Passive investors and the AI bubble

Good morning. Nvidia shares rose 2.5 per cent yesterday, bringing the shares almost all the way back to their all-time high of a few weeks ago. How much of the shares’ incredible resilience is down to passive investors, who continue to pile into US stocks without regard to price? My view below. I’d like to hear yours: robert.armstrong@ft.com. 

Passive investing, market efficiency and valuations

If you want to get a feel for how complicated markets are — how interconnected, how full of feedback loops — think about passive investing. Has the rise of passive funds made markets less efficient? Has it encouraged bubbles, most recently the AI bubble?

Try, in your head, to work out the answers to these questions. Even if you know quite a bit about markets, you may soon feel a bit uncertain. Like me, you may experience an urge to draw a diagram or deploy a spreadsheet, which may not be a lot of help. Reading some of the many papers on the topic may be of some use, but if you are hoping for professional or academic consensus on the topic, you will be disappointed.

Here’s a toy example I use to think about it:

1. Imagine, with an economist’s reckless love of abstraction, that all investor equity allocations were managed by four active funds of roughly equal size, A, B, C and D (the funds hold some cash, too, to facilitate trading and adjust risk levels).

2. One day, C and D announce they will become passive funds, holding all stocks at market weight. Half of the assets in the market are now passively managed. In other words, what has happened to the US stock market over decades has happened in our imagined market overnight.

3. There will now be a messy period, where C and D adjust their holdings to match the market (which includes themselves). This will take some back and forth, but eventually there will be a new equilibrium. 

4. Reaching equilibrium will have an effect on the prices of individual stocks, because C and D will have previously been overweight or underweight certain stocks, and they have to sell/buy those stocks to fix this. How much this adjustment moves prices will be determined by the elasticity of A and B’s demand for the stocks in question.

5. Overall valuation of the market doesn’t change in this process, though. Investor demand for equities has not changed as a result of C and D’s move to passive, and for every share that the passives have to buy to get to market weight (driving prices up) there is another they will have to sell (driving prices down). 

6. Once in equilibrium (and ignoring inflows of new money) market prices will move on disagreements about value between A and B — when A wants to own more of a stock and B will only sell it at more than the current price, for example, or alternatively A wants to own less of a stock and B will only buy it at a lower price. But to match these price changes, the passive funds don’t have to buy or sell anything. The price changes take care of the passives’ market weightings automatically.

7. New asset flows into the passive funds go into all stocks in proportion to the stocks’ size, so the increase in demand is the same for all of them, even though more money goes into the bigger stocks. There is a general price impact on all stocks from higher demand, but no relative price impact between stocks. 

8. There are now fewer people for the active funds to trade with. In fact, A and B have only each other to trade with, except when inflows or outflows force the passive funds to trade. Does this make the market less efficient? In this extreme example, of course it does. If A will only sell a certain stock at a price above what B is willing to pay, they will not transact, and the price will not move towards whichever side is more correct about the stock’s true value. The market has become more rigid, and probably more volatile in the face of inflows and outflows, due to the growth of passive investing. But of course this is a very extreme case. What we don’t know is what number and proportion of passive participants calcifies a given market, and there is no evidence that I know of that we have hit that level in the US yet. 

9. Even in our extreme example, however, there are other things that could happen to bring prices back towards fundamental value. An undervalued company might receive a tender offer for all of its shares, or might start buying its own shares back, driving up the price. Or a company’s earnings could rise or fall so much that A or B change their minds.

10. Is it important whether the fund managers who go passive are smart or stupid? Let’s say A, C and D were run by smart people, and B by stupid people. C and D switch to passive. Now A has only the dummies at B, who misprice things a lot, to trade with. A should make more money — but only on the assumption that B eventually recognises that it has been stupid, trades with A again, and the mis-pricings close. If the rise of passive makes the market collectively stupider, the immediate result ought to be a few people making it smart again, and making a lot of money doing so.  

This toy example makes me think that the rise of passive investing has not made markets much more inefficient, and has not encouraged the AI bubble. But I say this with lowish confidence.

I should also note, however, that the rise of passive investing could have surprising indirect effects on the collective behaviour of investors. In a recent piece, Ben Inker and John Pease of GMO point out an interesting possibility (which relates to step 4, above). Active fund investors tend to be inclined towards value stocks over growth stocks, Inker and Pease argue — which makes some sense, given that many stock pickers have been trained in fundamental analysis. Active managers may also be inclined to hunt for mispriced companies among small-caps and other stocks not in the biggest indices, whereas passive money goes for size, liquidity and the accompanying low fees. So when assets move en masse from active managers to passive ones, there may be a negative demand shock for value stocks and small stocks — and a positive demand shock for the Nvidia’s of the world.

Another area that has not been considered here is the impact of passive investing on corporate governance and what the late Paul Myners called “ownerless corporations”, and what second-order impact that might have on market efficiency. But my best guess is that the first-order impacts are pretty small.

One good read

Managing the president.

FT Unhedged podcast

Can’t get enough of Unhedged? Listen to our new podcast, 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.