Measuring Dividend Growth - ETFs
This article is the third (and last of the series), detailing a method for determining a dividend growth metric in the general sense. The first article, published as "Measuring Dividend Growth, Part 1: Exponential", addressed exponential (compounding) dividend growth, which accounts for the majority of stocks. The second article, published as "Measuring Dividend Growth, Part 2: Linear" covered linear dividend growth equities, exemplified by MLPs, REITs and other higher yielding equities (including Utilities and Telecommunications). The third (current) article will address ETFs, which distribute dividends from a collection of stocks (less fees). This area is unique because the population of equities contributing to distributions varies over time due to portfolio turnover.
Selecting ETFs to provide robust dividend growth is not an easy task. Most do not have a long enough track record to accurately judge. Their performance during and after the final crisis is out there for all to see. Contrast that with our own portfolios where history is out of sight. The portfolios we had during the crisis may have suffered dearly not only in value, but in dividend/interest cash flows. So can we honestly compare current portfolios with ETFs, or David Fish's CCC spreadsheets, for that matter? It would be interesting to know the 'portfolio turnover' in those spreadsheets during and after the crisis.
All this to prepare you for the shock of severe dividend cuts in the 2008-9 time period experienced by some of the ETFs discussed here
One would think that finding a good Dividend Growth [DG] ETF would be easy, just go into the marketplace and there they are. In the general sense, not true. The obvious place to look are among those that have the word "dividend" in their title. The two graphs below show the dividend record (100 shares) for 9 such candidates from 2005-2011.
ETFs featured are: SPDR S&P Dividend (SDY); Vanguard High Dividend Index (VYM); Vanguard Dividend Appreciation (VIG); Powershares Dividend Achievers (PFM); Guggenheim Multi-Asset Income (CVY); iShares Dow Jones Select Dividend Index (DVY); First Trust Morningstar Dividend Leaders Index (FDL); Powershares High-Yield Equity Dividend Index (PEY); Powershares International Dividend Achievers (PID).
The table below details some important metrics:
All data (in the table) are from Morningstar (Jul 2012). These ETFs are among those generally listed as dividend-types. I cannot say for sure if they stress dividend growth. If not, that is good because they don't, as the graphs testify. One possible exception is VIG (blue line, second graph). It may not stand out, but it didn't take a big hit during the financial crisis and recent dividend growths are OK. The problem with it is that the yield is too low, unless there is a firm commitment to dividend growth. With the low turnover, there may be hope. A quick look at specific holdings among these ETFs reveal many duplications, as one might expect. All appeared to hold top companies; PEY was an exception, with more les- known companies. I liked CVY for its mix; it had a few MLPs (Master Limited Partners) to boost yield, but turnover was a turnoff.
I think it is impossible to maintain an ever-increasing income flow, in the long term, with any appreciable amount of portfolio turnover. Think about it, for the question bears on our own portfolios. If you sell a stock because the dividend is reduced (or suspended), you have to replace it with one with higher yield to maintain cash flow because, more than likely, the stock price has also dropped.
In my opinion, an ideal DG ETF would have the following characteristics: 1) Number of holdings - 50 to 150, which provides enough for diversity, but does not have so many as to dilute performance (as in too big to succeed). 2) Yield - in today's world, about 3%, beefed up with higher yield-lower dG equities such as MLPs, Telecoms, REITs (Real Estate Investment Trusts). 3) Turnover ratio should be as low as possible. If this smells like 'buy and hold', well… Spend the effort in selecting the best, then leave them alone (just what you would do if you were hiring a person). It may be asking too much for a fund manager (even a 'passive' one) to keep hands in their pockets. 4) Dividend growth - target 10-15% with exponential growth. Once a portfolio is established, yield is less importance than DG. That is the focus.
Should we not expect to find the same investing principles in a DG ETF that we use in managing our own portfolios? After all, we have PTFs (Privately Traded Funds). Small voice stage right - "where's my bonus?" There is a big advantage over the long haul in having someone else automatically replace stocks that are slipping back. Otherwise, we have to perform that task for the rest of our lives. That is reason enough to keep searching for suitable ETFs.
If we cannot find DG in dividend ETFs (yet), what about looking elsewhere? The graph below shows dividend flow from 2004-2011 for Vanguard MSCI Emerging Markets (VWO) and Vanguard Industrials (VIS). VIS has all U.S. stocks. Graph and table data from Morningstar (Jul 2012). This graph is typical of many generic ETFs.
Average DG (rough estimate) for the 11 one-year positive dividend increases was 33%. OK, a lot was lost in one year, but was that a one-off event? Even if it happens once per decade, would that be acceptable? Surely, not a slam dunk, but not a reason to shun ETFs altogether, IMHO.
The graph below (2006-2011) has 3 more ETFs that didn't suffer much, if at all, during the financial crisis. They are: ishares S&P Global Infrastructure Index (IGF); ishares S&P Global Consumer Staples (KXI); Vanguard FTSE All-World ex-US (VEU).
Data from Morningstar (Jul 2012). Average DG (rough estimate) for the 12 one-year dividend increases was 37%, 20% without the big jump in KXI in 2008.
It should be obvious that ETFs could play a role in portfolios, if for nothing more than add diversity. Generally, they have a higher beta, which melds well with lower beta low yield-high DG stocks. Since some of the 'management' of specific stock selection/retention is outsourced, you should have less to worry about. Right?
The question arises as whether ETFs have exponential or linear dividend growth. There is not enough history to say for sure. Some plots lean one way, some the other. Theoretically, they should have exponential DG, since the underlying stocks have exponential growth. My hunch is that their growth is mixed, possibly an average between exponential and linear - which is what I am using in my retirement portfolio withdrawal studies.
ETFs have a 15% allocation in my portfolio in a 'Core' segment, 17 in number. Not all are there for DG, some are for capital gains. You know, the other term in the Total Return equation. The graph below depicts the 2007-2011 yearly income stream (re-scaled) from my Core segment. This is my current portfolio, not the one I had during, and before, the financial crisis. Yield average is over 3%.
The green line is the non-linear regression calculation, which is the 'best fit' exponential curve to the dividend data. The calculated dividend growth rate is 7.5%. The math involved is contained in Part 1 of this series.
Somewhere, sometime, somehow, someone in the ETF world will do the right thing, then we should pounce on it. Maybe it has already happened…