Regulators have historically had a tough time predicting systemic bank risk ahead of time.
Bank capital ratios, for example, were not predictive of the 1929 or 2008 financial crises. Risk-weighted and raw charts:
Capital ratios are of course critical for analyzing the health of a specific bank balance sheet, but historically failed to detect system issues in the bank lending sector.
That's why they were reformed post-GFC, to make them more granular.
Historically some combination of private debt as a % of GDP, monetary base as a % of total bank loans, and the percentage of bank assets held in nominally risk-free assets, has been more predictive.
For example, the monetary base as a percentage of bank loans was historically low leading into the 1929 and 2008 deflationary debt crises.
However, it also helps to check bank Treasury levels, which were also low in both periods.
These charts for example shows the % of US bank assets held in cash and Treasuries/Agencies leading into 2008.
Reserve requirements are there to prevent bank runs. Capital requirements are there to reduce the risk of bank insolvency.
However when looking at macro data sets, low bank reserves and Treasuries combined as a % of assets tend to occur prior to banking failures.
Private debt as a % of GDP and M2 were key aspects as well.
Unsurprisingly, high private leverage in an economy AND banks having low total allocations to risk-free assets, are a dangerous combination.
Basically, the 2020s banking system continues to look like the 1940s, meaning banks are stuffed full of cash and Treasuries, and rapid money supply growth is coming from fiscal deficit spending, not bank lending. lynalden.com/may-2021-newsl…
• • •
Missing some Tweet in this thread? You can try to
force a refresh
Here's a thread about social media decentralization.
A couple years ago, I tried to "like" one of Doomberg's posts about a platform, and I literally saw the heart fill up, and then drain out of it like blood. I'd never seen that before.
Twitter said me liking that was disabled:
Then, when I tried to search on Twitter for that certain platform that apparently can't be named, the search results would replace it with "newsletter" in my search of the network, which was kind of Orwellian:
Even now, I don't say the obvious visible name of that platform that starts with an S. Since an unconscious algo might derank it.
Maybe that's not an issue anymore. Probably.
The funny thing is that I didn't even post about that platform until I couldn't. Then I did a lot.
I keep seeing the chart float around of 23 million government employees, as though that's directly cuttable by the new Department of Government Efficiency.
Keep in mind that 3 million of those are listed as federal and the other 20+ million are state/local.
A thread. 🧵
Now, quantifying the actual federal workforce is actually nontrivial.
-Are we talking civilian, or military too (1.3M)?
-Are we including postal workers (550k)?
Along with Steve Lee @moneyball and Ren @0xren_cf, I co-authored a paper that analyzes the process and risks of how Bitcoin upgrades its consensus rules over time, from a technical & economic perspective.
Bitcoin is hard to change by design, and the methods of how it changes have evolved as the network has grown.
In the paper, we analyze what consensus is, and how different types of entities have different incentives and powers during the course of a potential consensus change.