Sliver of hope for dividends??

If you've been following, you know that I've been playing around with a new model that, among other things, calculates a discounted utility of lifetime consumption in the presence of a wealth depletion framework where spending is forced to income at wealth depletion. Several related posts:

• A trial run of a WDT game prototype
• Optimizing Spend Rates in a simple WDT model
• Supplement to "optimizing spend rates in a WDT model"

In this post I take another un-serious and preliminary look at the use of dividend income in the same analytical mix.  My bias from long ago was that div income was essential to my well being. My take away more recently is that it doesn't matter at all; total return is what it is as is consumption so there is little to no real impact over the long haul.  In my model I try again to make a naive stab at looking at div income in a really really rudimentary way especially now that I have flipped the WDT lifecycle model over to a probability weighted and subjective discount approach.  I thought it'd be worth another look. 


The assumptions are the same as the last two bullets above so will not be belabored or recapitulated.  The only difference is that returns are split into div and real growth components and so wealth is depleted ever so slightly slower but grows more slowly as well. 

The initial condition looks like this with the blue bars being consumption and the green bar is income from all sources...in this case div and SS.

The outcome (the sum of discounted, weighted utiles over a lifetime) looks like this for different spend rates with blue being the "before" and red(total return split) being the after:

Conclusions??

I'm kind of scratching my head on this. For all my protestations that spreadsheet models can be transparent, that is not the case here.  I dug through everything and I can't quite figure it out. Which is a fancy way of saying I am too lazy to try any harder...

Keep in mind that the results here are strictly an artifact of a spreadsheet and in no way represent anything real.  My guess is that the outcome has more to do with the way that the rows and columns and formulas are set up and the way that dividends and total real returns are conceptualized than it has to do with anything else. In other words it is a naivite effect not a real phenomenon.  Also this is all based on one artificial and not so realistic set of parameters.

On the other hand It's possible that the div income, by avoiding some wealth liquidation, even with the slower growth, even if it buys 6 months or a year, might do something to the utility calc on the margin.  The optimum is the same here but the utility calc is higher at the higher spend rates.  Modeling artifact or not, it's at least enough for me to give myself a sticky note to put on the inside wall of my head to remind myself to look at this later.