A simple model to think through (for flaws as well as merits): industrialization was a process with efficiency as the core competitive dimension, informatization is a process with learning as the core competitive dimension — in a coarse grained model this would almost correspond to the different dimensions in evolution; the first adaptation to the environment (efficiency) and the second adaptation to other adaptive systems (learning). Increasingly efficiency becomes table stakes, and competition, in this model, converges on red queen evolution, where it takes all the running you can do to remain in the very same place.
The fact that background extinction rates remain (somewhat) constant over large periods of time really brings this home, and suggests that we should expect to see continued failure rates remain constant in the economy. It also suggests that failure rates in the economy is the best measure of the economy’s health – so turns the focus, at least in some ways, to bankruptcy rates and the uniformity of these over the economy (i.e. they should not just end up in one sector, be clustered – or?).
Compare two ideal type economies – one has a uniform failure rate of about 25% and the other a clustered failure rate of 2.5% aggregated – which one should we invest in? And why?
The problem here is that the bankruptcy signal is noisy, and often seen not as an index of the health of an economy, but as a problem or an index of if the economy is going badly. The fundamental question of when companies go bankrupt is often answered with causes outside of the individual company, i.e. companies go bankrupt when the economy is bad.
If, instead, we could say – and I am not saying that we can today – that companies go bankrupt when their business model as a whole (including things like execution, logistics etc) fails, then the signal would be more interesting, and we should want some turnover in the economy, and look for ways of increasing it.
The European Union numbers suggest fewer bankruptcies than ever during the pandemic:
The US numbers suggest more:
But again, these are very noisy numbers and not sure there is a lot that can be read out of them – especially not on the back of a pandemic. But there seems to be something here that could be useful as a means of analysis.
Here is the long term rate in the US:
I did not find any longer than 1980, but an interesting question would be if there is a change in pattern of bankruptcies as we enter the information economy for real, and what that would mean — and if we are, maybe, looking at a shift in pattern here.
Now, of course, if you start looking at the delta things become more interesting (it is in the EU chart) for the US, here are new started businesses:
You see an intriguing divergence between newly started businesses and bankruptcies — how to understand that more in detail? One possibility is that we are also witnessing a shift to single person companies in the gig economy, so we should expect more companies in absolut numbers and so a lower percentage of bankruptcies (as these bankruptcies would in principle be personal bankruptcies and a different kind of signal). Intriguing.
There is noise there too, it turns out:
The numbers are affected by the shift in legislation to prevent abuse of personal bankruptcy (i.e. getting out of debt consciously assumed by declaring bankruptcy) — so not clear how to read or understand. Well, well — more to do, and generally: a focus on failure rates and failure modes is almost always an interesting way to examine any phenomenon. How things break / change / fall apart over time is a key to how they evolve (forgetting builds memory).