This article from FT ( demonstrating how markets create tail risks. Markets improve efficiency at the cost of resilience. We saw the same problem with lean global supply chains during covid.

Efficiency means using the least amount of energy/material/whatever for what you need. But if you get 'what you need' wrong you are left exposed.

Mitigating risk requires inefficiencies by definition. This means stocking more than what you need.

#economics #finance #risk


This doesn't represent very good economic analysis. A price-taking, expected-profit-maximizing firm, faced with price uncertainty, behaves in a way that is socially efficient and in particular does not minimize likely cost as your reference seems to suggest. That is, it is completely false that "mitigating risk requires inefficiencies" for precisely the same reason that buying insurance can be, and often is, efficient and rational. Buying insurance that you don't end up using does not mean you made an ex ante mistake. It is not efficient to stock more than you will *ever* need, but efficient to stock more than the minimum you will need, and expected profit maximization leads one to do exactly that.

So if markets are producing a lack of resilience -- and they are -- is it not markets directly that are doing it, but some problem with (1) price taking (e.g. market power) or (2) expected profit maximization. Market power doesn't, to my knowledge, suggest an underinvestment in risk mitigation, so while that might be a reason, it isn't a slam dunk, and might go the other way. There are lots of reasons why firms might not maximize expected profit -- executives and their boards are short-run focused would be a plausible example; rewards to firms are asymmetric so that firms wind up being risk loving. That is, it is not markets, but market imperfections, that lead to a lack of resilience.

@mcafee Your insurance example is exactly an inefficiency. It's far more efficient not to spend money on insurance. But not spending money on insurance increases tail risks. Insurance is a great example of how some reduction in efficiency can bring about resilience.

The point is, if you are not optimizing for resilience, you won't get it. Resilience is a cost, it's often neglected in markets. This is obvious with our Just-in-Time supply chains and globalized and highly centralized production.


Insurance is not inefficient. Let me give an example. I hike and usually need one liter of water. But sometimes, say 10% of my hikes, I'm really thirsty and want a second liter very much. So I always carry two. What I've done is insure against my thirst. Provided the extra value -- which is zero 90% of the time -- is greater than the cost of carrying the water, it is efficient to insure.

You have to do the net present value calculation.

Obviously if could perfectly predict my thirst, I would not need to insure. But we don't live in a world of perfect prediction.

@mcafee Right, but you have a model that 10% of your hikes you need extra water. The extra water isn't insurance in this case. Insurance would be taking three liters just-in-case even though you never needed three litres ever.


To bring more than I will *ever* need is not efficient, yes, nor is it insurance; efficient insurance is generally between the maximum need and the minimum need. But that is precisely what builds resilience: preparation for eventualities. And that was my point: resilience is efficient and is indeed a property of competitive markets. Now I am not saying our actual markets produced resilience, for mostly they did not, but the cause was not market forces but something else, e.g. short-run performance focus or a failure to calculate the probabilities.

@mcafee You are right, market agents figure out probable ranges and optimize for that. That's what makes them fragile!

That kind of optimization is NOT resilient because it relies on these models. That's my point. Tail risks are not possible to model precisely because they are rare events.

Global Warming is shifting these probability distributions in unpredictable and chaotic ways. There is no way to, as you said, efficiently insure this.

You must necessarily be inefficient to be resilient.



Well, not the way economists use the term efficiency, but there are multiple ways one might use the term efficiency and economists don't own it.

We also apparently use the term fragile differently; I use it to mean "likely to break," so that planning for most outcomes (optimizing for probable ranges) reduces fragility. Making supply chains very long made them fragile, because dependencies were hidden. This is how we came to have a yeast shortage -- we had a shortage of packaging. Many of the shortages were of the "for the want of a nail, the kingdom was lost" sort. These sorts of fragility are readily planned for and inexpensively thwarted when interest rates are low.

I think the challenges in prediction -- which are as you suggest probably growing -- is a red herring on the efficiency of resilience. Resilience means engineering for eventualities, not over-engineering for eventualities. Hotels face random demand; appropriately sized they handle more than the average demand but less than the maximum demand. That is resilient. Investing in good forecasting aids resilience.

There is nice work by Miles Kimball

on when the response to increased uncertainty is to invest more or less in resilience.

I encourage you to mull on the extra water in my hikes -- I cast this as a known risk to make it simple, but all risks look like this: sometimes water will be more valuable than others and I have to make preparations in advance of the realization of uncertainty. Extra spending in this regard, like me carrying more water than I typically need, is exactly what insurance looks like.

BTW, markets also usually help with increased forecast uncertainty, mainly because the person with better predictions can make a lot of money and their trades spread that information. I am not claiming that market forces incent efficient investment in prediction, however.

@mcafee Yes, I think this is where the disagreement is. Economists have a very narrow use of the term.

It will be impossible to predict the consequences of global warming. We are seeing this already, for example, the abysmal failure of the energy markets in Texas during unpredicted weather.

I don't think anything good came out of increased uncertainty in Texas. People still lost power. Companies still underinvest.

Markets have shown to do horribly in times of crisis.

Sign in to participate in the conversation
Qoto Mastodon

QOTO: Question Others to Teach Ourselves
An inclusive, Academic Freedom, instance
All cultures welcome.
Hate speech and harassment strictly forbidden.