Thread by @ben_r_hoffman

Ideology: Changes in treasury bond yield affects availability of credit via opportunity cost.
Mechanical explanation: Changes in treasury bond yield affects availability of credit via the market value of banks' capital reserves.
🧵

Lower Yield, Higher Price

Simple bond: an IOU for $10 in a year. If I buy it for $8 now, then it's as though it will go up in value by 25% in a year, so the effective yield is 25%, the same as if I bought an $8 bond paying 25% interest.

Market Price Yield
$10 0%
$9 11%
$8 25%
$7 43%
$6 67%
$5 100%

For any specific claim on future cash, a higher market price IS a lower yield, and vice versa.

What's the opportunity cost story? Any investment with a risk-adjusted yield below that of treasuries is strictly worse for banks. So the marginal loan should have the same risk-adjusted yield as a treasury bond.

On this account, when treasury yields go down (i.e. treasury bonds get more expensive), banks sell treasury bonds and make loans until the marginal yields match at equilibrium, and when treasury yields go up (and prices go down) banks make fewer loans and buy more treasury bonds.

In practice, I think there's a very different mechanism whereby higher/lower treasury bond yields lead to reduced/increased lending: capital reserve requirements. Banks are tightly regulated and have to hold some minimum % of their deposits as "safe" assets like treasuries.

Let's say a bank has $100 in deposits, and has to hold 20% of this, or $20, in treasuries, no matter how good the other lending opportunities are. Obviously this is consistent with a situation where the return on the marginal loan is a lot higher than the return on treasuries.

The Fed buys some treasuries, raising their market price by 5%. The bank, holding the same bonds, now has $21 in safe assets. It can sell the extra $1 of treasury bonds, and use that to make profitable loans.

It works the same way in the opposite direction - if the Fed sells treasury bonds, their market price goes down - so the bank now has to buy more treasury bonds to keep its capital reserves at the required level.

This is a totally different mechanism from the "opportunity cost" story, implies very different structural features of the financial economy, but makes the same predictions about the marginal effect of Federal Reserve actions on retail credit.

Accidental branching I'm reinserting here:

When the Fed buys treasury bonds it raises their market price through normal supply-and-demand mechanisms, and thus lowers their yield. When it sells bonds, it does the opposite.

Specifically, in the opportunity cost story there is a single credit market, securities like treasury bonds are one end of a spectrum along which capital flows efficiently, and towards the other end of which are things like credit card debt and payday loans.

In the regulatory-mechanisms story there are many markets, *within* which market yields may represent an efficient capital allocation based on the risk of the underlying investment, but *between* which interest rates differ for structural reasons like the one mentioned above.

Even capital allocation within a given credit privilege class may not correspond to any underlying fundamental profitability.

Agents with access to cheaper more reliable credit can drive less privileged agents into insolvency.

Example: US railroads were built on massive credit, often competed unprofitably to try to drive the competition out of business, and only became profitable when JP Morgan organized cartelistic price-fixing.

Example: Boeing drove all but one other state-backed competitor out of business, is now Too Big To Fail, and doesn't even have to be profitable.

Uber may be a more brazen instance of the same thing - it's selected for by an expectation of future access to credit, not by a clear model of how its future profits can cover its current debts, and that's enough to wipe out competitors that don't behave that way.

There's a strong structural analogy between this and Moral Mazes style coalitional strategies that favor nihilistic confidence and malleability of narrative over the sorts of things that actually make money for a company.

This is almost certainly also the mechanism by which Boeing was hollowed out by the managerial class at McDonnell Douglas.

The story of N is a good illustration of how white-collar "bullshit jobs" are much more accessible to central-case coalitional players than to autists or members of marginal groups. Credit scores measure this, not just conscientiousness.

https://twitter.com/ben_r_hoffman/status/1365309033922584576

N couldn't keep her job if she were thinking about lending mechanically. She has to think of it as a kind of dramatic scene being played out, and she can think about ways to make the scene more inclusive, but not about the actors as though they were true agents with interests.

To generalize, there are two very different types of debt:

A specific finite commitment, to be closed out in the future
A link in a chain, backed by future, bigger promises.

The finite-commitment notion of debt corresponds to a microeconomic story about efficient investment of real capital, e.g. lending grain to someone with the ability to farm it. The link in a chain notion of debt corresponds to a macroeconomic pyramid scheme.

Another worked example of promises as the sorts of things that get backed by future, larger promises, rather than concrete fulfillment of commitments: benjaminrosshoffman.com/effective-altr…

More relevant threads here:

https://twitter.com/ben_r_hoffman/status/1334966373223641090

@threadreaderapp unroll