Cool inflation, giddy markets | Financial Times

Good morning. The markets loved Thursday’s inflation report, and we did too. But we think the inflation fight is far from over. The fight over FTX is going to run for a while, too. Our thoughts on both below. Email us: robert.armstrong@ft.com & ethan.wu@ft.com. 

Inflation, cooler, but mostly the same

The stock market doesn’t rise 6 per cent in a day for no reason. Yesterday delivered one: decisively cooler inflation numbers that inspired hope the US Federal Reserve might back off and finally let a market bottom form.

The CPI report is indeed confirmation that the disinflationary trends visible everywhere from housing to job cuts is real. As a result, markets have lopped off a lot of inflation upside risk. Futures traders’ implied expectations for where rates will be in late 2023 fell hard, while expectations for the next few months dipped only modestly. The market thinks higher-for-longer rates are less likely today than they did on Monday.

What did the CPI numbers actually say? Last month, we offered four predictions, based on conversations with inflation specialists, for the October inflation data that have just come out (recall that the key services inflation categories are, from most to least important, shelter, medical care and transportation):

Core goods (ex food and energy) would start deflating.

Transportation services would stay volatile, but begin to gradually moderate as airfares settled down.

An annual update to the health insurance data used to price medical services would flip it to an inflation drag.

Shelter would stay pretty hot.

Yesterday’s inflation numbers delivered on all four counts. The chart below sums up. The top four rows are for October, the bottom four for September:

Core goods and medical services prices are falling, while transportation prices are growing a bit less quickly.

Shelter was a bit trickier this month. The high-level point is that it decelerated slightly, but there is a complication. As you can see in the chart above, the shelter category accelerated in October. But under the hood, the two most influential subcategories — rent (down from 0.8 per cent to 0.7 per cent) and owners’ equivalent rent (0.8 per cent to 0.6 per cent) — eased. The culprit was explosive acceleration in one little-watched subcategory: hotels (up 5.6 per cent in one month!). 

This looks like noise, though. As Omair Sharif of Inflation Insights explained to us, the hotels component doesn’t tell us much about shelter:

What really is in CPI is hotels near vacation spots, ski towns, beaches, things like that. Not the hotel down the road from your parents’ house. If there’s volatility around resort pricing, it’ll show up in hotels.

And with the overall hotel index right on its pre-coronavirus pandemic trend, we aren’t terribly worried.

Stepping back, yesterday’s report has not changed the picture much. Inflation has peaked; it is coming down; how fast it comes down hinges on shelter. And shelter prices are likely to come down slowly. As we’ve noted previously, there is a catch-up effect in rent levels that may well lift inflation for several months to come.

We’re glad inflation is not running wild, but the central hypothesis remains the same: inflation prompts the Fed to keep rates restrictively high for a while, causing a recession. What happens on the margin — inflation falling a bit more briskly, rates rising more gingerly — still matters. But the broader picture — hot inflation, punishing rates — matters more. (Ethan Wu)

FTX, again

From a Reuters story, published yesterday, about FTX and its chief executive Sam Bankman-Fried:

This May and June, Bankman-Fried’s trading firm, Alameda Research, suffered a series of losses from deals, according to three people familiar with its operations . . . Seeking to prop up Alameda, which held almost $15 billion in assets, Bankman-Fried transferred at least $4 billion in FTX funds, secured by assets including FTT and shares in trading platform Robinhood Markets Inc, the people said . . . A portion of these FTX funds were customer deposits, two of the people said

Here is the Wall Street Journal, also yesterday, on the same topic:

Crypto exchange FTX lent billions of dollars worth of customer assets to fund risky bets by its affiliated trading firm, Alameda Research . . . Chief Executive Sam Bankman-Fried said in investor meetings this week that Alameda owes FTX about $10 billion, people familiar with the matter said. FTX extended loans to Alameda using money that customers had deposited on the exchange for trading purposes . . . All in all, FTX had $16 billion in customer assets, the people said, so FTX lent more than half of its customer funds to its sister company Alameda.

I don’t know if it is legal for a person to transfer customer funds from their crypto exchange to their hedge fund, because I don’t know what a crypto exchange is, according to the laws of the Bahamas, the Cayman Islands, the Forest Moon of Endor or wherever the entities involved are domiciled.

Undoubtedly, though, such a transfer is wildly risky. It takes assets out of an institution that has on-demand liabilities and replaces them with an IOU. This raises the risk of a bank run. What is more, as the Reuters story has it, that IOU was largely collateralised by FTT, a token that the lender invented. That is insane in so many ways I don’t have space to lay them out here. Finally and obviously, such a transfer implies icy disregard for the exchange’s customers.

Bad stuff. But it does help to explain what is going on. FTX can’t meet customer withdrawals because it gave customer assets to Alameda and can’t get them back, at least not in a liquid enough form to be useful.

But this simple story is muddled by another account released yesterday. It came by Bankman-Fried himself, using Twitter. It contained two material claims. The first:

FTX International currently has a total market value of assets/collateral higher than client deposits (moves with prices!). But that’s different from liquidity for delivery — as you can tell from the state of withdrawals.

In other words FTX is solvent (in the limited sense of having enough assets to make its customers good eventually) but not liquid (in the sense of having enough easy-to-sell assets to make customers good soon). It just needs enough time to sell some assets that are hard to sell. And what mistake led to this liquidity problem? Not a huge transfer of assets to a hedge fund. Instead, FTX’s customers were much more levered than Bankman-Fried knew. That’s the second material claim:

. . . a poor internal labelling of bank-related accounts meant that I was substantially off on my sense of users’ margin. I thought it was way lower.

Users’ leverage was 1.7, not 0 as he had thought, Bankman-Fried said (this must be on some sort of net basis, because obviously lots of FTX clients had leverage, leverage is something FTX sells). So for every $1 in customer deposits, customers borrowed aggregate $1.70 from FTX. And the result was that there wasn’t enough liquidity left to meet the rush of withdrawals late last week.

But why would customers having leverage — owing FTX money, in short — cause a liquidity problem in the face of mass withdrawals? The only way I can think of is if the loans the exchange provided to customers were funded by loans from someone else, and that someone else wanted their money back at the same time customers did.

This could happen, but if it is what happened, Bankman-Fried’s mistake looks much dumber. He is suggesting he owed multibillion-dollar callable loans to a third party, and he didn’t know, even approximately, how big they were.

We still have more questions than answers. To wit:

Can the news reports and the Bankman-Fried Twitter thread both be true? Yes, in a narrow sense. It could be that FTX is solvent and illiquid, and that customer leverage made meeting withdrawals trickier than it otherwise would have been, and that FTX lent billions in customer assets to Alameda. It’s just that FTX is a lot less solvent and liquid than it would have been if it hadn’t made that loan. Of course, if Bankman-Fried wrote that long thread about being more transparent and how sorry he is, and just decided not to mention that he also lent billions of dollars of client assets to his hedge fund in return for the worst imaginable collateral, then he is a jerk of world-historical proportions. But that is a different issue.

Isn’t it odd that FTX says it is solvent (in the limited sense described above) but needs a whopping $8bn dollars of liquidity? You bet your ass it’s odd (and FTX may be bankrupt or bankrupt-ish already, according to Bloomberg). 

If FTX is solvent, why can’t it get someone to lend it the liquidity it needs? Maybe it can, and it will just take time. Perhaps the problem is that even though loaning FTX the money would be a smart financial move, no one with $8bn wants the reputational risk of dealing with Bankman-Fried, who, as noted, might be a major jerk. Also, lenders might not believe FTX’s books are accurate, given the history of, erm, poor internal labelling of bank-related accounts.

Can Bankman-Fried’s suggestion that higher-than-thought customer leverage led to the liquidity crunch withstand scrutiny? Yes, but a much simpler explanation is available: that FTX didn’t have enough liquidity because it made a dumb and grossly immoral loan to Alameda.

There is a lot we still don’t know.

One good read

Murdoch vs Trump.