…weird.

Mr. Biden, who has done a lot of good things while in office–including overcoming with actual medical and financial help Mr. Trump’s deer-in-the-headlights denial that the pandemic was happening, while the unburied dead stacked up in meat lockers–presented himself pretty much as the investigator of his taking secret documents home described him. That is, an elderly gentleman with a so-so memory.

In contrast, Mr. Trump, a fabulous reality star in his role as a real estate magnate, but whose major accomplishment during the epic boom in New York real estate in the 1980s and 1990s was to not lose all the money he inherited, appeared vigorous. And, as the old joke goes, I didn’t want to believe he was lying but his lips were moving–throughout the debate.

I thought it was an interesting case of the self-reference problem when Mr. Trump talked (again) about his “IQ test,” which he proclaimed he had passed with flying colors. It appears to have been, however, a test to detect dementia. So he is essentially trumpeting the idea that his doctors think he’s suffering from cognitive impairment, but not badly enough for him to flunk a dementia test. Another twist: how do we know he’s not lying about this as well.

But crypto hasn’t gone through the roof. And the stock market is up as I’m writing this.

Back in the infancy of the semiconductor industry close to a half-century ago, a fundamental split occurred between makers of analog chips using artisanal methods and makers of memory that could be mass-manufactured.

The former were blown away in the 1990s by the ARM-TSMC digital axis. The latter were originally Samsung, a number of Japanese companies, MU and Intel (INTC).

INTC quickly realized that the memory market was going to develop into a bruising, capital intensive, contest among a number of behemoths making what are more or less commodity products. So it shifted to making microprocessors, where it could (and did, until management lost its way in dreams of past glory) dominate the field.

That left MU, Samsung and the keiretsu to slug it out in a market that, to my mind, looks a lot like the one for construction or farm equipment, or washing machines. That is to say, a market strongly influenced by the business cycle. Yes, the secular growth prospects for memory chips are much better than for, say, tractors, as formerly dumb things like cars or refrigerators get computerized. But the competitive market structure means a cap on unit profits and the products’ use in expensive equipment both mean all the players are vulnerable to the ups and downs of the business cycle.

The business cycle isn’t an issue today, as far as I can see. And the memory industry in general is also getting a significant boost from the need for users to upgrade their computer equipment to make the best use of AI tools. But it seems to me that it’s a mistake to regard MU as an AI stock …AI-adjacent, maybe, but not AI in the way that NVDA or even TSM is.

If this is correct, the most important issues for holders (I’ve been one for about a year and a half) are, I think, what peak earnings for this business cycle will be and the multiple we should put on those profits.

According to Fidelity, MU is now trading at about 10x year-ahead earnings. If that’s the peak, then the stock seems to me to be fairly valued (if this were construction/farming equipment the peak multiple would be 5x or so, I think).

My guess is that those analyst estimates will end up being too low, not because I think the analysts are being conservative (they’re usually the opposite, in my experience) but because the memory industry tends to be boom or bust–and therefore has earnings that are hard to estimate with any precision. So I’m content to hold for now–but with the thought firmly in mind that the stock will likely peak (assuming, as I do, it hasn’t already) long before the earnings do.

So I find myself in the awkward position of being both in no rush to buy more, in no rush to sell and scratching my head at the apparent rationale behind today’s selloff.

A second question is whether this is a MU issue or one with broader implications for the AI space.

I own a small WMT position.

If you’d asked me even a year ago, I would have said WMT is a mature company whose management no longer had the entrepreneurial spirit of the company’s founders and whose latter-generation family owners were content to clip coupons rather than take the risks inherent in seeking superior earnings growth.

But then I noticed that, startling to me, I could find camera equipment online from Walmart that was cheaper than anyplace else. Then I read that the company was acquiring Visio, to create, I think, its own competitor to Roku. I also noticed that as the economy improved, customers who had migrated down market from Target (I own TGT, as well) to WMT were not moving back as fast as they usually do. So I decided to buy the stock.

I read today that WMT was successful in separating itself from credit card provider Capital One. This allows it to issue a credit card through Discover, which WMT bought in February.

Who knows what will happen from here. But the credit card move is more evidence that the Walton family no longer content to go the way of, say, the domestic auto companies or IBM–bureaucrats who look good in a suit rather than entrepreneurs.

This is a continuation of my post from yesterday, in which I argued that making a large bet in an equity portfolio on the direction of interest rates is particularly unappealing in today’s market. It’s not so much that Wall Street strategists have been consistently wrong on this front. It’s partly, but not totally, that this tends to be a grand bet-the-farm move that would send alarm bells ringing extra-loudly for any seasoned professional. It’s that, in addition, the waters are, to me at least, so murky.

If one wanted to make this kind of bet through equities, the way to do so would be with financials and nosebleed-multiple stocks, both areas that are extremely sensitive to rate moves. Overweighting either would be an implicit bet on rates falling, underweighting the opposite.

It seems to me that a better strategy for today, and the one I’m following, is to assume that rates will go sideways for a considerable period. (For what it’s worth, the only way I can see rates going up is through capital flight triggered by a Trump victory in November.)

If so, portfolio management becomes a game of analyzing individual stocks and finding ones that are attractive either on the conviction that they’ll have surprisingly good earnings or that they’ve been beaten down enough that a change in direction–or any other signs of life–will cause the stock to bounce (think: HOOD).

If we look in the simplest possible way, there’s a relatively strong relationship between the interest yield on Treasuries and the earning yield (1/PE) on stocks. That’s because stocks and bonds are the two main kinds of liquid long-term investment, and both react in varying degrees to rate shifts.

Remaining simple, that relationship is probably best described as: interest yield = earnings yield. So, taking the 10-year T note as a proxy for bonds in general: if the interest yield is 4.27%, the PE that puts the S&P 500 in equilibrium with Treasuries should be 1/.0427x , or 23.4x current earnings. If we take the 30-year at 4.4%, then the equivalent PE is 22.7x.

Based on consensus estimates, the PE on coming 12 months’ earnings is about 22x. If that’s correct, and based on my simple model, stocks and bonds are roughly in equilibrium today.

If the 10-year yield were to drop to (I’m just making up a number) 3.5% over the next year, then the equivalent equilibrium PE for stocks would be 28.5X current earnings, or a level about 25% higher than where they are now.

How likely is that?

The Fed’s announced target for inflation is 2% (I believe a more realistic one is 2.5%, or maybe a tad higher). Add to that a real yield of 1.5% (meaning a return in excess of compensation for inflation), or about what the real yield was during the ten years before the pandemic, and the Fed’s nominal target is 3.5%. Getting to 3.5% is certainly a possibility. But the Fed has been trying its best to land there for about two years, without total success. The domestic economy has been too strong.

This continues to be the situation, despite Wall Street’s most heralded strategists continuously predicting that a sagging GDP, triggering a sharp drop in rates, was just around the corner.

In my view, there’s no reason to believe that Wall Street’s forecasting ability is going to miraculously improve. I think it’s highly unlikely that rates are headed up. More than that, I have no clue. My main takeaway is–great if rates start to decline, but this can’t be the heart of my equity investment strategy. Instead, it’s got to be finding hotspots of growth …that are mostly bets on individual company earnings coming through in a way that the consensus has underestimated.

more tomorrow