The average lifespan of a company is 10 years.[0] The average lifespan of companies on the S&P is maybe a little more than twice that. Each year you have a 1/100 chance of seeing a hundred year flood. With a short lifespan the odds are decent that your company simply won’t see one. The question is then: how good of an idea is it to spend to be robust to them? Of course it is viciously circular: planning to not be robust to white swans is probably why the lifespans are so short and shrinking.
As for the turn to founder control, one counterexample is Tim Cook who doesn’t have voting control of Apple but is sitting on a massive shock absorber made of cash earning insanely low rates of return. The idea that you need control of the board to be robust doesn’t seem to be necessary. It is probably helpful, apple may be sui generis, but it isn’t necessary.
It should all come down to a cost/risk/reward/importance factor.
If you are a bakery that sells very close to 100 loafs a day, you would want to make about 100 per day. Some event happens randomly and you have demand for 400. You could call it a supply chain failure or you could call it a success to turn away 300 customers once a year vs throwing out 300 loafs every other day.
At the end of the period, all of this redundancy spend gets binned so you gained nothing from it. Some essential services like healthcare could make sense to spend extra on redundancy since turning away 300 "customers" is a lot more serious than not getting your loaf or gaming pc.
I literally worked this scenario. We procured extra proofer and oven capacity as compared to what we needed daily. The cost was marginal, there was some labor cost savings (less cleaning) and added resilience.
But on those busy days, the store would turnover 4-5x when competitors were out of product.
I'm not sure if Cook is a counterexample. Is it really a shock absorber for their supply chain?
Cash on the balance sheet is not inventory. Cash won't make the bill of materials magically appear; if a weather event knocked out 5nm production at TSMC, I'm not sure any amount would be able to address their supply chain woes. There just aren't enough ASML machines in the world, let alone everything else in a state of the art fab.
I wouldn't rule out their ability to build a fab in some sort of incredible Manhattan Project-esque feat with their outstanding funds, but at that point you'd have to ask "at what cost?" Then again, maybe that's a more efficient way they could use their cash than sitting on it.
Your raise an interesting question about whether private firms are the right actors to invest in long-term infrastructure. Alternatives include governments (whether via policy or public infrastructure) and also longer-lived industry consortiums.
With that said, citing averages on companies can be pretty misleading. For example, family-owned businesses tend to last longer, with 1/3 reaching 60 years [0]. That naively suggests a mortality rate of ~1.8% per year, not that far from the risk of a 100y storm. There may be other categories of private firm that are more optimized for longevity, or it may be possible to structure them in order to make better infrastructure actors.
All good points. The age of firms is definitely fat tailed. Further, it is hard to count firm survival (mergers don’t obviously destroy firms, is IBM the same company it was, etc). Also what a firm is has changed over time (legally, structurally, etc).
It would be interesting to see how many family generation changes family firms can survive. Obviously there are very old firms (some of which are closely tied to the gov: the British Royals, Aramco) but I would guess you have a step-like survival function.
Private firms (public or not) can and do build long-lived infrastructure. High voltage power lines are an example (maybe not a great example since they are heavily regulated). Expensive infrastructure has a long-term financing problem because DCF with any reasonable discount rate will essentially turn all earnings 30 years from now into nothing. The persistence of some government that will recognize the debt (or persist some legal entity that will receive the cash flow to service it) makes possible long-term financing. One solution is to make infrastructure cheap but that is hard (and maybe only possible in software).
Governments need to be built up to handle 100 year floods. Businesses standard life time don't last long enough to spend a ton of resources on betting towards those kinds of time scales.
Sadly the incentives government officials face also often favor short time horizons. Therefore I favor any marginal increased investment for 100-year floods from any part of society
Government can’t even let private sector build houses for 1 year demand. I don’t forsee them being able to manage anything on 100 year timescales.
Just look at the climate change stupidity, any city in the world could build a 6 foot sea wall and give the middle finger to climate change in a year with minimal cost but instead we have all manner of stupid at an international level.
I used to live in a city that was in a rainforest and would run out of water like clockwork during the two months of the year it wasn’t bucketing down rain. It even had mountains with snowpack as a backup but they could never figure out how to build a dam an extra 6 feet higher so we wouldn’t run out of water. We had enough money to buy people heroin but couldn’t figure out how to not run out of water. There was never a heroin shortage, always plenty of that, they’d just buy more but if you wanted to water your lawn, get rekt bud, we don’t have any water.
To be fair regarding the heroin, it did cost about $500,000 per person to obtain enough heroin.
See walls don't work for every city in the world. Miami, for example, sits on a porous karst substrate so building a sea wall wouldn't do much to prevent flooding.
The short time horizons for governments is due to election cycles. Projects that win votes in the next election are prioritized over things that will benefit future administrations. This often leads to wildly inefficient uses of resources, such as projects being regularly cancelled and restarted, and ignorance of maintenance costs.
Running the government "like a business" is meaningless rhetoric and no one can agree on what that actually means, but typically those in favor of the idea want to privatize government services and generally reduce governmental 'bloat' where whatever policies they don't like or utilize are defined to be inefficient. Timescales are typically irrelevant to that discussion.
It probably helps that Tim Cook had a lot of supply-chain experience before he became CEO. They also ship their phones from China by air rather than sea.
I duck-duck-go'd for a source and happened upon this article from 2015 :)
Small correction: Each year you have a (1-0.5^(1/100))≈0.69% chance of seeing a hundred year flood, if I'm not mistaken (and assuming hundred-year-floods happen independently of each other).
At 1% chance each year, it would actually be 'sixtynine year floods'.
No, if they arrive independently then they are best described by a poisson process, and the expected time to arrival is exactly 1/(event probability per interval). Not sure where you got the 0.5 from?
The 0.5 derives from 'in a hundred year interval, there's a 50% chance to observe the event', which was my intuitive interpretation of a 'hundred year flood'. But as so often in statistics, that intuition was wrong -- 1/100 is indeed correct. (I'm at least not alone in that erroneous interpretation: "A common misunderstanding is that a 100-year flood is likely to occur only once in a 100-year period."[0])
The better way of looking at it is probably: In a 100*n year window, you expect to see n such events, which makes it pretty immediately obvious that the yearly incidence probability is 1/100.
As for the turn to founder control, one counterexample is Tim Cook who doesn’t have voting control of Apple but is sitting on a massive shock absorber made of cash earning insanely low rates of return. The idea that you need control of the board to be robust doesn’t seem to be necessary. It is probably helpful, apple may be sui generis, but it isn’t necessary.
[0] https://www.sciencedaily.com/releases/2015/04/150401132856.h...