Me waiting for Irma on Sunday 9/10/17
During my recent encounter with Hurricane Irma, an experience that ranged somewhere between hair raising and terrifying, I had the opportunity to ruminate on some of the similarities between hurricane forecasting and retirement planning and not just because hurricane forecasts are really Monte Carlo simulations -- just with more complex modelling and bigger computers. I thought I'd pass along some of those reflections.
----------------
1. 100 days out to seven days out (T-100 to T-7) was like my early pre-retirement years.
In my first year in Florida (let's call it late 2008) I didn't really think about hurricanes. I didn't even know what those funny little vertical locks were in the exterior doors. Last year, though, Hurricane Mathew which ran up the entire length of the east Florida shore around 50 or 100 miles out was a big wake-up call. It was the first time I had to do those runs for batteries, water, canned food, etc. I vowed I'd do better next season which runs from late summer through November. 100 days out from Irma would have been around June. In June, I actually remembered and registered the thought that I should start getting prepared. Except that next week seemed like a better idea than this week. And then another week... You get the idea. Seven days out I could see the spaghetti models pointing a storm at me. But I was already a little late because good practice is to watch storms forming off the coast of Africa in order to get a head start. The map showed a strong storm pointed my way but potentially bending north like Matthew did (no worries, right, because last year was just peachy?). This is my approximation of the real projection ~seven days out; the yellow star is me:
Yes, of course I need to get those D batteries. I'll for sure do it tomorrow. Or maybe the day after. The affinity with retirement here is that this is an awful lot like the retirement planning I did in my 20s, 30s, 40s, and even 50s: I'll deal with it tomorrow even though I really know what is coming.
2. T-minus-5 days was like what happened when I finally started to see my retirement risk.
This is where the Irma forecast changed. My house was now being projected to be run over by what was described, with an entirely straight face on the Weather Channel, as the biggest and most powerful storm in the history of the Atlantic (Jim Cantori on the WC would add !!!!). I saw the change on the maps in real-time and drove immediately for supplies. But they, the supplies (e.g., water), were already being attacked by locals like piranha on a wounded capybara. There were no D batteries. I topped off my gas which was a stroke of genius based on what happened over the next 48 hours. The retirement-affinity here is related to my early 50s when I was already a year or so into a semi-voluntary retirement (now that really is a story) and I suddenly realized I knew almost nothing about the mathematics of retirement or retirement risk and I scrambled to figure it out (In case you are wondering, my fail rate estimates were completely red-lined when I finally looked, foolish me) This, by the way, is also where I started to binge on several years of analytics related to retirement finance which enabled the creation of most of the content on this blog. I have described my process to others by saying it was like waking up and realizing that I was standing on a tightrope in the fog. That's bad enough but when the fog burns off I'm balancing over a 10,000 foot gorge[1]. Fear becomes real. Here is the map now at T minus 5. I have not rendered it precisely but you get the idea.
3. A short lesson on constant spending.
One of the reasons I wrote this post is that I once, with my very own eyes, saw an otherwise capable and well educated and well known retirement writer describe constant-spend plans as something typically chosen by the risk-averse. This has it exactly backwards. Constant-spenders are, counter-intuitively perhaps to some, very very active risk-takers. They commit to one path and do not make changes given circumstances that are changing around them, except for maybe inflation. This is exactly like committing to a hurricane risk-mitigation plan based on T-7 forecast information and then refusing to change when new information becomes available, new information that happens to put a Cat5 hurricane path right over the top of your house. In a past post on "constant spending games" I illustrated it like this when looking at a 4% constant spend and then re-calculating forward fail rate estimates at each step as the years unfold (I don't remember the underlying return/vol assumptions but I'll assume that they were relatively reasonable)[2]:
To me this means that constant spending (at least at certain levels), like a stubborn adherence to a hurricane plan based on a T-7 forecast, is not risk-averse it is risk-accepting. It isn't really even a plan. What is it? What it is is "easy." Easy to remember, easy to budget, whatever. At the risk of insulting constant spenders I have to say that there is a fuzzy grey line here between easy and lazy.
4. Some thoughts on "uncertainty cones" and early retirement.
The difference between T-7 and T-5 reminds me of something related to early retirement. With T-7 the path is still long, the cone of uncertainty (a phrase that can be used as a drinking game trigger when watching the weather news) is wide, and the number of times that one might need to adapt in the future is very, very high. That, in a nutshell is early retirement. T-5 might possibly be better but even if it is, which it probably won't be, it won't be by much.
5. T-minus-1 felt like a retirement plan gone awry after human capital has been depleted.
Three days or so out from Irma and there was no gas to be had anywhere. I had to get up at two in the morning to search for both gas and short lines (tracker.gasbuddy.com ended up saving me; lesson learned). I remember (say thurs or fri night, I can't remember; I think it must've been very early Fri morning) burning an eighth of a gallon of gas on about a quarter-to-half remaining at 2:45 am while finding nothing anywhere near me and feeling, over the entire search, like I wanted to throw up; my palms were wet...I mean I still had kids at home[3]. I was forced around this time to make a decision on evacuation and had to assimilate a lot of disparate inputs like: statewide gas supply lines (really bad), NHC surge maps (I live on a coral ridge that runs north-south and provides some height, so good), the inclination of my significant other to evacuate (no), traffic northbound for 500 miles (usually 7 hours, some made it in 20. Also this is a long flat peninsula...so bad), risk of getting stuck gasless-foodless-roomless in rural FL when the wind hits (moderately high; my neighbor who evacuated described widespread abandonment of cars along the way after he came back), risk of social unrest on the road (low to moderate), pet friendly hotels that are also available to the north (unlikely at the last minute), structural integrity of the home (high...ish), am I in a mandatory evacuation zone (yes, but don't tell), presence of a critical mass of rational and seasoned and provisioned neighbors that were staying (yes), access to nearby height over storm surge (my neighbor in the two story house gave me his key! Thanks Brian), if evacuated do I have the ability to stay proximal for return and recovery (not great), etc. Making a bad choice among bad choices, I decided to stay but that was under the gun and too late and I probably won't make that choice again. By T-minus-1 I knew this beast was right on my back and the options became almost exactly zero. T-1 was also where those that had struck out for Fort Myers or Tampa (adaptive behavior!) based on the T-5 track realized they had to turn around and come back, except the problem was that some of them got caught flat-footed and were stuck out there for whatever reason (another after-the-fact reason to be wary of some evacuations). In retirement terms, this strikes me as vaguely akin to burning through too many resources by a late age when it is awfully hard or impossible to fix one's situation with paid employment or the charity of others. What exactly do you do? This is a little bit of a metaphor-stretch but you understand. Here is T-1 in red:
Another thing to note here is that relative to a fixed location (in the retirement metaphor we can maybe refer to the "fixed location" as a fixed planning date or horizon, something like "terminal age" or "age of annuitization" perhaps) the cone of uncertainty narrows and certainty rises over time. In my case this was relatively good news because the eye-wall was now certain to pass west of me. This redeemed part of my foolish decision but this is also like winning the lottery when one is 80.
6. The reliability of forecasts for both hurricanes and retirement are notoriously bad or fickle.
With the exception of those that are currently inside the eye-wall of a hurricane, no forecast is accurate and the further away you are from the hard-hit-date the worse it gets. Each minute after a forecast the value of the forecast depreciates precipitously. Everyone I know here in south FL knows exactly-to-the-minute when the new models are recalculated and published and each person knows and has affiliated themselves with a particular model (European model, Canadian, NOAA etc) like one might do with NFL teams or particular winemakers. Entire lives are changed and redirected and utterly disrupted based on the flow of this new information. This, my friends, is real adaptive behavior when the stakes are at their very, very highest. Here, for what it's worth, is my fake rendition of forecast reliability:
This kind of thing is why I have always considered constant spending in retirement to be an active risk-taking proposition (some retirement researchers will often footnote/disclaim their otherwise-savvy but oblivious papers by saying that no one in practice really runs themselves completely to ground). It is the adapters, I say, that are constantly seeking ways to shed or avoid risk and usually succeeding.
7. It is only rarely that a forecast will actually match the real path of a hurricane (or life).
Extending the previous point a bit, it is clear to me that forecasts are borderline-useless except when being used in a "continuous process" of adapting. Like I mentioned before, only those inside the eye-wall of a hurricane can predict with any certainty what will likely happen, though by then one wonders if it really matters all that much anymore. Even after time T-1, when certainty was relatively high, I still saw the center of Irma drift a little further east than expected (towards me! remember Wilma! and Andrew for that matter...). Almost nowhere between T-7 and T-0 did the real path follow the expected (average) path (i.e., like retirement), though the general cone forecasts were kinda-sorta ok. Here, for example, is my weak rendition of Irma's final path superimposed over the various forecasts. Note that the north coast of Cuba in particular got put through a meat grinder that the forecasts did not really ever predict.
The final comment here by me is that in both hurricanes and retirement the stakes are very high (maybe an edge to the hurricane, to be honest) and that adaptive behavior is always going to be a better choice than stubborn adherence to a constant plan. That's my plan anyway.
------------------------------------
[1] The truly wealthy don't know this fear, of course. I also have to presume, though I don't really know, that the relatively impoverished don't have this fear...or at least in a sudden way. It is generally the ones in the middle that walk a razor's edge.
[2] I suppose to carry the metaphor the blue and red lines might be like evacuating or living in Colorado.
[3] My ex thankfully evacuated them out of state.
I think you nailed it. :-)
ReplyDelete