Fun little artifact from the "Spending, Rules and Habit" post

I don't know if this is just a coincidental, statistical artifact kind of thing or something more meaningful but thought I'd compare -- from the last post Spending, Rules and Habit -- the following:  

(a) The mean real spend from the "close to optimal" scenario using the fake parameters (50k spend on $1M with .05/.12 r and sd and a little bit of spend adaptation) used to generate one of the illustrations in that last post linked above. I took the mean real spend for each of the first 30 periods of the simulation scenario mentioned in that post.  Might be better to use median. Also, note that the distribution is defective btw because some percentage of the spends "fail" or snap to income at some point which creates two distributions, the regular one and the ones that fail.  Mean probably makes less sense in this case but at least it makes a pretty picture below, and   

(b) the Blanchett  (2016) regression formula that fits, or tries to fit, observed evidence for how people spend in real life in retirement.  I was just curious but now the overlay looks interesting. Since the regression model is sensitive to the level of spend I used 50k since that is what we did in (a). Here is the Blanchett (2016) Formula (p 15):  

Charting (a) (orange) and (b) (blue) over 30 periods, it looks like this when rendered in absolute real dollars

Does this mean anything? Probably not but if it did, maybe it means that real retirees are not all that irrational when they try to spend in practice over a lifetime.  Maybe one doesn't need a PhD to get it right after all...  

----------------------

Blanchett, D. (2016), Estimating the True Cost of Retirement, Morningstar