September 19, 2023

I Was Wrong About Post-SVB Treasury Market Strains- Here’s Why

I Was Wrong About Post-SVB Treasury Market Strains- Here’s Why

On March 16th 2023, the Thursday after Silicon Valley Bank Failed, I published a piece entitled “What's going on with Treasuries? Silicon Valley Bank and the incoherence of the Federal Reserve's (lack of) an interest rate policy this week.” The central premise of this piece was that a lack of forward guidance was creating uncertainty in the treasury market as participants were unclear whether the Fed would be hiking because of inflation, holding because of financial stability or even outright cutting interest rates. This uncertainty in turn, I argued back then, was causing treasury market issues. I argued it was those issues that led to a breakdown of liquidity similar to 2020 and so called “repo madness” in September 2019. There is nothing logically wrong with its central argument. The problem with my old argument is simply that it's empirically false.

No one has publicly criticized this piece, or said that it was in error. However, I’ve heard through backchannels that I was wrong. A fortunate part of my success over the last three and a half years is that I have all sorts of contacts among people on Wall Street, in government or involved with public policy in general. These contacts are people who I can run things by, and who talk to me periodically. Sometimes it's because they are sympathetic with my political inclinations, while a surprising number are simply interested in my analysis. They talk to me because they’re attracted to conversing with people who want to “get it right”. I have the great privilege now of being able to draw from, or hear from, these people periodically. That input is most valuable when I get things wrong, as I did in that piece.

The contacts I talked to thought that the impact of a lack of forward guidance I was pointing to was potentially plausible, but in practice they don’t think it's something financial market participants would think in terms of. This is interesting in itself: because if the Fed can’t even get treasury dealers to think about the impact of its lacking forward guidance, what hope is there of businesses or households doing so? Anyway, that’s a conversation for another time. So what did happen? The best guess from the people I talked to was a combination of two things:

  1. The role of high frequency trading desks and hedge funds


  1. Government securities dealers attempting to preserve balance sheet space for the bank holdings of treasuries they expected to be sold.

Let’s take these in turn.

High Frequency Traders and Hedge Funds

The role of high frequency traders and hedge funds in 2020’s liquidity freeze has gotten a lot of attention. Much has especially been made of the increasing role of hedge funds. I haven’t historically focused on this topic because I’ve felt that it missed the more fundamental issues. The key is that before the international financial regulatory agreement Basel III, dealer balance sheets could expand to an unlimited extent in order to absorb “risk free” treasury securities. I wrote about this multiple times in 2020. Hedge funds, or high frequency traders, backing out of these markets during periods of instability didn’t seem to me to be particularly surprising. Nor did I think limiting the extent their balance sheets could expand to absorb treasuries would improve liquidity, even if it did make these institutions “safer”.

I don’t think any of those conclusions are wrong, but I’ve become convinced over the past six months that it is interesting in and of itself that these actors have stepped into the breach. Some have become referring to them as “shadow dealers”. A recent interesting paper on this comes from Columbia Law professor Lev Menand and New York Fed Senior Policy Advisor Josh Younger. They focus on the history and current state of the treasury market, with a special focus on repo financing of treasury holdings. Following a lot of the alternative monetary analysis of the last 15 years, they treat repo liabilities as a form of money. That means they come to the conclusion that the U.S. has never stopped “money financing” treasury holdings.

What changed after the Fed-Treasury Accord of 1951 was instead the type of money that financed their holdings and which institutions began holding them (and emitting money liabilities). In short, we went from central banks and banks purchasing and holding treasuries by issuing settlement balances or emitting bank deposits to dealers emitting repo liabilities. The paper, “Money and the Public Debt: Treasury Market Liquidity as a Legal Phenomenon”, is well worth a read (and not just because it positively cites me multiple times, including making the argument the paper makes in 2020).

So what does this have to do with today’s topic? In reading this paper I have become convinced that I have underrated the role of these high frequency trading desks and hedge funds. The core issue is one of monetary design, but these institutions reflect that lack of monetary design. Rather than being sideshows, their behavior in recent years reflects the opportunity to partially take the place left behind by more rigid balance sheets of dealers, who remained part of bank holding companies. At the same time, since “shadow” dealers are further outside the reach of the “regulatory perimeter”, they don’t have the same obligations as licensed dealers. They don’t have to regularly participate in treasury markets, and provision liquidity. Nor do they have access to the backstops that licensed government securities dealers have.

These HFT desks and hedge funds run from treasury markets at the first sign of trouble, and that’s precisely what they did after SVB failed. On the other hand, I was not wrong in 2019 or 2020 to think that regulating these entities more wouldn’t solve the fundamental problem. Making their balance sheets more rigid will just drain liquidity from treasury markets further, which would need to be offset one way or another. One way is the discount window, and the other way is to increase treasury purchases by the Federal Reserve. As we discussed in the UK context last year, they wouldn’t be “QE”. They would be “market functioning” purchases which were done regularly before the laxity of balance sheets under Basel II made them far less necessary. Meanwhile, bringing them into the regulatory perimeter but allowing them to keep their balance sheet flexibility would be a strange choice. If you are going to do that, why not simply provide more flexibility to existing dealers?

Balance Sheet Space For Licensed Dealers

Which brings me to the second reason I’ve heard back from wall street and other contacts: the licensed dealers. This is far more straightforward than the first reason. Following their expectation that banks were going to dump a lot of treasury holdings in an attempt to meet anticipated runs, dealers expanded the distance between their bid prices (the prices they are willing to purchase at) and their ask prices (the prices they are willing to sell at). This is known as their spreads, “dealer spreads” or straightforwardly “bid-ask spreads”. Since spreads are a critical way that participants and observers judge the state of liquidity, this by definition goes a significant way to explaining the treasury market dysfunction.

The other interesting thing about this point is that it means one part of my analysis in that March 16th post was half right. I thought the Bank Term Funding Program (the 13(3) facility the Fed launched) would keep banks from selling treasury securities. I was right about that. What I was wrong about was the balance of its liquidity impact. Writing feverishly and in a rush, I didn’t spend enough time thinking about whether chartered banks were significant dealers in government securities on the “sell side”. If they aren’t (and they aren’t), then keeping them holding treasury securities doesn’t impede the typical circulation of treasuries.

On the other hand, I didn’t realize that dealers were increasing spreads on the expectation of big selloffs and BTFP, by making selloffs less likely and more gradual, would lead them to shrink their spreads again — and improve overall liquidity. Credit where credit was due, they responded appropriately to the situation on a technical level even if the disturbing political economy issues remain and the monetary design issues are completely unresolved. I still think forward guidance should have been provided when SVB failed. But “no harm, no foul” as they say since it didn’t impact treasury market liquidity because treasury dealers, both licensed and “shadow”, don’t pay any attention.


It may seem strange, but being wrong in this way made this piece quite engaging and exciting to write. Being wrong can be a painful or embarrassing experience — but sometimes it's quite an enlightening and thought provoking one. This case is certainly an example of that for me. I also think writing about when I’m wrong helps readers understand the debates I’m engaged in, better than my opinion on its own. Revisiting and reassessing something I wrote in the midst of a crisis is also enlightening , so I think I will also be revisiting some of my writing from spring 2020 over the rest of the year.