Quantitative tightening is easy

Good morning. Consensus has moved quickly from “this rally is great” to “this rally is great but a little spooky.” Unhedged’s view on this came in bits and pieces last week: we noted that we are seeing a broad move up, not just a tech party; that stocks are plain expensive now; that earnings are fine but not exceptional; and that sentiment is peaky. It’s really only the last point that gives us serious pause, but we are keen to hear what readers think: robert.armstrong@ft.com and ethan.wu@ft.com.

The Fed takes stock of QT

Quantitative tightening — the shrinking of central bank balance sheets — is badly understood. Until recently, there had been just one attempt at it: the Fed’s ill-fated 2017-19 QT push, which ended when the repo crisis shook funding markets. With such a skimpy record, you could reasonably imagine QT as a big disruptive force or as a minor source of background noise. No one knew any better.

But since the Bank of England began QT in February 2022, global central bank balance sheets have been trimmed aggressively. Enough has happened for us to take stock, and that was the aim of a much-watched monetary policy forum in New York on Friday, centred around a splashy new paper by MIT’s Kristin Forbes, Columbia’s Wenxin Du and Deutsche Bank’s Matthew Luzzetti. It is one of the first major attempts to assess QT’s record. And as follow-on speeches from the Dallas Fed’s Lorie Logan and Fed governor Chris Waller made clear, central bankers are listening.

The paper is long and detailed, but here is how Forbes summarises the big takeaway, in her Financial Times op-ed today:

QT programmes have, so far, been working as central banks intended. They are “in the background” and not seen as an active tool for adjusting monetary policy. At the same time, they have provided a small degree of support for central banks’ efforts to tighten financial conditions, with minimal impact on market functioning and liquidity. QT has worked in the opposite direction to quantitative easing, but the effects are much, much more muted.

In other words, QT and quantitative easing are not symmetric. One measure is the impact on long-dated bond yields. Whereas QE may depress long yields by as much as 100 basis points, QT barely tightens financial conditions. Over a one-year horizon, the authors estimate QT raises long-dated bond yields by 4-8bp.

This partly stems from when they are used. QE is crisis-time monetary policy, meant to credibly commit central banks to loosening and drive down long yields when rate cuts alone aren’t enough. QT is peacetime monetary policy. As soon as running QT becomes a problem, it can be shut off, as the BoE demonstrated during the gilts market crisis. So a new QE programme sends a louder signal to markets (Fed to the rescue!) than a new round of QT. In theory, markets could even price in the eventual onset of QT in an initial QE announcement, suggested Lorie Logan. Good news for central bankers, then. They can unwind the balance sheet without causing an asset price crash.

There is more to say, and we’ll return to QT in tomorrow’s letter. But the big point is that up to this point, QT has proven not only feasible but relatively painless. It was not obvious at the outset. As recently as 2022, prominent academics were arguing that QE might be a one-way ratchet, easy to rollout but agonising to undo. But as Forbes, Du and Luzzetti argue: “the analysis in this paper should give central banks more confidence to unwind asset purchases in the future”.

An accident could absolutely still happen as QT drains liquidity from the system. But as of now, it looks like QE is on its way to becoming a standard part of the central bank toolkit. What began as an unconventional, extraordinary intervention has become anything but. (Ethan Wu)

A reply to Uber’s Dara Khosrowshahi

Regular readers will recall that I recently wrote several disapproving things about Uber’s buyback plan. Last week, the Uber’s CEO replied to my arguments in an admirably clear email, which we published. Now I’m closing the loop, replying to his reply.

Khosrowshahi makes three points and I’m sympathetic to two of them. Responding to my view that Uber seems likely to have higher-returning uses for excess capital than a buyback, he emphasised that fixed investment is subject to operational constraints set by company culture and employee morale:

It is true that we can invest more to hire engineers to build new features that could return 20%+ ROI. However . . . I have to bet that my engineering pod cannot only return 20%+ next year, but also earn a similar 20%+ each and every year afterwards. Ship one great project, identify another, ship again, rinse and repeat. If my engineering pod fails to return its cost of capital, I can’t sell it — I’d have to lay them off. That comes with a financial cost but more importantly a human and organisational cost. It has been my experience that providing a stable, challenging but consistent work environment is hugely beneficial to our team.

This makes sense to me, and I should have thought about this more. Corporate finance nerds like myself tend to think of deploying excess capital in a simple, quantitative way. If a company has money on hand, there are six options: bank it; pay down debt; invest internally; make an acquisition; buy back shares; or pay a dividend. You simply pick the one that best maximises long-term shareholder wealth. But management is not accounting. Companies are made out of people, who need to be treated a certain way if they are to be productive.

I also agree with Khosrowshahi’s point that executives are not always good at knowing when their stock is cheap or expensive:

Non-professional-investor CEOs tend to be quite poor at buyback timing. They are typically overconfident when their stock has positive momentum (‘of course my stock price is going to rise!’) and scared/under-confident when their stock loses momentum — and in that instance they tend to hoard cash unnecessarily

Buybacks are demonstrably procyclical, meaning that companies tend to repurchase more of their stock when it is more expensive. But I would draw the opposite conclusion from this than Khosrowshahi does. That executives tend not to be valuation experts and to fall victim to bias in assessing their own stock suggests to me that companies ought to articulate an impersonal, valuation-based approach to determining when they will buy back stock. While Khosrowshahi says the company might buy back more shares if the price were to “dislocate,” overall Uber has chosen the “humble” approach of just buying consistently, “dollar-cost averaging” its purchases in hopes of creating value in the long term. 

This seems wrong to me. Buybacks only create value if completed below intrinsic value. Yes, it is hard to know what intrinsic value is, but that difficulty is not a licence to buy indiscriminately. No criteria will be perfect, of course. Until 2018, Berkshire Hathaway had a rule that they would only buy back their own stock when it was under 1.2 times book, which made sense for a conglomerate made up of mostly mature businesses (in 2018, the rule was dropped and replaced by the discretion of Warren Buffett and Charlie Munger. I suspect this happened because Berkshire’s other options for excess capital grew less attractive and Berkshire’s just about stopped ever trading down to 1.2 times).

Uber is not alone. Very few companies specify even a rough-and-ready strategy for valuing themselves ahead of buybacks, which is a bit shocking when you think about it. I can think of two explanations for this.

First, as I have argued in my pieces about Uber, is that stock-based compensation creates a strong bias in favour of stock buybacks. Companies treat stock comp as (at best) a non-cash expense or (at worst) not an expense at all, by excluding it from adjusted earnings measures. A rising share count gives the lie to this shoddy practice, by revealing how stock comp dilutes earnings. Price-insensitive stock buybacks keep the game going.

Secondly, I can see how having public criteria for buybacks might be awkward for management, because explaining those criteria almost amounts to telling investors what you think the company’s stock is worth. Berkshire pre-2018 was saying, “our best guess is that the company is probably worth a moderate premium to book, that’s it.” Executives, particularly at growth companies, might not want to make that sort of implicit commitment. It might put a lid on investor excitement about the shares, and it might be used to put management’s feet to the fire at bad moments. Managing public companies is really hard.

One good read

The risks of management by “hard numbers.”

FT Unhedged podcast

Can’t get enough of Unhedged? Listen to our new podcast, hosted by Ethan Wu and Katie Martin, for a 15-minute dive into the latest markets news and financial headlines, twice a week. Catch up on past editions of the newsletter here.