Good morning. I’m not sure what to make of the negative first-quarter GDP number, which was a disappointment, but included a huge shift in the trade balance, which may be noise. But we are seeing slower growth elsewhere, so in any case the report was not wildly anomalous. The employment cost index, out this morning, will matter much more to markets. Meanwhile, three tricky topics to end the week. Email me: robert.armstrong@ft.com

Are soft Amazon earnings an ill omen?

Amazon reported first-quarter earnings yesterday afternoon, and the shares fell as much as 10 per cent in after-market trading. Amazon Web Services did great, as it always does. There seems to be two other culprits: weak sales guidance for the second quarter, and slow growth in Amazon’s online retail business.

Historically, Amazon guidance has not told us much, so I think we can look past that — it could just be conservative. But here is growth in online retail sales:

Should this worry us, in the context of Amazon itself, or American consumption generally? In a call with journalists, the company’s CFO put this dreary pattern down to tough comparisons with the peak pandemic period, when online shopping boomed:

We don’t see the slowness in the growth rate in dollar basis from prior quarters. What we’re seeing . . . is that there’s a very high growth period from about May 2020 through May 2021, where our growth rates went from 20 per cent to over 40 per cent. And then the next calendar year, which we’re coming up on the anniversary of, creates a lot of difficult comps

I’m not sure whether I buy this or not. It makes sense that people might have bought a lot of stuff online a year ago that they are back to buying in stores now (or not buying at all). At the very least, though, that undercuts the notion that the pandemic permanently changed buying habits. Whether there is also an underlying growth deceleration, we just can’t know at this point. The pandemic makes all sorts of things hard to analyse.

Against using Tips yields as a proxy for real interest rates

It is very standard to treat the yield on inflation protected treasuries (Tips) as a proxy for the real rate of interest. It makes some sense to do so: the Tips yield is the risk-free after-inflation return a pretty large group of investors is willing to accept.

But it is also very standard to point out Tips yields have been negative since the start of the pandemic, and that the idea of a negative real rate is very strange. Is capital cheaper than free? Should people pay to be rid of the stuff? And so on. People who make these kinds of points tend to think that Federal Reserve buying of Tips has hopelessly distorted their prices.

Dhaval Joshi of BCA Research offers a different argument against Tips yields as a real yield proxy. It goes like this:

1. The Tips market is small relative to demand for inflation-protected, risk-free assets: $1.5tn, compared to a $25tn Treasury market.

2. When inflation surges, investors cram into Tips, distorting their yields. “The ultra-low real yield on inflation-protected bonds captures nothing more than the massive imbalance between huge demand for inflation hedges and tiny supply.” 

3. Here’s the proof of the distortion. The difference in yields between nominal Treasuries and Tips is often called “inflation expectations” or “inflation break-evens” because it seems to represent the yield investors are willing to give up to receive inflation protection. Note, though, that this difference neatly tracks the oil prices. Joshi’s chart:

4. But the idea that inflation expectations should track the oil price is absurd. The higher the oil price is now, the less it is likely to go up later, pushing inflation up. Indeed, empirical evidence shows that the oil price and subsequent inflation are inversely correlated. And the logic of this argument should hold for prices broadly.

5. So the Tips yield cannot be a good proxy for real yields.

I can’t see what is wrong with this argument, though the inflation expectations-oil price relationship looks a little messier if you go back before 2015, where Joshi’s chart begins.

This strikes me as a pretty important argument, if it’s true, given how much we lean on break-even inflation to make sense of markets.

Utilities, dividend stocks, and rates

It used to be that utilities were thought of as a sort of substitute for Treasuries. Utilities are very safe, given that there is always a certain level of demand for power and the companies’ returns are often government regulated. And they tend to pay a solid dividend yield. So, as a rule, utility stocks prices went down as Treasury yields went up — the logic being: why should I take even modest risk on a utility stock when I get a good riskless yield on a Treasury? The same logic ran the other way too.

This is not true lately, though (hat tip to Nicholas Bohnsack at Strategas whose work alerted me to this). Here’s a chart that shows the performance of the utility sector relative to the S&P 500, plotted against 10-year yields:

As you can see, when yield rise, utilities have generally underperformed (see especially 2012-14, 2016-2019, 2020-21). But now yields are shooting higher, and utilities are outperforming at the same time.

The difference is inflation, of course. If yields are going higher because inflation is rising, rather than because the economy is strong, it makes sense to sell bonds (which get killed by inflation) and buy utilities (which have a yield and can to a degree pass higher input costs on to consumers). Inflation upends previously reliable market correlations.

This got me wondering if a similar thing is happening to dividend stocks generally. Were they historically sensitive to Treasury yields, and is the relationship changing now? Well, here is the relative performance of the iShare Select Dividend ETF, compared with the 10-year yield:

I can’t see much of a relationship over much of the past decade. Dividend stocks underperformed pretty steadily, whatever yields were doing. Since the pandemic, though, there is a clear correlation. As inflation has risen, it makes sense that dividend stocks start to look like a good bond substitute.

Many expect inflation to subside from here, but to remain higher than pre-Covid levels. If that is right, might not dividend stock valuations rise?

One good read

When Joerg Wuttke talks about China, listen.