The Line We Can't Cross | Mike Green on the Passive Investing Endgame

Every single day, these index funds and ETFs receive inflows or outflows occasionally. That actually means that they are continually trading. They are continually executing in markets, and as a result, they don't at all fit the definition that Bill Sharpe provided in the arithmetic of active management.

That number of about 83% is actually a static number that if you get to that level of passive, effectively the market becomes so impossibly volatile that it becomes an inevitable event that it will eventually cause its closure. My estimate is actually we're about 54% passive by market share, and we picked up about 4% last year. So, you know, if that pace continues itself, we would be looking at somewhere in the neighborhood of five years out, the world comes to an end.

It feels like I was just here.

Oh, I don't agree with that at all. But not that Dave is not highly intelligent and doesn't offer highly intelligent questions, but you guys very thoroughly prepared. You sent me some really nice notes.

Well, it's interesting. So the inelastic market hypothesis refers to a paper by Xavier Gabaix and Ralph Koijen that came out in 2020. My work on passive actually preceded that. I started in 2016, and really I credit Lasse Heje Pedersen at AQR for writing the paper that ultimately, in my opinion, broke the mental logjam around how passive influences markets. In 2016, he wrote a paper called Sharpening the Arithmetic of Active Management, which refers back to Bill Sharpe's seminal paper in 1991, The Arithmetic of Active Management, which is the source of all the language that you hear that active and passive, at the end of the day, own the same things in the same proportions, and therefore the only difference between the two is ultimately going to be the fees that active managers charge, that help to explain or provide the fundamental explanation for the outperformance of passive strategies and encouraged people to deploy assets in that direction for the simple reason that they could become free riders onto the system.

Lasse Pedersen pointed out something very, very important, which is the definition that Bill Sharpe provided for passive versus active. His definition of passive management was somebody who never transacts. Never transacts. Can't get in, can't get out. And he did it as a thought experiment where he articulated that, well, let's presume that they trade in the final hours when the markets aren't really open or closed. That's obviously impossible. And Lasse was correct to highlight this. What he noted was that there are distinct events, index reconstitutions in particular he focused on, where the end portfolio of the passive investor has to change, and so they have to transact. And during that time period, they become active managers and are no longer subject to the rules that Bill Sharpe articulated.

That actually is absolutely correct. It's turned out to now become the largest hedge fund strategy is actually index arbitrage, basically taking advantage of passive strategies, trying to use statistical arbitrage to project or predict what is going to go into an index or leave an index and to trade those appropriately. That's why we have things like the run-up in Tesla in December of 2020 in anticipation of it joining the S&P 500. And as I said, it has become an incredibly large business for hedge funds. The second source of portfolio change, though, is actually caused by the end customer. Anytime you contribute cash to a portfolio, you're changing the composition of that portfolio and forcing it to transact to match the benchmark. That was really my contribution to the process and recognizing that every single day these index funds and ETFs receive inflows or outflows occasionally. That actually means that they are continually trading. They are continually executing in markets, and as a result, they don't at all fit the definition that Bill Sharpe provided in the arithmetic of active management.

Basically, none of his work has any theoretical framing in a world in which their portfolios are changing. That's what really set me off on that, because all of a sudden, the minute you realize that, you recognize what they actually are, which is just a systematic algorithm that operates off of the world's simplest algorithm: Did you give me cash? If so, then buy. Did you ask for cash? If so, then sell. And that becomes a much easier thing to understand in terms of how it's going to impact the markets. So that's where the whole thing starts. Now, by 2020, the academics began to wake up to components of this. Xavier Gabaix and Ralph Koijen wrote the inelastic market hypothesis that challenged another precept to the efficient market hypothesis, which all index funds are based on. In the efficient market hypothesis, it's presumed that markets are highly elastic, that they can basically absorb nearly any amount of supply and demand change with very limited impact on either individual securities or certainly the market in its totality.

The actual specification within the efficient market hypothesis is that a dollar into the market, because every buyer has a seller, really should only impact market capitalization by basically the difference in the bid-ask spread, typically about a penny. Therefore, a dollar into the market only creates about a penny of additional market capitalization. Gabaix and Koijen, using some very sophisticated mathematical techniques, derived that that number wasn't even remotely close. They estimated it was between five and eight. They were using a time period average and so didn't actually capture the trend of that rising over time. We have subsequent research from Valentine Haddad and others, and I've done my own research on this, that suggests that that's actually a rising factor, that we're seeing larger and larger impacts associated with the market's loss of elasticity tied to the growth in share of passive investing.

Yeah. So Hari shared some of our initial findings over Twitter, and it's important to specify, you know, the characteristics that we assumed in that. So one of the assumptions is actually that's not having any inflows or outflows. And so that number of about 83% is actually a static number that if you get to that level of passive, effectively the market becomes so impossibly volatile that it becomes an inevitable event that it will eventually cause its closure. In an unrestricted market like XIV, where we were at very similar levels, you had the capacity to go to zero with the circuit breakers in place and the restrictions on that type of trading in place on the US equity markets.

You know, our underlying presumption is effectively that the markets would stop clearing, they would close the markets. And that this would happen over a repeated, you know, period of days as people tried to make redemptions in response to the market being closed. You know, that's a very conservative model. I want to emphasize a couple of features that are being put in there. One is that we are assuming that anyone who is not passive is fully active, effectively having a historical elasticity of about 75 to 80. Secondly, we are presuming no flows, as I mentioned, so that doesn't include any redemptions, it doesn't include any contributions, etc. It's simply an endogenous feature that plays out in the market. And some of this work was actually a replication of work that was done by Andrew Lo on a similar topic, where he came to a very similar conclusion: between 75 and 80% passive, the market would effectively cease functioning.

If we make some assumptions about declines in elasticity for active managers, which have absolutely happened, I'll share a few slides on that, that number moves significantly lower. If we start incorporating components of flow into that, that number also begins to fall dramatically. And so we basically treat that 83% as an outer limit. To your point about where we are right now, my estimate is actually we're about 54% passive by market share, and we picked up about 4% last year. So, you know, if that pace continues itself, we would be looking at somewhere in the neighborhood of five years out, the world comes to an end. Again, this is, you know, we have high confidence in the outer limits of the model. Again, it replicates some of the work that I did with Peter Thiel. It also replicates some of Andrew Lo's work. And so, you know, I would consider this basically the starting point, not our end point.

Well, this is part of the problem, right? Is once it becomes a mathematical certainty, right? If you actually have built a closed-form solution that says this is just what the math says will happen, then, you know, we're effectively debating math. And to my knowledge, math doesn't fail. You know, the assumptions could certainly fail. And so, you know, that's kind of the point that I would emphasize. Like, could they change the rules? Could the Federal Reserve inject trillions of dollars into the US equity markets? Could we have a fiscal package that includes an exchange stabilization fund, etc.? Those are all legitimate, you know, responses. It is also legitimate to close the market and say, everybody has to deal for a while. We've done that before. We did it at the start of World War I. You know, regardless, the point I guess I would emphasize is, you know, do I think it will happen? Yes, I do. Do I think that there will be extraordinary responses in response to extraordinary events? Yes, I do.

No. So what the data suggests is an acceleration, which unfortunately makes sense because we know that more than 100% of the marginal flows that are coming in are passively managed assets. You know, effectively what that means is that the active managers are facing a shrinking pool of assets. Their resources are deteriorating. I think there's going to be an imminent restructuring of the active management space as the reality of trying to manage significantly less assets at much lower fees begins to clash with the business models that have historically had this as a highly remunerative industry where, you know, you were willing to spend a significant amount of money.

Yeah, it is slowing these flows, right? And so this is actually one of the interesting features that happens as more and more of those approaching retirement actually own passive vehicles and start to sell those passive vehicles to finance their retirements. Those flows will slow and they will begin to reverse. I think it's actually really critical to understand, you know, across the mutual fund complex, and mutual funds are the predominant vehicles in 401ks, those retirement assets that we talk about when we talk about 401ks are almost inevitably in mutual funds. You are seeing growing use of ETFs in there, and so I do want to be very cognizant that there is a difference there. So there is a degree of inevitability to those flows turning and beginning to decline. That doesn't mean money is going to flow into active management, though, just to be very, very clear, right? It actually means that you're going to see simultaneous redemptions out of active management and passive management. And that's the scenario that concerns me most. Effectively, those who should be searching for value are deprived of capital at the same time that those who are trying to convert equities into cash face a very challenging situation because they are trying to sell to people who don't have cash to meet those redemptions. Let me pull this up here quickly. What we're seeing is we're actually now starting to see redemptions coming from passive mutual funds. This is a change, right? This tells you that those retirements are beginning to accelerate.

What's happening to the positioning is that you're seeing Vanguard index mutual funds get redemptions. Now, that is an indicator that the boomers are retiring. At the same time, we have more than offset flows into the ETF. Currently, if you look at something like the positioning in Nvidia or the positioning in Apple or Microsoft, you're seeing evidence that there are redemptions coming out of these funds. And the net right now is getting quite a bit closer to zero. If that flips, then that changes that calculus we talked about before. I want to be very, very clear. It's possible that Vanguard is losing share. It's possible that other firms are taking share from them given Vanguard's relatively sparse representation within ETFs. But I don't think that's what's going on. I think we're actually starting to see indications that we are actually beginning to see outflows. And that should be concerning. That pulls those numbers closer. That pulls those dates closer.

It depends on what happens to active managers. It depends on what happens to active managers, right? Because if they continue to lose at the pace that they have been losing, last year active managers lost assets somewhere in the neighborhood of $600 billion worth of redemptions, while passive pulled in somewhere in the neighborhood of a trillion dollars worth of inflows. And that $600 billion is a larger share of the active manager share at this point than the change that we would expect to see within the passive vehicles. So it continues to show share gain.

Yeah, we actually know that. So, you know, somewhere in the neighborhood of 95% of the contributions that are coming in through the younger generation are flowing straight into target date funds through their 401k retirement plans, which they are automatically defaulted into. That's less true for the boomers. That's, you know, if you basically go from boomers who never really had forced allocations to 401ks and hence were the source of the active manager flows into mutual funds, all the way down to the current crop of those entering the labor force. Those entering the labor force today, it's almost exclusively passive.

I think that's stupid. So you can pull up the chart that I have that shows the composition of trading volumes. What they are ignoring and focusing on is strictly the direct trading that is going into passive vehicles. You had Figure 6 from their paper, which shows the index trading volume adjusted for cash flows, which means the inflows that I was talking about, right? So all they're capturing when they adjust for cash flows is basically the index reconstitution trading, which by definition only happens three or four times a year and is effectively meaningless, right? These are very small changes that they're capturing in their data, where they are effectively looking at things like new issuance or share repurchases that may modestly change the overall positioning, new index inclusion, etc. What they're excluding is those inflows, right? That's what the cash flow adjustment means. And so they know the impact that they're having. They are being intentionally misleading in this chart. If you go to the one before that from Hagstrom in 2013, that and Market Structure Edge, you know, you can get a much more holistic view of what ultimately is going on here, which is that passive trading now makes up roughly 80% of market volume on a daily basis. That's tied to the direct trading, the market making that facilitates ETFs and index funds trading close to fair value, their NAVs, the options that need to be traded, and the hedging around options that are struck against indexes, etc. You know, I candidly think that Vanguard should be held liable for misleading people at this point, but that's what I think they're doing.

If I go back to 1995, active management made up 80% of the trading activity. My estimate today is it's around 6 or 7%.

And part of that is also happening because the active managers themselves have shared work on this. The mechanics of this type of motion, where money is flowing into a very specific algorithmic allocation, is effectively pushing those securities higher at rates that exceed the overall market. I'm not the only one who has identified this. Jiang at Michigan State, for example, wrote an entire paper on it. You know, the implications of that are quite profound, but one of them, of course, is that from an active manager perspective, in order to maintain performance close to the benchmark and not run the risk of losing their jobs, they are themselves becoming more and more index-like and closet indexers. That's driving reduced elasticity amongst the active managers. I could share a chart with you on that as well. But before I do that, the other component that I would just emphasize on it is this is mechanically why active managers are underperforming more and more, even as theoretical frameworks like Grossman-Stiglitz suggest that, you know, this should be a field day for active managers because we have all these non-thoughtful investors that are going out and buying things at ridiculous value. You know, if I, let me just quickly show you the chart.

This is again from Haddad. This is looking at the change in elasticity for those they continue to designate as active managers. And so this is another one of these interesting pieces of research that talks about how much larger the passive share actually is than is measured. But they are treating anyone who doesn't have elasticity effectively of zero as an active manager. And so the active managers themselves used to have elasticities in the range of 75 to 80. Today that number is well below 50%. Part of what's happening here is more and more active managers like Fidelity are emphasizing index funds. One of the reasons why that's happening is the regulatory framework changed radically in 2006, that was implemented in early 2008, that required 401ks to default into passive vehicles, and any selection of non-passive vehicles actually exposed the sponsor of the 401k to liability. This has become one of the nice businesses for trial lawyers. And so we're seeing firms basically become more and more passive in their construction, effectively mimicking the successful species if we think about it in evolutionary terms.

Yeah. No, it's, by the way, I would say the same thing about forecasts for small caps, you know, resurging and rotations into emerging markets, etc. There's a lot of people who are hanging their hat on a very substantive change. You know, one of the points I would make is like if people do decide to rotate their portfolios and pursue alternative approaches, that also is a mechanism effectively of reasserting flows out of passive vehicles and just pulling that point closer and closer.

I think there's a couple of things, and again, I think they're being intentionally misleading. So when they're describing index funds in this context, they're taking a slightly different approach. They're including index funds that would be like the REIT index or a nuclear power index, right? Those push the ownership, you know, of quote-unquote passive strategies that are not at all passive strategies, not even remotely close to what we're talking about in a market-cap-weighted framework. And many of those are actually being forced into mid-caps by virtue of concentration limits that exist within portfolio construction. And so often you'll see something like a REIT index that has a modified cap-weighted exposure, where if they were to simply buy in proportion of the market cap, it would be, you know, three stocks for all intents and purposes. By conforming to the modified cap-weighting formulas to meet the diversification requirements of the 40 Act, it pushes them into large ownership of some, you know, mid-cap companies. You know, those absolutely have an impact on those mid-cap companies, but nobody in their 401k has the, you know, the REIT ETF as their default investment vehicle. So almost everything that we're talking about, they are intentionally avoiding and obfuscating with this analysis.

Yeah, again, they're being intentionally obfuscating. Right. What they're highlighting here is that after they trade, they own the same share of every company. There's no discussion about the impact of that trade on the price and the market capitalization of the individual companies. They intentionally avoid that. They're simply showing us the math of market-cap weighting. What we know, and you have this directly below in the notes that you sent me, is the work of J.P. Bouchaud on market impact. Market capitalization plays no meaningful role in the provision of liquidity. What matters for liquidity is the dollar value of trading and the size of the order relative to the idiosyncratic volatility of the individual name. If you actually pull that formula up on screen, which I assume you're going to do in the edit of this, the impact of an order of size Q is going to be a function of the volatility of that either market or individual security times the square root of the quantity, the dollar quantity traded, divided by the average daily volume that's traded. This makes intuitive sense. If I go into a market that on average trades a...

million dollars and I try to execute a billion dollar trade, I'm going to have a big impact on that market. That impact is going to be larger for highly volatile stocks because they have effectively less anchor to them. So what this tells us is that we should be looking at dollar volume of the size of the order relative to the dollar volume of the trading liquidity. This is part of what I actually ran through on one of my Substacks. Let me just pull that up quickly. This was in a piece that's on my Substack called 'Pay Attention to Your Privates.' If you take the example of Apple and CarMax, which at the time that I wrote this were respectively close to the largest and close to the smallest stocks in the S&P 500, if you put in a billion dollar order to buy the S&P 500, and that's not a typical order for an index fund, about $70 million of that order is going to go to Apple and about $100,000 is going to go to CarMax. Right now their trading volumes are such that that differential, that 700 to one differential, isn't even close. This is about a 60 time differential. So it's a roughly order of magnitude difference as a share.

Their volatilities are somewhat similar. The volatility of CarMax is slightly higher. So if we run through this formula, what we discover is on a daily basis, Apple is receiving a positive impulse. And again, you know, Vanguard's inflows last year were about $350 billion. That would suggest about a billion dollar order every day. So we're looking at every single day Vanguard gets inflated by that order by .167%. Right? CarMax gets inflated by .006%. If you scale that to an annual number, that works out to about a 6% increase in CarMax and about a 23% increase in Apple.

By the way, that's actually more than Apple gained last year. So, you know, there are offsetting components to it. Among them, the fact that candidly, Apple is not growing anymore. It doesn't have superior profitability or investment opportunities. It's failed in almost all of its new product introductions. Sure, there may be something fantastic out there, but that's a really big leap to make at the scale that we're talking about with Apple. So what we should be seeing in Apple is lower multiples. We should be seeing lower valuations reflecting the fact that the law of large numbers has largely caught up to Apple and it should be trading at lower valuations. Instead, it's trading near the highest valuations in its history. And I would ascribe that to the impact of passive investing. Vanguard, of course, would say we have no role in it whatsoever.

My hunch is that there are people smarter than me that have taken my research and converted it into this. I'll be totally candid with you. It took me basically nine years to get here, which reinforces why I call myself the dumbest man alive on a regular basis. And it really was kind of a breakthrough where I realized, wait a second, the S&P 500 isn't a single security. When I did my initial trades around this, I did it on the XIV, which was a security that referenced a single underlier, two underliers, but that's splitting hairs. There's 503 securities within the S&P 500. And so it actually turns out that there's much more impact occurring at the individual stock level and that's driving behaviors within the index that will ultimately turn the index into something that looks like a single stock, which is increasingly what we've seen with the rise in concentration. Again, a forecast that I made all the way back in 2017 that everybody thought was insane at the time, but has continued to play out.