The Signs Were All There | Mike Green on When Passive Flows Meet the Largest IPO in History

Even the discussion around Tesla's inclusion in September of 2020 paled in comparison to the discussion that's going on now. This is recognized as something different, and we've dropped any pretense that they are quote-unquote passive indices. My analysis is it is mostly the passive flow. As I've shown elsewhere, that number is now accumulating to impact the largest US stocks by about 18% a year. Over 50% of Google's profits in this last quarter alone were tied directly to the price appreciation of its investment in Anthropic. People are putting a multiple on those gains. I mean, these are one-time gains. You either have the cash to make the payment on your debt or you don't. And that's the real question I think that's going to play through. We'll look back on it if it actually plays out and we'll say, 'Well, these signs were all there and nobody was paying attention.'

It's great to see you, Matt. I love the intros every single time. We were all products of the 60s and 70s. Mike was the most common name for an extended period.

It is. It is something to be proud of when your name is Mike Green and you can actually find yourself on Google. So, there you go.

This is the least interesting beat-to-death topic.

And you know, you think you learn. My work obviously suggests there's a mechanical component to much of this. The behavior of the indices actually would seem to validate that at this point. And so, you know, we're just kind of all waiting for what happens next because this has not happened before. We actually don't know how this will play out. We can build models of it and we understand the bid that is coming for SpaceX from the passive indices, the demand that that creates. The Heisman from the S&P 500 was actually a remarkable event. I candidly was a little bit surprised, as I described in my Substack. I was actually in the middle of an interview discussing SpaceX when the news came out and had no quicker or made the observation that I couldn't understand why S&P was doing it. I did understand why NASDAQ was doing it. I couldn't understand why S&P was doing it when the reporter I was speaking to said, 'Well, guess what? S&P just announced they're not going to include it.' And I think this is actually a really critical thing to understand is the role that these index committees and the index construction committees actually play from a fiduciary standpoint. That's become the most interesting thing in markets right now is this basic discussion around what sort of role should we expect passive to play? How should we expect it to impact it? It's a radical change from the discussion I was having a decade ago or even five years ago, which is people refusing to admit that indices had any impact whatsoever. So, you know, we've come to an interesting conclusion that feels a little bit like a logical endpoint to some of it, but again, we just don't know. We haven't seen it before.

I completely agree with you. Even the discussion around Tesla's inclusion in September of 2020 paled in comparison to the discussion that's going on now. This is recognized as something different. And we've dropped any pretense that they are quote-unquote passive indices. It's become very clear that they play a very active role in the behavior of capital markets.

Well, again, you know, because this has not happened, it is difficult to figure out exactly what the pattern is going to be. The indications that we have is the low float that is being offered combined with the buying from the NASDAQ should result in something that looks something like two times the underlying float in terms of demand that's created, expectations that the IPO will do relatively well. The impact that it will have on remaining securities in the NASDAQ 100 and the Russell 1000, etc., is certainly going to become the primary focus. One of the factors that is actually playing through right now, I would argue, is the locking up of capital and margin associated with the IPO itself is probably part of the cause that we're actually seeing for some of the recent market weakness. The concerns around that, the actions by the S&P. You know, I correctly articulated that this was actually a pretty big deal, that this suggests that financing may be in question for some of these AI companies, but you know, we can't know that fact yet is kind of the easiest way to put it.

Well, I think it has impact across both existing publicly traded companies where companies like Google have recognized that this may be one of the best opportunities to equitize some of the debt obligations that they've entered into. That in turn places pressure on additional supply from existing publicly traded companies. Meta is now exploring an offer, for example. It's hard to imagine that these can't get done, but all of it plays through in an increase in supply. And that really was the focus of the piece that I wrote this weekend, which is for the first time, we're actually seeing a meaningful shift in supply that is quite different than what we've seen up to this point. By and large, companies have been repurchasing shares and shrinking the quantity of equity that's available even as the demand from 401ks and passive entities have shifted outward the aggregate demand, driving prices and valuations much higher. It's natural to conclude that if we see anywhere close to the quantity of supply that is proposed with the Anthropic IPO, the Google issue, the Meta issue, how far behind can an Amazon issue or even a Microsoft issue potentially be behind this? We simply don't know again. And so this is one of those points where we've come to a fork in a road. And so the answer is always take it.

We're not sure which one that actually is though, and that's the problem, right? We can look for signs of it. It does feel like large-cap IPOs of this magnitude is a road less traveled, but again, this is just one of these situations that, you know, it is a predictable outcome, it is the anticipated source of supply. Marco Sammon, I highlighted his work, 'Who Clears the Market When Passive Trades', has highlighted that the vast majority of supply for passive demand is coming from companies themselves, and now we're just seeing it on a scale we haven't yet seen.

Yeah, so this is actually some of the more recent stuff that I've done and it ironically has been facilitated by the ability to use LLMs to rapidly build models and test some of this stuff. The really critical insight of passive is to think of it like a fire hose. It's concentrating a flow of capital into a crowded theater. The behavior of the participants in that crowded theater are kind of what you need to be thinking about. And what we've done is articulate using market impact work that was created by Jay Busch all the way back in 2014, was the subject of a very interesting paper that he wrote in, I believe it was 2024, originally called 'Ponzi Funds'. It was highlighting the impact of flows of capital that becomes a self-induced momentum when the capital flows are large enough that the funds themselves become a significant source of the liquidity. He focused on the ARK Funds example and how that created its own endogenous liquidity by creating capital that then had to buy the securities that underpinned her holdings and driving those securities higher. The same phenomenon plays through in the S&P once you aggregate together all of the S&P funds. They've now become a very large fraction of the total holdings. NASDAQ 100 is obviously an extension of that, slightly more concentrated, more tech-oriented, and as we've discovered, more prone to change itself to facilitate new IPO listings. But what we're really trying to do with that paper is identify how we should expect passive to impact markets. And so, if you think of that fire hose, it's basically sending the largest flow of liquidity towards the largest stocks. It has the largest impact on the largest stocks that have the highest volatility, Perold's framework. We see the evidence of that in this chart where historically the large-cap high stocks have been significant underperformers. This is part of the work that Cliff Asness at AQR correctly identified in the betting against beta framework. Basically, individuals overpay for access to highly volatile securities because it's an embedded form of leverage that they typically can't obtain in their accounts. That means those securities have a tendency to become overvalued and therefore their forward expected return on a discounted cash flow framework is expected to be lower. In a passive world, that's reversed because they've become more richly valued because they are more volatile. The impact of those flows that are allocated simply on the basis of market capitalization has a larger impact on those securities, and if that flow becomes the underlying dominant flow as it has, we would expect those securities to suddenly start outperforming the remaining portions of the index, and that's exactly what that chart is illustrating.

The objective of using a rolling 10-year is honestly somewhat arbitrary. We're simply trying to capture a long enough period that it typically will capture a business cycle. That's the objective of using something like a rolling 10-year framework. If we were to use a rolling five or a rolling 12-year framework, it would result in very similar outcomes. We actually tested for that to see if the choice of period really mattered and it ultimately didn't.

My analysis is it is mostly the passive flow. As I've shown elsewhere, that number is now accumulating to impact the largest US stocks by about 18% a year. That's more than many of them have actually been appreciating for the past several years and suggests actually that it is really passive driving this phenomenon. I'd also note that the vast majority of these companies that now claim AI provenance really didn't have any AI components to them a couple of years ago. And so like, you know, are memory stocks all about AI? No, they're really about memory and memory is used by the large-scale LLMs. But interestingly enough, we are seeing the forces of capitalism at work, the forces of innovation as the price of memory has begun to rise. We're seeing more and more LLMs that are capable of operating with less memory. They're becoming more efficient in their usage of memory. This is totally predictable. It was exactly like the wavelength division multiplexing and the amplification that occurred alongside the fiber optic buildout that allowed us to radically increase the quantity of capacity that we were able to obtain in terms of data transmission relative to the same amount of capex. I think this is going to be no different. And so, you know, is it AI? I think that's a convenient narrative. I do think that the earnings growth that people are pointing to is becoming increasingly circular. When Nvidia uses its stock to create an investment in CoreWeave, CoreWeave then has to use that investment to purchase Nvidia GPUs. That creates a circular financing component to it that is unfortunately identical to what we saw with the vendor financing that existed in the dot-com and the fiber optic buildout and candidly has been a feature of every capital spending bubble in basically recorded history. My analysis would suggest that Nvidia's profits are radically overstated on this. Over 50% of Google's profits in this last quarter alone were tied directly to the price appreciation of its investment in Anthropic. People are putting a multiple on those gains. I mean, these are one-time gains. Maybe they will continue forever because it certainly does feel like this is never going to stop. But to put a multiple on one-time gains and call that quote-unquote earnings is, you know, a very surprising outcome to me for most market participants.

Was it 12% of earnings growth or 12% of earnings? I think actually it's 12% of earnings. Yeah. Yeah. Exactly. Right.

Basically all the earnings growth, and you know, many other people have pointed this out, that, you know, the ex-Mag 7 or ex-AI earnings growth in the S&P is basically non-existent.

We actually have seen some shrinkage of the 401k flow, but we've actually seen the exact opposite in terms of the aggregate flows. So, as performance for the S&P has gotten better and better, again, in my analysis, supported by the growth of passive, as well as the demographic characteristics of boomers now having to effectively forward financial assets rather than try to replace income, as I've discussed in several recent pieces, you know, that has increasingly morphed into what feels like a discretionary buy-every-dip type phenomenon. And we absolutely saw that play through in the April to give or take end of May flows in which flows into ETFs actually exploded to the highest levels we've ever seen. Do I think that was tied to retirement flows? Not really. So the evidence that we have is there's some weakening on the 401k component. But as you would expect as you get towards the end of any bubble, you're seeing an increase in discretionary participation as more and more people basically just say, well, this only goes up. It's a license to print money. And I think that was a bigger impact combined with the systematic flows, the reversal of short positions, the CTA trend-following strategies which continue to attract additional capital, as well as things like volatility control funds stepping back in for the very simple reason that we didn't actually realize that much volatility. It's one of the things we actually talked about at the firm that I'm an advisor on, Tier One. As we looked at the market in April with the extraordinary skew and the bid for protection that was in place, it was behaving like a market that had already crashed. And we're starting to see signs of that again that people are starting to bid up fixed strike volatility. The VIX is in the mid-20s. Implied correlation is in the absolute toilet as more concern is on single-stock components of it. These are all hallmarks of a market that is uncertain about what happens next, is somewhat desperately bidding for protection. If it plays out as it did in April, although I think this could, you know, the increase in supply could be more material this time. If it plays out like it did in April, the odds are that we get another drawdown and then we get pushed to even new, you know, even higher highs in the absence of anything material developing. The flip side of that is as we look at the economy, the increase in gasoline prices in particular is hitting the lower end of that K consumer and we're seeing sentiment deteriorate rapidly. We're seeing savings being drawn down. And you know, my personal favorite was a piece that came out from the New York Fed that was highlighting the link between inflation or loss of purchasing power and consumer sentiment and came to the absolutely shocking conclusion that households that had found new jobs that paid more money and allowed them to maintain their purchasing power without additional effort weren't reporting the same degree of depressed consumer sentiment as households that had to obtain additional working hours or get a second job. Like this was to me just an astonishing statement that came out of the New York Fed that they were surprised that people who are forced to work harder and work more jobs in order to maintain their purchasing power are more depressed than people who have found new jobs that allow them to do the same with less effort. It's shocking the insights that we develop at the highest levels of economic research.

Okay.