Thread by @policytensor

If the term spread predicts looser financial conditions, why did QE work?

Macrofinance thread 👇

1) Had an interesting discussion with @RobinBrooksIIF on the implications of higher yields on financial conditions and economic activity. What prompted my tweet storm was this tweet:

https://twitter.com/RobinBrooksIIF/status/1371904040871804931?s=20

2) A greater term spread is assumed to imply higher cost of borrowing and therefore should dampen financial conditions, as Robin assumed. But this is not the case.

3) Wider term spread is contemporaneously correlated with looser financial conditions. On the y axis is National Financial Conditions Index from the Chicago Fed; on the X is the term spread (T10Y2Y).

4) Wider term spread also predicts looser financial conditions next week. F=37.99, P<0.0001 in the Granger causality test.

5) A wider term spread predicts looser financial conditions because of the channel identified by Adrian and Shin's classic paper on why yield curve inversion predicts recessions.

6) The logic: higher term spread -> higher net interest margins -> higher bank net worth -> greater ease of finance.

7) The Adrian and Shin logic works in reverse: yield curve inversion -> lower net interest margins -> lower bank net worth -> tighter financial conditions.

8) What about the cost of capital logic that @RobinBrooksIIF assumed dominates? Don't higher yields imply a higher cost of capital?

9) No. Firms cost of capital are completely independent of bond yields. Yes, the cost of external finance matters, but in only as much as it contains information on general credit conditions.

10) In fact, while yields have steadily collapsed …

11) The hurdle rate of return has increased over the same period. From Sampson and Shin (2020), onlinelibrary.wiley.com/doi/full/10.10….

12) OK, so if the bank capital channel completely dominates the cost of capital channel and the term spread predicts easier financial conditions, then why did QE work? Wasn't the point of QE to push down the yield curve?

13) That's where it gets interesting. There is not a shred of doubt that QE worked. But it worked despite the fact that it lowered banks' net worth. Lower spreads eroded banks net worth margins, making them less willing to lend. Yet, credit conditions eased. Why???

14) The answer is that QE worked not through lowering the PRICE of duration risk but by lowering the QUANTITY of duration risk in the market.

15) The logic is that the Fed pushed out real money investors further out along the risk curve by taking tens of billions of dollars. Yield hungry institutional investors had no choice but to chase ever riskier assets, thereby generating a financial boom.

16) The asset price boom in turn directly eased financial conditions since credit conditions are coupled with the valuation cycle through the collateral channel — more can be lend against the same collateral when its value goes up.

17) With their higher valuations and the compression of credit risk premia, firms found it easy to fund expansions. This stimulated economic conditions to an extend.

18) But economic activity and firms' investments were demand-constrained, not credit constrained. And Bernanke's 'wealth effect' barely stimulated demand because risk assets are owned by the rich who hold on to capital gains rather than spend it.

19) This ties in to yesterday's discussion with @nikhil_palsingh on the frustration of the technocrats, who were pushing on a string because fiscal policy was inert.

20) The moral of the story is that the dual variables, price and quantity, are relevant for credit conditions and economic activity. In particular, to understand the channels through which monetary policy works in our upside-down good-news-is-bad-news world. ~end of rant