Here is another reflection after 12 years of retirement finance. Me? I feel like I took this stuff pretty far for myself, from: rules-of-thumb to formulas to spreadsheets to automated spreadsheets to R code to esoterica like backward induction and pseudo-reinforcement-learning. But I don't think, with the remove of a year or two now, that those latter efforts were as additive as I thought they were at the time. I would call them more "confirmatory" of things already known or at least more easily known.
If I think about it carefully, the most useful tools I ever found, if we ignore the scrappy human virtues like adaptability and judgement, all came very early in my journey. The earliest of them came from the Milevsky book I read back in what? 2012? "7 Equations." "But RH, those are dorky deterministic things!" Yes, but they are simple, transparent, amenable to scenarios and easy parameter variation: you can see what is going on and why. Opaque code in R or excel or VBA or Python or AI platforms really cannot deliver that kind of transparency and flexibility even thought they can take us pretty far.
Of those 7 equations, and I am not going to go back and look so this is from memory, the most useful were "how long will it last" (Fibonacci) and "how long will I live" (Gompertz). 90% of retirement finance is contained in those two questions and a retiree, and maybe even his or her advisor, could amuse him or herself for a very long time with those. Then, if one get's a little wise one can shift, by way of a little coding, from the single point answers coming from the 7Eq formulas to distributions. But we are still speaking the same language of "how long this or that" when we do that, we are just fleshing out the dimensionality a little. Then, with even a little more insight one can combine the two questions/distributions into a life-weighted portfolio longevity thing which is one of the other 7EQ formulas, the Kolomogorov eq for life probability of ruin. I mean, we are still back at the beginning of my journey 12 years ago, though. Not much new here.
So, there we are. And note here that we have already stepped out of the domain of most financial advisors because we have: 1) folded in consumption, and 2) gotten a little actuarial with longevity math. But what we haven't done is step far enough away from the domain of asset managers and accumulators because are still talking about piles of wealth and returns and ruin. Meh. This 'meh' insight lead me later to my next epiphany or additive streak which was centered on consumption utility. I wasn't and am not a huge fan of econ utility but the focus there on consumption and income in retirement, as opposed to wealth and asset management, I found to be quite useful. Also, I think the convexity of the CRRA math -- as one example of utility math, and in the sense that low consumption is punished, as is volatile consumption -- is useful even though I don't 100% trust it. And let's ignore here higher orders of risk aversion where I have, in the past, joked that the coefficient of risk aversion is probably less important than therapy at higher coefficients. Different story there.
So, now what at the end of all these years? The what is that I don't perseverate on retirement math too much anymore. I mean, I probably would if I was closer to the edge of a cliff but I don't think I am there yet. I do still do a few things, though. I keep my data. I keep a balance sheet. I run an income statement. I track my spending as a rate (and in detail too). I trust my ability to monitor and adapt. I trust my judgement. I assume (and this is fading a bit) my ability to go back to work. I presume that there will be no civil war2 and that assets will still prospectively produce value (don't look at post-89 Nikkei) and that the productive and entrepreneurial capacity of Americans will be sustained for my few years left. What I don't do is obsessively code new tools or programs anymore. And anyway, as I hinted in a past post or two, a lot of that geek-quant effort was from curiosity, yes, but I was also trying to escape an haranguing ex gf via staring at my computer longer than necessary: her loss, our win. This was not a healthy response btw. Otoh, I should thank her some day for all the tech productivity that was incentivized by that particular social regime. But, as above, that is another story. The better question is whether I should give my current gf of 4 years a hard time for the life-peace that comes from her support, affection, loyalty, stability, and care...a peace that does not really foster a drive to code giant new piles of quant-dreck. Heh.
No comments:
Post a Comment