My favorite reading site for individual investing is Seeking Alpha.  Certain authors bring out tons of comments. Recently, one of the better authors touched on the commonly debated question of whether it makes more sense to purchase lower yield, higher dividend growth companies or higher yield, lower dividend growth companies, or both. This is a topic that interests me a lot, because it can inform the next purchase, or drive the evolution of a portfolio, even if you have eschewed pure growth stocks, or low yield stocks that don't appear to have any intention of growing their dividend.
This post is an attempt to get the issues down on paper and examine them

First there are the things that we can all agree on;  head to head, a higher yield stock will produce more cash flow at the front end, irrespective of the dividend growth rates. Lower yield, higher dividend growth stocks take some time before their rising earnings result in an absolute cash payment that matches or exceeds the payments of the higher yield stock.

Second, you can model that tortoise/hare thing using simple spreadsheets or tables. The point where both investments are paying the same cash output is typically a fairly long time in the future. The point at which the lower yield stock actually has produced more dollars in dividend payments is even further out in the future.

In order to reach that point of dividend equivalence, certain things have to happen;  First, one has to assume the earnings growth rates, dividend growth rates and dividend policies remain the same over time.  Unless these things are assumed, there's no way to know if and when that parity will be reached. For a low yield, high dividend growth company to exceed the higher yield company in dividend dollars per share paid, or total dividends paid, it's earnings growth needs to be much higher and one would expect it's capital gains would produce a much larger position for the same dividend production. Those who advocate this kind of stock for investors who have a long horizon and no need for a great deal of cash flow are essentially advising for total return and not assuming that dividends will contribute all that much to total return.

Second, one has to choose whether dividends will be taken as cash, used to reinvest in the same stock or other stocks. If the higher yield stock is completely reinvested in itself, the point of dividend parity is pushed even further into the future, irrespective of what occurs with the lower yield company's dividend.

Which assumptions are more likely to be true over the long haul?
A low yield, higher dividend-growth stock will continue to behave the same over the long haul?
A higher yield, lower dividend-growth stock will continue to behave the same over the long haul?

What about that low yield, high dividend growth company? High earnings growth supports a high dividend growth rate, but high earnings growth eventually yields to the law of large numbers, when the growth prospects moderate. High dividend growth eventually produces a higher yield.  The low yield, high dividend growth company becomes a higher yield, lower dividend growth company. If one depends on the board-room to continue raising the dividend, eventually that comes increasingly from a rising payout ratio. At some point, the terminal growth rate and terminal dividend yield should be fairly constant, at the rate of inflation. Essentially, that means there is no longer ANY inflation adjusted growth. In that circumstance, the only way one can build wealth is to reinvest dividends and compound the number of shares owned. If one reinvests shares in this company during it's rapid growth phase, one accelerates the total return, the dividend growth and the growth in share count, but it starts from a very low, almost insignificant point, so most of the increase in value comes from earnings growth and the P/E assigned by the market. As the company becomes that mature, lower growth company, dividend payments become a larger percentage of the total yield and retained earnings no longer lead to higher growth, or higher dividend growth rates.

What about that higher-yield, lower dividend growth company?  First, one needs to look at the payout ratio. In the extreme, a company could be a cash generating juggernaut, but not growing at all. In that case, it might have a 100% payout ratio, delivering all earnings to the owners as dividends. There is no dividend growth whatsoever. Earnings yield equals dividend yield equals total return.

What about the company that pays out 80% of earnings and grows slowly? That company compounds it's earnings internally slowly. As such, it's dividends grow slowly, as most of earnings are already distributed and it can't grow any faster than the small fraction of retained earnings allow it to grow.  If that company retains earnings and continually purchases other companies, it may be able to grow it's top line that way. Depending on how it manages the merged shares, it may produce rising dividends per share, so the owner sees rising dividends. The company may also repurchase shares, driving the dividend per share up by distributing earnings amongst a declining share count.
Irrespective, if dividends are taken as cash, the dividend growth is anchored by earnings growth and the company's policy around repurchasing shares and issuing stock as incentive pay to insiders.

If one needs every cent of dividend payment to pay expenses, then the owner is dependent solely on composite performance of earnings and board-room decisions for growth in yield. If one sells shares to generate income, the ability of the portfolio to generate more income is impaired to that degree.

If some dividends can be turned back into new purchases, share count rises and dividends also rise.
If one is in the accumulation phase, all dividends can be turned back into share accumulation. So, if that is occurring, which horse wins the race, the one that grows more rapidly or the one that churns out more cash? It depends on your definition of winning, and how far it is to the finish line from the start line. If the model runs to infinity, the lower yield, higher dividend growth company will always win. However, in the real world, there is no endless growth. For the individual investor, there is a finite horizon. At some point, the investments must become the replacement for the paycheck.

The sweet spot for most investors is a level of savings such that dividend payments cover all living expenses and still allow for reinvestment as well. Few individuals escape gravity in this fashion.

So, should one take dividends as cash? Should one bank on steady conditions over many years and gamble that the lower yield, higher dividend stock will eventually catch and outrun the higher yield, lower dividend growth company and deliver adequate cash on which to live?   Would higher capital gains and selective liquidation of stock be a better plan than trying to live solely on dividends? Should one reinvest each dividend in the company that paid it? Should one collect dividends and selectively reinvest? Somewhere in-between? For each investor the answer will be different.

What about me?  First, I have chosen to use dividend reinvestment plans to turn all dividends back into the companies that issued them. If one is careful with the definitions, one can distinguish between current yield (dividend per share), dividend per originally purchased share (yield on cost), raw dividend growth (dollars collected within the position) or composite dividend growth (dollars collected within the entire portfolio)  Some individuals need every horse to be accelerating at a minimum rate, or that horse is replaced with another. Some need the whole team to be accelerating at or above a threshold rate, or the slowest horses are retired and replaced. Some are simply happy to add horses and accelerate at the same rate, but pull a bigger wagon.  Some are willing to sell off horses as the finish line approaches.

I have generally emphasized that I'm making a portfolio bet, and any "rebalancing" that I do is with new money, buying more shares within positions to prevent any other position from becoming overly influential in the composite performance. That means collecting dividends as cash and selectively reinvesting isn't for me.

I also like the advice of one participant in the conversation to never sell, unless forced by a buyout. That demands more due diligence on the buy side, but also simplifies the decision making on portfolio holdings. I'd prefer to use the "rarely sell" approach. I have a few positions that are approaching double the mean position size. That means 4% rather 2% of the portfolio, with my roughly 50 holdings. That hardly worries me, but if one of those companies melts down, it does have a disproportionate effect on my portfolio's performance. There isn't enough new money to raise every other position to the 2x mean level and establish a new mean. So, I live with a degree of imbalance, or I sell some shares. So far, I'm living with the imbalance.

One thing to remember; high valuation moderates the impact of dividend reinvestment, as each share costs more money. Dividends are paid per share, and don't track price. They do track earnings, roughly speaking. Low valuation enhances the impact of dividend reinvestment, since lower share price results in more rapid share accumulation. Thus, the DRIP plans serve to maintain a degree of balance in the positions.

I am also very skeptical about those who predict long term high growth rates. Certain companies have achieved it (like Nike) over decades.  If one wants to predict long term growth rates, you're more likely to be successful with very mature companies that produce essential goods and services, generally with low, steady growth.  I am more interested in the near term performance than the 30 year performance, because I know I can redeploy assets if the story changes sometime in the future.

It's a fact that dividends are always positive,  that valuations can very widely, and that companies that have paid dividends over a long period of time are likely to continue paying equivalent or larger dividends,  unless something goes south with the business. That suggests that compounding in share count may be more reliable than the internal compounding of retained earnings. Companies don't necessarily allocate capital into equivalent or higher returning endeavors.  However, if that capital is given to me as an owner, I know exactly what I will do with it. I will reinvest it in either the company that paid it, or another company that I think has better prospects. Thus far, I'm not reinvesting one company's dividends into other companies, but if I did, it would still be a form of share compounding within the portfolio. My stake in a company that pays a high dividend but has low growth can still grow at a meaningful rate ( T, D, S). If the company is using cash to repurchase shares and the share count is declining as a result, it enhances the performance of my position, irrespective of the growth rate of the company.

One wise senior investor who talks alot about dividend paying stocks says that a carefully selected growth stock can produce total returns that far outstrip the impact of dividends, as long as you correctly identify the candidate.  Another wise senior investor whose portfolio is full of dividend paying stocks says that the first rule of investing is to avoid losing money, and the second rule is to always remember the first rule.  I'm not a gambler. I don't know which growth stock will run like a rocket to the moon.  I do know that I can identify companies whose long term growth and dividend policy are fairly predictable, so the chance that I will experience a permanent loss of capital is quite small.  Being a risk adverse investor, I'm betting on the steady-eddy approach, rather than attempting to shoot the moon.