Berkshire or the S&P?

Good morning. Unhedged is back at full force. Nvidia took up all the oxygen last week, but don’t sleep on Eli Lilly. Strong earnings fuelled by weight-loss drugs have pushed the stock up 32 per cent this year. Citi expects Nvidia and Eli Lilly alone will make up a quarter of S&P 500 earnings growth in 2024. How’s everyone feeling about this rally? Email us: robert.armstrong@ft.com and ethan.wu@ft.com. 

Berkshire and the S&P 500

If your entire equity portfolio had to be in either the S&P 500 or Berkshire Hathaway, which would be the better choice?

With my colleagues Oliver Ralph and Eric Platt, I put this question to Warren Buffett, Berkshire’s chair, in an interview five years ago. His response:

I think the financial result would be very close to the same . . . if you want to join something that may have a tiny expectation of better [performance] than the S&P, I think we may be about the safest.

In his latest annual letter to shareholders, which appeared over the weekend, Buffett says something similar but not identical:

[We] have no possibility of eye-popping performance . . . Berkshire should do a bit better than the average American corporation and, more important, should also operate with materially less risk of permanent loss of capital.

One reason for the modest outlook, Buffett writes, is that his company is so big now — 6 per cent of the S&P 500’s total shareholders’ equity — that there are no companies in the US or abroad that are (a) big enough to make a meaningful impact on Berkshire’s earnings, (b) possible to reliably value, (c) have trustworthy and competent management, and (d) are available at a reasonable price. There are no big moves left to make. A similar story applies to purchasing stakes of public companies. Berkshire is a large conglomerate of high-quality businesses that will just churn along. 

But shouldn’t the quality make a difference? That is, shouldn’t the safety that Buffett referred to both in the recent letter and in the 2019 interview give the company an edge? If an asset has less risk for the same level of return, you can just leverage it up so it has the same risk and a higher return. You can own Berkshire with some borrowed money. 

But notice that in the more recent statement, Berkshire is making a comparison to “the average American corporation,” not the S&P 500. The difference is important because an average corporation might go bust; the S&P 500 never will. Some members of the index will peter out, but others will soar. Its diversification makes it intrinsically safer than its average member from the point of view of capital preservation. Buffet’s more recent claim is even more modest than the older one.  

I don’t want to read too closely into the wording of either statement, but I do think we should at least take Buffet at his word: there really is no reason to expect meaningful outperformance from Berkshire over the long term. It’s too big. Its ability to provide expensive capital to stressed businesses in moments of crisis must be balanced against the drag on returns from its low leverage and high cash holdings in good times. Buffett is not under-promising so he can over-deliver. He’s just being honest. 

Berkshire bulls might object that the company has delivered slightly better performance than the S&P since 2019: 15.7 per cent annually to the S&P’s 14.1. That 1.6 per cent difference, compounded over time, could add up to something meaningful. But remember that in the decade before 2019, Berkshire underperformed by about the same small but meaningful margin. It probably just doesn’t matter, from the point of view of long-term returns, whether you own Berkshire or the S&P.

Why, then, should Berkshire exist? When a virtual mega-conglomerate can be owned for a few basis points, what is the point of an actual mega-conglomerate? Here we move into non-economic territory. Institutions matter: they can create trust, bind people together, and pass along wisdom and values. Berkshire is such an institution in American life, and my guess is that it does a lot of non-economic good. Back in 2019, my colleague Oliver asked Buffett if he should put Berkshire or an S&P tracker fund into his young son’s college fund. Buffett replied that “I think your son will learn more by being a shareholder of Berkshire.” There is something to that, even if it cannot be translated directly into wealth.

Inflation and the money supply

Most economists missed the 2021 inflation spike. Even those who were right that the pandemic economic cycle was different, such as Jan Hatzius of Goldman Sachs, were too sanguine about inflation early on. Equally, the few who correctly predicted inflation’s rise, such as Larry Summers, were too pessimistic about how entrenched it was. Inflation is poorly understood, and devilish to forecast.

That said, look at the chart below. It shows a measure of year-over-year money supply growth against the Fed’s preferred measure of core inflation. If you had simply followed what the rapidly expanding money supply was telling you, you’d have seen inflation’s rise perhaps 10 months ahead of time. And you’d have been worrying about inflation a year and a half before Jay Powell said it was “probably a good time to retire” the “transitory” label:

Monetarism, the idea that changes in the quantity of money drive inflation, has fallen out of favour in economics. Yet the eerie accuracy of the money supply in predicting the 2021 inflation wave spurred interest in considering money again. As Martin Wolf wrote in 2022: “Just as the financial crisis showed that banking matters, so this inflationary upsurge shows that money matters . . . we cannot steer the economy via the money supply [but] we cannot ignore it either.”

Was this conclusion premature? The money supply did not expand in isolation. It coincided with a mammoth fiscal stimulus effort and a global supply chain logjam. These alone might plausibly explain why inflation shot up and then collapsed. That the money supply mattered is not self-evident.

In a new paper, two Dallas Fed economists, Tyler Atkinson and Ron Mau, downplay the usefulness of money. Their premise is simple: if money matters, then measures of its growth should be able to forecast inflation better than a rudimentary forecast based on lagged inflation from 12 months before. If money-based forecasts can’t outperform crude extrapolation from inflation’s recent past, what use are they? 

As it turns out, money-based forecasts fail to outperform lagged inflation. Since 1969, they have done roughly 11 per cent worse in predicting core PCE inflation over the next 12 months. Money-based forecasts occasionally do much worse, too. As the chart below shows, in 2011 the money supply was signalling deflation, while a lagged-inflation-only forecast accurately stayed close to 2 per cent:

(One data note: readers are probably most familiar with M2, the classic money supply measure. The problem with M2 is that it adds together very different types of “money”; interest-bearing, locked-up CDs are treated the same as a checking account. The Dallas Fed authors instead use the Divisia measures, which weight payment types based on how cash-like they are, and therefore how much they contribute to economy-wide liquidity.)

Atkinson and Mau do concede that money-based forecasts slightly outperform lagged inflation if you look solely at post-2020 data. Why might that be? They offer no explanation, but former New York Fed chief Bill Dudley has. In his interview with Unhedged last year, Dudley told us:

M2 is going to be correlated with the shift from QE to QT. But if you go look at M2 growth after the GFC, you saw a lot of QE, you saw rapid growth of M2. And there was no inflation, no consequence for growth. M2 just doesn’t have much relationship to economic activity.

People just don’t understand how the Fed’s operating model has changed. Quantities of money don’t really matter very much. What really matters is the interest rate that the Fed sets on reserves.

Put another way, in a world where liquidity is everywhere, changes in the money supply don’t signal much about the availability of credit. Whatever link might’ve existed in a pre-QE world is no longer there. So the best explanation of the post-2020 correlation between the money supply and inflation is that it’s largely coincidental.

It’s possible that a sophisticated model might extract signal from the money supply data. We’re not saying it doesn’t matter at all. But for investors interested in a quick temperature check of the economy, money may not have much to offer. (Ethan Wu)

One good read

McKinsey proves, once again, that it is not good at managing geopolitical conflicts of interest.

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.