Yeah. Experts have been bickering for years, so how can a couple of bloggers resolve the debate within days?

Well, I don’t want to carry this any further, and your points are noted. For future reference, I think it’s important to recognize that everyone (and I mean EVERYONE) is taking the past-performance and leap of faith approach to investing. Indexers assume the market is pretty efficient. That’s a leap of faith even though empirical evidence has proven otherwise a la tech bubble. And what about the notion that currency exchange will even out over the long-haul? That’s a leap of faith that many people seem to hold dear.

As long as everyone is playing by the same rules, remember to look into the mirror before passing judgments.

Eugene Fama and Ken French have constructed a three factor model to explain the size and value premiums in equities and suggest that out performance of value is simply a compensation for extra risk in small and value stocks. Larry Swedroe in his book “Rational Investing in Irrational Times” devotes an entire chapter to explain that value stock out performance can be attributed to their extra risk (based on Fama and French). Others like Siegel believe the excess returns are due to inefficiency. A third camp questions even the persistence of out performance of value stocks (Malkiel and Bogle are in this camp). This is not a settled debate and despite the heated debate it is not clear (at least to me as a non-expert) who is right.

I’m in agreement with you that that rebalancing every month is dumb. The other studies (including Siegel,) however, only rebalance once a year.

>>”So, that’s one buy and one sell.”

It’s not necessary to sell. Take advantage of the higher dividends. Use them to top up the under weighted stocks. If all some stocks remain in the top decile, you don’t sell everything. Only sell the ones that got knocked out of the decile. Also, let the reinvested dividends jack up the adjusted-cost-base.

>>”why can’t I be skeptical about the net returns? When I express an opinion that the turnover is much higher and trading commissions and taxes will eat up the excess returns, all I hear is “trust me. It’s not that much more””

Okay, let’s come up with some numbers. I’m going to assume Canadian tax rules even though I know Americans pay less taxes than us; I used to live there. However I’ll use BC taxes instead of Ontario. Let me know if you object.

Assumptions:
Realized capital gain per turnover = 20%
Capital gain inclusion rate = 50%
Marginal tax = 30% (on $37k to $70k income)

High Yield Portfolio (100 stocks in $100k):
Turnover rate = 20%
Tax-drag = 0.6%.
MER = 0.12%
Total Expense = 0.72%

Index:
Turnover rate = 6%
Tax-drag = 0.18%.
MER = 0.09%
Total Expense = 0.27%.

The difference in annual drag is 0.45%.

I didn’t account for the dividend-tax-credit-refunds or the higher adjusted-cost-base due to dividends from both portfolios, although the high yield portfolio will benefit more.

In the same Siegel study, the gross out performance of high yield stocks over S&P500 was 3.09%, or a 2.64% net advantage per year without accounting for tax refunds and adjusted-cost-base.

None of the studies tell you how the portfolio is constructed (I’ll have to try and borrow Future for Investors and the David Dremen book) except one: High Dividend, Low Payout study. In page 13, this is how they tell us the portfolio was constructed: “equal-weighted decile baskets based on dividend yields as of each month-end”.

Let’s see how to actually construct such a portfolio: say a dividend investor holds just 2 stocks, TD Bank and Royal Bank. TD Bank is trading at $50 and Royal Bank at $50 this month. And the portfolio has 100 stocks of each (equal-weight).

Next month, they are trading at $52 and $48 respectively. How would you replicate the study? The total portfolio has the same total $10,000 value we started with the previous month. To have equal weights, the study would split it into $5,000 each and end up with 96 stocks of TD and 104 stocks of Royal. So, that’s one buy and one sell. They do this for every month for 26 years and find that dividends do outperformed the index.

When Tweedy Browne produce a study like this shows the gross returns of a dividend portfolio is better and you call it “evidence”, why can’t I be skeptical about the net returns? When I express an opinion that the turnover is much higher and trading commissions and taxes will eat up the excess returns, all I hear is “trust me. It’s not that much more”, which I have no doubt is true about your portfolio but isn’t true of the study. I think it’s unfair to say that my mind was already made up.

1) “If all market participants realize that high dividend yields are their ticket to beat market returns, they will take advantage of it such that it won’t work in the future.”

2)”Theoretically, a dividend investor with a large portfolio can take advantage of Siegel’s study but so are other smart investors. That is, a systematic advantage that existed in the past may not work in the future. ”

3) “But I question the claim that it is oh-so easy that anyone can do it successfully or all you have to do is follow strategy X (in this case, assemble a portfolio of high-yield stocks) and you will be successful. ”

4)”if dividend stocks did produce outperformance in the past (i.e. excess returns without more risk) what makes you think market participants won’t try to exploit it in the future now that cheap ways to capture high yielding stocks exist.”

I’ll respond for the fourth time. Most participants lack the discipline to stick with a strategy if it doesn’t produce exceptional results each and every year.

>>”…you don’t have to sell Royal Bank for the way you invest. But that’s how the study is constructed, so it is relevant to talk about how exactly rebalancing is done in the study and question where the expenses are accounted for.”

Show me the part of the study where the investor must sell and repurchase the same stock even if it remains in the top decile next year. In the unlikely event that you actually find it, it wouldn’t matter, because we’re all problem solvers. The study summarizes the gross returns without fees. In real life, we can simply mimic that by hanging on to the stock without contaminating the results.

>>”It means don’t extrapolate trends that existed in the past unless there is a good reason to think conditions exist for history to be repeat itself.”

I already presented the reason. The condition is human psychology, and it’ll repeat itself well into the future.

>>”Perhaps, we should look at the performance of dividend ETFs 10 years from now and see how they fared.”

We don’t need to wait 10 years. I already have 15-years of Canadian data in my blog.

But do you believe it would’ve make a difference anyway? One of my unanswered questions to you was under what conditions would you be convinced that high yielding stocks do beat the market over the long-haul. There’s no answer because there were none. Your mind was already made up.

I know the drill. I’d present my evidence. You’d try to poke holes in them. But if all else fails, there’s always the “past performance isn’t indicative of future performance” umbrella offense. The merit of dividend investing was refuted before the trial began. I’m in a no-win predicament.

We are talking in circles here. I totally agree with you that you don’t have to sell Royal Bank for the way you invest. But that’s how the study is constructed, so it is relevant to talk about how exactly rebalancing is done in the study and question where the expenses are accounted for.

“… it doesn’t agree with the real world”

Sure it does. It means don’t extrapolate trends that existed in the past unless there is a good reason to think conditions exist for history to be repeat itself.

“… we observe that market participants tend to chase trendy stocks”

Including dividend stocks. For example, look how they chased the Dow Dogs in the early nineties. And that’s exactly my point: if dividend stocks did produce outperformance in the past (i.e. excess returns without more risk) what makes you think market participants won’t try to exploit it in the future now that cheap ways to capture high yielding stocks exist.

Perhaps, we should look at the performance of dividend ETFs 10 years from now and see how they fared.

My full question was why dwell on obsolete expenses (ETFs and dividends,) when both indexing MERs and trading expenses are lower today? It’s more fruitful to compare gross returns, because we can then start deducting the updated expenses on both sides of the comparison.

>>”If it was equal weighted, you have to rebalance pretty much every stock, every year from 1957 onward, which means you sell every stock and buy it again.”

I mentioned before that a high yielding stock (e.g. Royal Bank) will generally remain that way for many years. Rebalancing doesn’t mean sell all your Royal Bank shares only to repurchase them back.

>>”Warren Buffett quip…”

It’s a witty statement but it doesn’t agree with the real world. Our own Canadian billionaire investor, Seymour Schulich, have read 2,500 books in his life time. I’m sure his home resembles a library.

>>”Where is the logical explanation for dividend investing is superior to the index? ”

There isn’t a logical explanation for a behaviour problem, but we observe that market participants tend to chase trendy stocks instead of good old boring dividend stocks.

>”Maybe dividend investing is a form of value investing, hence is riskier (as measured by volatility) and hence provides higher returns.”

I don’t know where you’re getting your information from. I’ve read many investment literatures before, and this is the first time I hear someone says dividend/value investing is riskier.

On page 157 of Contrarian Investment Strategies (1998,) David Dreman found that high yielding stocks lost on average 3.8% in all the bear markets between 1970-1996. The market lost 7.5%. I’m not cherry picking as this book happens to be on my table. I don’t have time to compile all the studies right now, but here’s an article that answers some of your questions:

http://www.investopedia.com/ask/answers/06/pricevolatilitydividendsvsnodividends.asp