Valuation is the more volatile component as you suspect. But dividends vary quite a bit as well. And you have to be careful about how you count dividends. When shares plummet the yield of course goes up for a fixed dividend amount. But the yield can still go up even if the dividend drops just as long as the share price drops more. So if you just look at yield you might fool yourself into thinking dividends did OK when in fact the dollar amount of dividends actually dropped a lot. I think in the 2008-9 crash dividends dropped around 20% for reference - and since price dropped more than that yields improved. But an investor still had a fair bit less in dividends after the crash.
Right, it makes getting the table I want a little harder to produce. I guess one easy way is to add another column to the table, which is "inflation-adjusted change in dividends relative to the previous year". In other words, year one total dividend payout is X, year two is Y, and year two has Y/X*100 as the value of that additional column. Or I suppose you could devise some kind of a "normalization constant" on which to base the dividend payout percentage.
Maybe my searching skills are weak or I'm not patient enough, but I can't find a table of S&P 500 or total US equity dividend returns, at least not
absolute returns. There's
this S&P 500 dividend yield graph, but that's all relative to the share price, i.e. suffers exactly the same limitation as you pointed out.
This S&P 500 Dividend chart might be what I'm looking for, but I'm not sure... It says, "12-month real dividend per share — May 2013 dollars". I don't know what that number means.
My basic AA strategy for FIRE is to basically have a 25/75 (bond/equity) portfolio, and declare myself FI when my dividend returns exceed my expenses. This is taken straight from the
Your Money or Your Life recipe, although IIRC, YMOYL suggested using long UST bonds.
Either way, the emphasis is on actual income, rather than saying "I need an investment portfolio of X dollars." I get the impression that a lot of people are using the latter metric to define FI. In particular, the 4% rule, so if your annual expenses are X, then your portfolio needs to be X/0.04, or 25X. But I believe the 4% rule has baked into it an implicit assumption of using total returns, which means selling principle to fund retirement. I personally don't want to do that, since early retirement means a much longer time horizon than a traditional retirement. It seems all the academic studies focus on a traditional retirement period (30 years or less). And what is the impact of combining a much longer time horizon (maybe 50+ years) with worst-case sequence of returns (i.e. retiring just before a major market crash)?
I know MMM's personal philosophy is that flexibility is a big part of the equation. And I don't want to discount that, but I'd also like to be able to have a more robust passive income stream available. As such, defining FI solely in terms of real dividend returns means you never touch your principle, and presumably takes some of the volatility out of forecasting (if absolute dividend returns do have lower variance than valuations). And I would
hope that, over the long-term, the valuations would at least keep pace with inflation. In other words, my grossly over-simplified view of equity returns is that the value portion simply tracks inflation, and the profit comes from dividends (and you use only the profits to fund your lifestyle).
As things are now, this basically implies a 2% SWR, or 50 times your annual expenses. Definitely moves the FIRE goalpost farther away, but may be worth it depending on personal preference.
But then again, my simplified view/strategy of equity returns isn't too different from that of rental properties. You invest some fixed amount in the property, and over a long enough time line, the property value should track inflation, if not do better. But you can't sell part of a rental property to fund your retirement, so you're forced to live only on the rental income. (OK, technically you could sell part of a property if it's a multi-unit, or you could sell part of the ownership stake to another person... but those are very high "friction" transactions relative to selling stocks... in all likelihood, not something you want to do.)