividends and towards stock buybacks seems a rather fundamental change in dynamics. I am also puzzled by the fact that neither dividend strategy did particularly well in the recent years, while I would have expected quite some inflows (hence price inflation) towards dividends stocks from investors getting disgruntled with bonds and chasing yield.
Quality in general is another option, instead of focusing on dividend quality. Surprisingly, by relative PE, quality generally hasn't seemed to grow expensive even with recent research on the factor and new ETFs. https://interactive.researchaffiliates. ... ty-quality Also close the left panel and select Sector Neutral Quality to get closer to QUAL, and I suppose avoid some of the problems with a simplistic comparison of PE to decide if a style is cheap. Gross Profitability in the Factors section also has below median valuation. For Achievers (Dividend Appreciation), Defensive (Strategies section) might also be worth looking at. I don't know about Growth and Stability.

User avatar
siamond
Posts: 3323
Joined: Mon May 28, 2012 5:50 am

Re: New Dividends Appreciation data series - backtesting

Post by siamond » Fri Aug 11, 2017 11:25 am

lazyday wrote:Quality in general is another option, instead of focusing on dividend quality. Surprisingly, by relative PE, quality generally hasn't seemed to grow expensive even with recent research on the factor and new ETFs. https://interactive.researchaffiliates. ... ty-quality Also close the left panel and select Sector Neutral Quality to get closer to QUAL, and I suppose avoid some of the problems with a simplistic comparison of PE to decide if a style is cheap. Gross Profitability in the Factors section also has below median valuation. For Achievers (Dividend Appreciation), Defensive (Strategies section) might also be worth looking at. I don't know about Growth and Stability.
Last night, I added iShares QUAL (and the corresponding MSCI index) data series in my working version of the Simba spreadsheet while I was at it. And since we already have iShares MTUM for momentum (and the corresponding MSCI index) in there, I completed the process and also added the two other MSCI-based iShares Factor ETFs (namely USMV for minimum volatility and VLUE for Value). Note that VLUE is strongly skewed towards large value.

Then I looked at the correlation matrix between such factor funds and the dividend strategies. I also included the US market and Large Growth (somebody had a question about it). This is a rather coarse view of things, based on annual returns only (would be better to compute correlation based on monthly returns). The highest correlation between dividend strategies and factors was with Min Volatility and Value. That would be a good explanation for the solid Sharpe/Sortino ratios and performance premiums displayed by those dividend strategies.

Note that Vanguard did more advanced math for factor attribution in their dividend strategies study, and found more emphasis on Quality for Dividend Growth.

Image

lazyday
Posts: 2994
Joined: Wed Mar 14, 2007 10:27 pm

Re: New Dividends Appreciation data series - backtesting

Post by lazyday » Fri Aug 11, 2017 11:59 am

A minor clarification:

In this thread, and in the Vanguard paper linked, "dividend growth" usually refers to either the concept of dividend growth, the Dividend Aristocrats (25 years), or Dividend Achievers (10 years) which the Vanguard Dividend Appreciation index fund follows.

Only rarely does "dividend growth" refer to the Vanguard Dividend Growth fund. :)

EDIT - corrected, Vanguard's fund follows Achievers not Aristocrats.
Last edited by lazyday on Sat Aug 12, 2017 1:17 am, edited 1 time in total.

snarlyjack
Posts: 426
Joined: Fri Aug 28, 2015 12:44 pm
Location: Montana

Re: New Dividends Appreciation data series - backtesting

Post by snarlyjack » Fri Aug 11, 2017 12:13 pm

Siamond,

I wanted to let you know I' am going on vacation tomorrow (Saturday) I'll be gone for a week.
"I'll be back..."

Here is a dividend study by "Heartland Funds" that you might find helpful.

http://www.suredividend.com/wp-content/ ... eturns.pdf

garlandwhizzer
Posts: 1604
Joined: Fri Aug 06, 2010 3:42 pm

Re: New Dividends Appreciation data series - backtesting

Post by garlandwhizzer » Fri Aug 11, 2017 1:14 pm

Great post, siamond, and great feedback, nedsaid, among others. Very interesting reading, good example of how much one can learn on the Forum.

My two cents worth. The time period we are evaluating is basically the last 30 years, a period of ever lowering interest rates/inflation and recently sluggish economic growth. Dividends as sources of income have become increasingly important ever since the bursting of the tech bubble, a time when dividends were at historical lows, I believe. The flight to dividends for yield hungry investors has been dramatic particularly over the last 8 years as interest rates have hit all-time historical lows. When you have a tail-wind your returns look nice.

If you believe that we have entered an long secular period of slow growth/low inflation/low interest rates, investors may well continue to value dividends quite highly. As Vanguard says in their paper on dividend strategies, returns of these strategies are highly period-dependent. If we soon enter a period of increasing economic growth/increasing inflation/increasing bond yields, dividend strategies in general might not outperform reliably, same with REITS and other strategies whose appeal is largely yield. If safe bonds get attractive yields, high dividend strategies which have inherent risk may not do so well. High dividend strategies tend to include weak companies that have high payout ratios (more financial stress during bad times) and limited profit growth opportunities (utilities, commodities, consumer staples, etc.). They tend to be much lower on quality/profitiability than dividend growth strategies which tend to be solid, growing companies with less financial stress. Therefore as siamond pointed out, in the tough times dividend growth suffers less than high dividend in terms of drawdown just as quality/profitablity suffer less in lean times than value.

All you need to know in order to pick the winning dividend, cap-weighted index, or factor strategy going forward is not so much the results of the last 30 years, but what will be the big macro-economic trends of the next 30 years in terms of interest rates, inflation, domestic and global macroeconomic growth, etc.. No one knows that reliably up front. You put your money where your belief system is and hope.

Garland Whizzer

snarlyjack
Posts: 426
Joined: Fri Aug 28, 2015 12:44 pm
Location: Montana

Re: New Dividends Appreciation data series - backtesting

Post by snarlyjack » Fri Aug 11, 2017 2:01 pm

Siamond,

One last study for you before I leave on vacation.
"The Importance Of Dividends" dated 11-6-2016

Enjoy...

https://www.thebalance.com/the-importan ... nds-416840

Post Reply