19-20 March 2014
Chuck Carnevale and Larry Swedroe discussed dividends /and pants where money can be hold 8-)/ from contradictory points of view in the following SA papers
I think that both sides wrote not perfect articles (but I admit and strongly appreciate their efforts) and pointed some weak /IMO/ points in emotional comments to the articles (reproduced in footnote 1 below).
One aspect of Chuck/Larry debates (as well as few early 2013/2014 SA articles/comments) is dividends related risk.
"First of all and practically speaking, once I receive a dividend from a company I own, I have less money at risk precisely proportionate to the amount of my dividend check. Therefore, I understand that I simultaneously have reduced the risk of owning that stock simply because I now have less money at risk. Moreover, I did not have to sell any shares to receive that cash back, therefore, my beneficial ownership interest in the company remains intact. More simply stated, I still have all my shares."
"This concept fails to understand the simple math that shows that a homemade dividend, created by selling shares equal to the amount of a dividend, produces the same result. By selling shares you reduce your exposure to the stock's risk in the same way, and by the same amount, as a dividend does."
Larry also stressed that Fama-French & Co multi-factor models ( kind of engineering improvement of Capital Asset Pricing Model - CAMP) works quite well without dividends for explanation of stock portfolio returns.
Because dividends and their growth represent ~ 95% of total stocks return in long run and ~ 80% of return in average 5 year period (see fig. 1 below constructed from prof R. Shiller data (econ.yale.edu/~shiller/data.htm) for US stock market.
Dominance of dividends in long-term returns (well, are 5 years are really long? - not for me) is not specific US phenomenon. In some countries dividends and their growth are even stronge esp. if change of valuation is negative as shown in the fig. 2 below:
I think that something is very fishy with so-called Modern Portfolio Theory (MPT) and CAMP if it ignores dividends.
On another hand, the return is linked with risk in MPT/CAMP so if dividends are irrelevant to the return they should be irrelevant to the risk.
Unfortunately I do not have access to databases, so I'll use single example to show that MPT/CAMP cannot explain reality.
If we presume that MPT/CAMP is a scientific theory and use Karl Popper's ideas (en.wikipedia.org/wiki/Falsifiability and en.wikipedia.org/wiki/Karl_Raimund_Popper) that even single violation is enough for a theory refutation (see also footnote 2)
As it well known the prime risk is MPT and CAMP is stock price volatility. It seems interesting to compare volatility of stock of firm ABC which keeps all cash with firm XYZ which distributes all cash to shareholders. I know and own such XYZ firm - it is First Financial Bancorp from August 2011. Current yield of FFBC is 3.4%
Because FFBC is a bank (not favorite industry for some investors and academics who also dislike utilities) it is hard to find close competitors which do not pay dividends. Nevertheless OTTW , TFSL, TAYC, UBMI, CLDB, FNHM are close enough and I used this banks for comparison. I also compared FFBC with other regional US Midwest banks with similar market cap (FMBI with current yield 1.6%, PRK with yield 4.8%) P/E ratio (KLIB with yield 3%, OTTW with zero yield), P/B ratio (FITB with yield 2.1%, BUSE with yield 2.7%), and LT debt/equity ratio (FFNM with yield 1.5%, THFF with yield 2.9%) using today's (19 March 2014) data from biz.yahoo.com/p/414priu.html and www.buyupside.com/calculators/volatility....
These banks pay dividends but they keep some cash in hands (their payout ratios are below 100%).
The results are presented in the table 1
See also Footnote 3
FFBC was less volatile than ABC-type banks and than banks with similar characteristics and payout ratios below 100%. Hence from MPT/CAMP point of view risk is smaller for FFBC.
Bingo! Investors probably view risk differently than MPT/CAMP pundits (see also footnote 4).
I think that S&P database beyond FASTGraphs (favorite Chuck's tool) allows perform more massive and detailed studies. One warming - 100% payout should be the company's policy, not just an accident.
Another aspect of Larry/Chuck debates is importance of money location (in company pants or in investor pants). We are talking here about cash generated by company from its business and not strongly needed by the company.
There are few possibilities for a company with these money (I use small letters)
a) company keeps money (including investment in safe instruments)
b) company distributes money as dividends
c) company uses money for internal projects
d) company pays debt
e) company acquires other firms
f) company buys its shares back
g) company increases executives salaries and bonuses
h) company increases workers salaries and bonuses
<well I use "i", "ii", etc for different sections of this blog, so no "i" here 8-) >
j) company buys shares of competitor(s)
k) company buys shares of firms in totally different sectors (like KO buys INTC shares)
l) company supports some politicians
m) company participates in philanthropy projects.
If shareholder receives dividends (and in some cases pays taxes) there are few possibilities for the shareholder with these money (I use capital letters and keep the same letters for similar activities)
A) shareholder keeps money (including investment in safe instruments)
C) shareholder uses money for internal projects (e.g. to buy food)
D) shareholder pays debt
E) see K)
F) shareholder buys shares of the company which paid dividends (like DRIP)
K) shareholder buys shares of firms in totally different sectors (like KO dividends buys INTC shares)
L) shareholder supports some politicians
M) shareholder participates in philanthropy projects.
It seems that a person (shareholder) has less possibilities than a company but let's compare a)... M).
IMO a person can keep money (a/A) more safely than a company e.g. in FDIC protected CD (each CD has protection limit ~ 100K$ which is nothing for a company and quite good for a person). Any bond is not safe (see http://huff.to/1gr0hoq , http://bit.ly/1gr0hot and seekingalpha.com/article/1752002-the-nex...).
I love b) and I hate double taxation of dividends and frequent change of US tax rules for dividends (both are crime IMO). The reasons for my "love" are in my SA comments and blogs (esp. in seekingalpha.com/instablog/725729-sds-se... and seekingalpha.com/instablog/725729-sds-se...).
If for-profit company beleives that a project will be quite profitable it should fund the project. Investor can fund profitable and non-profitable projects.
Debt should be paid according to schedule or in accelerate rate if interest rate is higher that any "for sure" investable project.
There are too many initially hidden problems im M&A. I prefer natural grown but I trust company top management esp. if it has M&A experience.
See seekingalpha.com/instablog/725729-sds-se... for my opinion on buybacks. Shortly I'm strongly against buybacks. Based on some similar ideas I don't DRIP but it might be a good choice for an investor whose total dividends from portfolio are still small to expand the portfolio. I rarely use dividends to increase my stake in a company by "independent" share purchase.
g) company increases executives salaries and bonuses / h) company increases workers salaries and bonuses
I'm usually agaist g) and for h).
I don't have a strong opinion, there are cons and pros.
I'd be very suspicious if company buys shares of firms in totally different sectors. I beleive investor should do it.
Company must NOT support any politician. IMO supports of politicians (l/L) should be illegal for companies and for trade unions but legal for individuals and non-formal groups.
Warren Buffett has very good policy for philanthropy projects in BRK.A but most companies do it by just ONLY executives will. IMO the approach of Berkshire Hathaway should be adopted or owners of public for-profit company should vote on proposed philanthropy projects.
Therefore I prefer to have money in my pants' pocket that in company one.
What to explain?
Larry stated that FF "model explains almost all of the differences in returns of diversified portfolios" without dividends.
Well to me it is more important the stock return itself than "differences in returns". As far as I know MPT does NOT explain the stock return itself (e.g. why it is ~ 10% annual in USA). Because dividends represent ~95% of the total return (see figs 1 and 2) they are the explanation. Financial pundits might ignore dividends, label them as puzzle, say that they are irrelevant, etc... - I do not care. Fact is 95%. A good science rather explains 95% first and then explains reminder. Well probably it would be hard to create thousand of PhD theses for 95% explanation in this really simple case of dividends.
1. Somehow SA can loose comments when there are "too many" for an article. For example I do not see my initial comments to Chuck and LArry articles cited above ( but I can see my replies to comments of other readers). SO I reproduce these comments here:
My comment on "Debunking The 'Dividends Don't Add Shareholder Value' Myth"
Thank you for excellent article.
I didn't read comments yet, so probably will say nothing new below. Anyway let me play a critic...
1) Although I agree with you I think the proof your presented isn't perfect: MPT proponents should argue that earnings (your orange line) should be higher if company use all earnings to grow business and so price should follow this new orange line. This argument is valid if all projects the company performs have NPV and for most dividend-paying firms it is not the case. It is known that dividends imposes discipline on firm management not to spend money for junk projects.
2) Price behavior on ex-dividend day must be random. Although stock exchange adjust opening price, dividends are NOT news in this day. Price reacts on dividends in the announcement day if the announcement is not expected (e.g. dividend freeze for DG company or higher than several previous dividend increase, etc...). Academic studies (some of them not so bad) show that price-dividend correlations occur in couple days before the announcement /I guess because of leak of information/ and last few days after the announcement. Commonly ex-dividend day is in couple weeks after the announcement and all correlations decay at this time.
3) In my pure English bonus is extra pay due to good performance (bit.ly/1koPzQb). If company has long dividend history and not so good performance during a short period /e.g. last year/ investors could expect dividends to be paid. Of course long underperformance reduces probability of such expectations. John Lintner noted back in 1956 that dividends are sticky at least in USA. Such stickiness reduce to some extend bonus-like character of dividends IMO.
My comment on "Dividends And 'The Magic Pants' "
Thank you for good article.
Let me start from your summary
1) While dividends may comprise a significant portion of total returns, they don't add any explanatory power to future returns.
Yes dividends and their growth comprise ~95% of long term return (bit.ly/xFpiPl)
Yes dividends probably don't add any explanatory power to future returns in the framework of MPT. Different taxonomy can be created probably which would explain the result I just mentioned above. You know that FF model is not unique.
2) Approximately 60 percent of U.S. stocks and 40 percent of international stocks don't pay dividends.
By definition from Warren Buffet purchase of stocks which do not pay dividends is speculation not investment. See also bit.ly/1l045k8
2a) The total returns to investors come from capital gains.
Well paper money is not wealth. Capital losses happens…
3) Corporate dividends can be replaced with self-made dividends.
Oh Larry, Larry….. I don't know that block exists in your head but you still cannot distinguish milkman from butcher (see my early comments). Probably brainwashed course in your college was quite good and concept of time (which is absent in economy) became totally foreign for you. Take your photo made 20 years ago and look in a mirror. Does time exists? For most retired folks the problem not to outlive their money is so serious that they loss their health thinking about it and in result live shorter…. So self-made dividends proposal is at least anti-human.
NB: I assume that Larry is a financial advisor (actually he is more - he wrote several books on investing /some of them are good/) who should be aware about his clients financial wealth.
Now to the article. I'd not agree with all Chuck's statements but let's analyze yours article.
OK I don't know who say the maxim I'd like to rephrase "Prediction of expected returns makes astrology a serious science". There is signal/noise ratio in engineering and if it is significantly smaller than 1, signal (i.e. stock price) detection becomes impossible (see e.g. bit.ly/Nr421M).
"The bottom line is that while for some stocks dividends compromise a significant portion of the total return, the evidence demonstrates that their return isn't impacted by the dividend policy."
This is probably incorrect - no academic paper I aware considered different dividend policies except primitive separation pay/no-pay. If you know such papers I'd appreciate copies/links.
Pants analogy: Did you grow up in communist country as I did where government tried to convince citizens that your and my pair pants is the same pair? In capitalist countries there is big difference if dollar sits in CEO pants or in yours one. If it doesn't matter for you, can you keep your cash in my saving account? I promise return you these money (sorry without interest)….
"$1 is worth $1" - Sure? Take $30 and try to buy pair of pants in Saks Fifth Avenue and in Walmart. Do you see difference? It is the same as cash in "CEO pocket" versus cash in my pocket. Full disclosure: I own Walmart shares but buy pants elsewhere /actually in my wife's favorite store to save my time during her shopping tour/ 8-).
There is no such thing as "financial theory" yet - see bit.ly/R7h6d4
"Now, this doesn't even consider the lost interest income the company would have earned had it not paid out the dividend and invested the cash in bonds."
Investor can do the same…. Just remember that any bond is not 100 safe - even US Treasury had defaulted - see recent self comment to bit.ly/Upp1C5.
If company can invest money better than average investor it should NOT pay dividends, but good companies generate more cash than they can invest.
If company has $30 billion cash and no debt "cost of capital" is irrelevant for it because the company should borrow money ONLY at cost lower than it can earn. If person has $3M cash the person should borrow money ONLY at cost lower than it can earn.
'using the cash to buy back shares" - it is often a stupid idea - see bit.ly/XRf3gk
"The two companies have a beginning book value of $10…."
Are you sure that person can ALWAYS make home-made dividends at value?
"Company B now has a somewhat higher expected growth in earnings because it has more capital to invest."
You assume that B can invest ALL cash at the same good rate. It is rarely the case for cash-rich companies.
"They find it hard to accept even the obvious because somehow it goes against their accepted beliefs"
Oh yes, human brain isn't perfect…. For example it is quite weak with risk concept and obvious is hidden under noise - see bit.ly/102frew.
"Assume Company A and Company B have identical cash flows and have identical opportunities to invest in profitable projects. Company A pays a dividend, B doesn't. If company A wants to invest an equal amount in new projects as Company B, it may have to borrow money since it has paid out some of its cash flow"
Prudent investor understands such case and allows company to stop dividend payments for some period - see bit.ly/Nr4du0
"dividends do provide one benefit. They reduce what is called agency risk"
Not only this case is considered in academic literature. I suggested you to read books, it seems you ignored.
Let me not discuss US taxes here, I'm not a polite person 8-).
2. There are a lot of facts that show MPT/CAMP weakness, actually Fama-French developed their 1992 3-factor model to save CAMP, now 5-factor models are considered. IMO this is probably good engineering but not a science because Occam's razor principle is violated.
3. I also collected data for some of these banks from Finvis.com - see table 2 below:
4. I'm trying to understand risk in blogpost seekingalpha.com/instablog/725729-sds-se....