The prior piece received a much stronger reaction than I expected. The topic is complicated, with ideas that are difficult to adequately convey in words, so I’m going to use this piece as a follow-up. I’m going to look at the specific example of the Tech Bubble, which is a clear case in which the broad adoption of a passive approach would have removed negative skill from the market and substantially increased market efficiency. I’m also going to offer some thoughts on the topics of differential liquidity, niche ETFs, and the drivers of inefficiencies in the area of asset allocation.
To begin, let me offer a brief comment on the topic of market efficiency. Recall the definition:
Efficiency: The extent to which all securities in a market are priced to offer the same expected risk-adjusted future returns.
In this definition, we have to emphasize the term “expected” because returns have a random component, and investors can only evaluate returns based on information that is available to them at the time. We also have to emphasize the term “risk-adjusted”, particularly when we try to talk about “efficiency” in terms of the relative pricing of securities that have different risk profiles–think: equities, bonds, cash. We don’t expect an efficient market to price these asset classes for the same return, but we do expect any differences in return to be justified by other differences in the securities: differences in liquidity, volatility, loss potential, and so on.
Sometimes, you will see the term “efficiency” defined in terms of the relationship between a security’s price and its future cash flows. For example:
Efficiency: The extent to which a security’s price matches the discounted sum of its future cash flows.
But this definition is not helpful, as the discount rate is not specified. Any cash-producing security can be described as “efficiently” priced if we get to “make up” our own discount rates from nothing. Now, to improve on the definition, we have the option of imposing the constraint that discount rates be applied consistently across different securities. But then the definition reduces to the original definition, because a “discount rate” is nothing more than a required rate of return for an investor. When we require that all cash flows in a space be discounted using the same rate, we are requiring that all securities offer the same rate of return, which is exactly where we started with the original definition.
The original definition of efficiency, which is the definition proposed by the person that invented the concept, is the best one for current purposes, because it directly connects “efficiency” with the concept of it being impossible to outperform in a space through active security selection. Pick whatever securities you want in a space. If the space is efficient, you will not be able to outperform, because every security you might pick will be offering the same return, adjusted for risk, as every other security.
Now, in the piece, I tried to make the following argument:
- Passive investing increases the efficiency of the space in which it is being used.
I underline that last part because it’s crucial. Passive investing does not increase the efficiency of spaces in which it is not being used, nor should we expect it to. To use an example, investors might be bullish on U.S. equities relative to other asset classes. They might express that bullishness using $SPY. But the passivity of $SPY is incidental to what they’re doing. They are expressing an active preference–a preference to overweight large cap U.S. equities and underweight everything else–using a passive vehicle. The end result, a situation in which large cap U.S. equities are overvalued relative to everything else, would have been no different if they had expressed that same preference using an active vehicle–say, Fidelity Magellan fund. Nor would it have been any different if they had simply gone into the space of large cap U.S. equities and made the picks themselves, based on whatever they happened to like. The cause of the inefficiency has nothing to do with the passive or active nature of the vehicle used, and everything to do with the underlying bullish view that is driving its use–a view that is not passive, and that we are assuming is also not justified.
Now, the primary difference between using a passive vehicle and an active vehicle to express a view lies in the relative impact on the names inside the space that the vehicle is being used in. The passive vehicle will tend to have a minimal relative impact, because it is buying a small amount of everything in the space, in proportion to the amount in existence. The active vehicle will have a larger relative impact, because it is buying only what the manager–or the stock-picker–likes in the space.
To make the point maximally precise, suppose that the space in question is the space of [All U.S. Equities]. Note that when I put the term in brackets, I mean to refer to it as a space, a collection of securities:
[All U.S. Equities] = [$AAPL, $MSFT, $XOM, $JNJ, $GE, $BRK-B, and so on…]
We’re comparing two different ways of investing in [All U.S. Equities], active and passive. To invest actively in that space is to go in and buy whatever securities inside the space you happen to like, those that you think will perform the best. To invest passively in the space is to own a little bit of everything in the space, in proportion to its supply, i.e., the “amount” of it in the space, i.e., its”weighting.”
When an unskilled or negatively skilled participant chooses to invest actively in [All U.S. Equities], he reduces the efficiency of the space. He makes bad relative picks. In doing so, he creates opportunities for other participants, those that do have skill, to make good relative picks at his expense, as his counterparties.
To illustrate using the momentum factor, suppose that we have an unskilled or negatively skilled participant whose strategy is to buy those stocks that have gone down the most over the last year.
“When I invest, I keep it simple! I look for whatever is the most on sale! The biggest and best bargains!”
This participant has no understanding of the fundamentals of the companies he ends up buying, no understanding of the attractiveness or unattractiveness of their valuations as businesses, and no knowledge of the momentum patterns that markets have historically tended to exhibit on 1 year lookback periods. All he knows is the price–the more it has fallen, the more he wants to buy. Obviously, this participant is going to create opportunities for skilled participants–those that are familiar with the fundamentals and valuations of the companies being traded, and that are aware of the market’s historical momentum patterns–to outperform at his expense.
To be fair, maybe you don’t believe that there is such a thing as “skill” in a market. If that’s the case, then ignore what I’m saying. You will already be an advocate of passive investing–anything else will be a waste of time and money on your view. The argument, then, is not directed at you.
Now, suppose that Jack Bogle, and Burton Malkiel, and Eugene Fama, and the Behavioral Economists, and the robo-advisors, and so on, manage to convince this participant that he should not be investing in [All U.S. Equities] by making picks in the space himself–either of individual stocks or of expensive managers–but should instead be investing in the space passively–for example, by buying $VTI, the Vanguard Total U.S. Equity Market ETF. To the extent that he takes their advice and does that, he will no longer be introducing his flawed stock-picking strategy into the space, where other investors are going to exploit it at his expense. The efficiency of the space–i.e., the difficult of beating it through active security selection–will therefore increase.
Unfortunately, this is the place where things get muddled. If we’re talking about the space of [All U.S. Equities], then passive investing, to the extent that it is preferentially used by lower-skilled participants in that space, will increase the space’s efficiency. But, apart from that increase, it will not increase the efficiency of other spaces–for example, the broader space of [All U.S. Securities], where:
[All U.S. Securities] = [U.S. Equities, U.S. Bonds, U.S. Cash].
To illustrate, suppose that investors get wildly bullish on U.S. equities, and try to allocate their entire portfolios into the space, all at the same time. Suppose further that the lower-skilled participants of this lot, in making the move, choose to heed the advice of the experts. They shun the option of picking winners themselves, and instead buy into a broad equity index fund–again, $VTI. Their approach will be passive relative to the space of [All U.S. Equities], but it will not be passive relative to the space of [All U.S. Securities]. Relative to [All U.S. Securities], it will, in fact, be extremely active.
If lower-skilled participants express their demand for U.S. equities by buying $VTI rather than picking stocks in the space themselves, they will have eliminated the market inefficiencies that their picks inside that space would otherwise have introduced. Consequently, the efficiency of the space of [All U.S. Equities] will increase.
But, and this is the crucial point: beyond that increase, the space of [All U.S. Securities] will not become more efficient. That’s because the approach of putting an entire portfolio into $VTI is not a passive approach relative to [All U.S. Securities]. Relative to that space, the approach is just as active as the approach of putting an entire portfolio into Fidelity Magellan, or an entire portfolio into [$FB, $AMZN, $NFLX, $GOOG] “because I buy what I know.” U.S. Equities will still get expensive in comparison with other asset classes, because a preferential bid is still being placed on them. The fact that a passive vehicle, rather than an active one, is being used to place the bid is not going to change that result.
Now, a great example to use to concretely convey this point is the example of the market of the late 1990s. Hopefully, we can all agree that that the market of the late 1990s had two glaring inefficiencies:
- Inefficiency #1: In the space of [All U.S. Equities], some stocks–for example, new-economy technology stocks–were wildly expensive, well beyond what a rational analysis could every hope to justify. Other stocks–for example, stocks with ties to the old-economy–were reasonably valued, especially in the small cap space. You could have outperformed the market by buying the former and shorting the latter, and you could have known that at the time, assuming you weren’t drinking the kool-aid. Therefore, the market was inefficient.
- Inefficiency #2: In the space of [All U.S. Securities], which again consists of [All U.S. Equities, All U.S. Bonds, All U.S. Cash], the U.S. Equities component, as a group, was much more expensive than the other two asset classes, priced for much lower returns, especially after adjusting for the dramatically higher levels of risk. Again, you could have outperformed the market by holding cash and bonds, shunning equities, and you could have known that at the time, assuming you weren’t drinking the kool-aid. Therefore, the market was inefficient.
Recall that back then, passive investing wasn’t even close to being a “thing.” In fact, it was frowned upon:
“Vanguard? What? Why would you want to dilute your best holdings with companies that don’t have the potential to grow in the new economy? Doing that will significantly reduce your returns.”
“Yes, you need to diversify. But there’s a way to do it without sacrificing growth. What you need to do is this: mix your more established growth names–e.g., your Microsofts, your Intels, your Ciscos, your Yahoos, basically, the reliable blue chip stuff–with higher-risk bets–something like an Astroturf.com, or an Iomega, or a JDS Uniphase, that will offer high-powered upside if things work out for those companies. That will give your portfolio balance, without sacrificing growth. And if you need someone to help you set that up, we can certainly help you do that.”
Remember?
Since then, passive investing has become significantly more popular–so much so that a number of top investors have publicly described it as a “bubble” unto itself. And to be fair to those investors, the increased popularity does have certain bubble-like characteristics. Gurus on TV advocating it? Check. New companies springing up to exploit it? Check. Rapid increase in AUM? Check.
Now, suppose that passive investing had been as popular in 1999 as it is today. Instead of expressing their cyclical equity bullishness by utilizing their own stock-picking expertise–or their expertise in picking stock-pickers, which was just as flawed, evidenced by the fact that everyone was piling into the same unskilled Technology managers–suppose that investors had simply used a broad market index fund: say, $VTI. What would the impact have been?
With respect to inefficiency #2–i.e., the inefficiency that saw the U.S. equity market wildly overvalued relative to other asset classes–that inefficiency might not have been ameliorated. Bullishness for an asset class tends to lead to overvaluation, regardless of the vehicles used to express it, whether passive ($VTI) or active (Fidelity’s top performing fund, you going into your E-Trade account and buying into a stock that your friend at work just made a bunch of money on).
But inefficiency #1–the inefficiency that saw certain parts of the U.S. equity market trade at obnoxious prices, while other parts traded quite reasonably–would definitely have been less severe. It would have been less severe because the segment of the market that was doing the most to feed it–the lower-skilled retail segment that, in fairness, didn’t know any better–wouldn’t have been playing that game, or feeding that game by picking unskilled managers who were playing it.
This example, in my view, is a perfect example of how the widespread adoption of a passive approach can make a market more efficient. Granted, it can only do that in the specific space in which it is being used. Right now, it’s not being appreciably used in the broad space of [All U.S. Securities] or [All Global Securities]. But it is being used in the space of [All U.S. Equities]. We can debate the degree–and I acknowledge that the effect may be small–but I would argue to you that the use has made the space more efficient, even as we speak.
If you disagree, then do this. Go find your favorite value investor, and ask him what his biggest gripe right now is. This is what he will tell you:
“The problem with this market is that everything is overvalued. Everything! In 1999, I was able to hide out in certain places. I was able to find bargains, especially on a relative basis. Right now, I can’t find anything. The closest thing to an opportunity would be the energy sector, but that sector is only cheap on the assumption of a rebound in energy prices. If we assume that prices are not going to rebound, and are going to stay where they are for the long-term, then the typical energy name is just as expensive as everything else in the market. Yuck!”
Here’s a question: what if that the widespread embrace of a passive approach with respect to U.S. equities in the current cycle is part of the reason for the broadness of the market’s expensiveness? Personally, I don’t know if that’s part of the reason. But it certainly could be.
Now, before you point to that possibility as evidence of the harm of passive investing, remember the definition of efficiency:
Efficiency: The extent to which all securities in a market are priced to offer the same expected risk-adjusted future returns.
On the above definition of “efficiency”, an expensive market that gives investor no way to escape the expensiveness–e.g., the 2013-2016 market–is actually more efficient than an expensive market with large pockets of untouched value–e.g., the 1999 market. If the increased popularity of passive investing has, in fact, reduced the dispersion of valuation across the U.S. equity space–the type of dispersion that lower-skilled participants would be more inclined to give support to–then the outcome is consistent with the expectation. Passive investing has caused the U.S. equity market to become more efficient, harder to beat through active security selection. The suckers have increasingly taken themselves out of the game, or have been increasingly forced out of the game, leaving the experts to fight among themselves.
A number of investors seem to want to blame passive investing for the fact that U.S. equities are currently expensive. Doing that, in my view, is a mistake. The expensiveness of U.S. equities as an asset class is not a new condition. It’s a condition that’s been in place for more than two decades–long before passive investing became a “thing.” If you want to blame it on something, blame it on Fed policy. Or better yet, blame it on the causes of the low-inflation, low-growth economic condition that has forced the Fed to be accomodative, whatever you believe those causes to be. The Fed has responded entirely appropriately to them.
It’s hard to blame the current market’s expensiveness entirely on the Fed, because there was a long period in U.S. history when the Fed was just as easy as it is today–specifically, the period from the early 1930s through the early 1950s. In that period, equities did not become expensive–in fact, for the majority of the period, they were quite cheap. Of course, that period had an ugly crash that damaged the investor psyche, at least in the earlier years. But the same is true of the current period. It has had two crashes, one with a full-blown crisis. Yet our market hasn’t had any problem sustaining its expensiveness.
If we had to identify an additional factor behind the market’s current expensiveness, one factor worth looking at would be the democratization of investing. Over the last several decades, investing in the market has become much cheaper in terms of the transaction costs, and also much easier in terms of the hoops that the average person has to jump through in order to do it. It makes sense that this development would increase participation in the equity markets, and therefore increase prices.
Does it make sense to hold cash at an expected future real rate of, say 0%, when you can invest in equities and earn the historical real rate of 6%? Clearly not. In a period such as the late 1940s, the barriers to having a broad section of the population try to arbitrage away that differential may have been large enough to keep the differential in place. But today, with all of the technological innovation that has since occurred, the barriers may no longer be adequate. A 6% equity risk premium may simply be too high, which would explain why it has come down and stayed down. To be honest, I’m not fully convinced of this line of reasoning, but it’s worth considering.
Returning to the current discussion, it’s important not to confuse the democratization of investing–the development of broad online access to investment options at negligible transaction costs–with the increased popularity of passive approaches. These are fundamentally different phenomena.
It may be the case that in the current cycle, investors are using the familiar passive vehicles–$SPY and $VFIAX–to drive up the market. But so what? The passivity of the vehicles is not what is fueling the move. What’s fueling the move is the bullish investor sentiment. If, instead of being sold on the wisdom of a passive approach, investors were skeptical of it, then they would simply express their bullishness actively, piling into popular active mutual funds, or picking stock themselves–which is exactly what they did in 1999, long before the average investor even knew what $SPY was.
Before concluding, let me briefly address a few caveats:
First, liquidity. Passive vehicles are not sensitive to differentials in liquidity. So, to the extent that those differentials exist across a space, widespread use of passive vehicles in a space can create exploitable distortions.
To give an example, if a micro-cap ETF sees net inflows, it’s going to use those net inflows to make purchases in the space. Given its passive approach, the specific purchases will be spread across the space, regardless of the liquidity differences in the individual names. So, if there is a part of the space that is highly illiquid, prices in that part will get pushed up more than in other parts, creating a potential inefficiency.
If net flows go back and forth into the micro-cap ETF, it will engage in costly round-trips in the illiquid security, thereby losing money to active participants in the market–specifically, those that are making a market in the illiquid security. In contrast with a passive ETF, an active ETF or mutual fund can avoid this outcome because it is able to direct its purchases towards liquid areas of a market, or at least factor the cost of illiquidity into its evaluations of potential opportunities.
As a rule, passive vehicles tend to generally be less “disruptive” in their purchases than active vehicles because they buy securities in proportion to the supply of shares outstanding. In a sense, the demand they inject is intrinsically tethered to the available supply. One caveat noted by @SmallCapLS, however, is that the actual supply that matters, the floated supply in the market, may be different from the total shares outstanding, which is what determines the weighting. In situations where the float is tiny, and the outstanding share count very large, passive vehicles have the potential to be more disruptive.
The market impacts of these potential liquidity-related inefficiencies are minimal. Not worth worrying about, especially for those investors that are not involved in illiquid spaces.
Second, niche ETFs. No one has raised this specific objection, but after thinking the issue through, it would be hard to deny that certain highly niche ETFs–for example, the proposed 3-D printing ETF $PRNT–have the potential to create distortions, if not simply by giving inexperienced investors bad ideas, i.e., suckering them into making bets on spaces that they should not be making bets on. But again, we shouldn’t pin the distortions onto the passivity of the funds. With repect to 3D printing, for example, there’s an active mutual fund that offers (or offered) the exact same exposure: $TDPNX, the 3D Printing, Robotics, and Tech Investment fund. If you look at the active mutual fund graveyard of the last 20 years, you will find plenty of the same types of funds, each of which managed to pull investors into a fad at exactly the wrong time. It’s not a passive vs. active thing, but a feature of the business–gather assets, in whatever way works.
Third, when I said that nobody is using a passive approach relative to the broader space of global securities–i.e., all foreign and domestic equities, all foreign and domestic bonds, all foreign and domestic cash, owning each passively, in proportion to its outstanding supply–I did not mean to say that this could never be done. It can definitely be done–there just isn’t a demand to do it.
Consider a hypothetical ETF–we’ll call it $GLOBE. $GLOBE holds every liquid financial asset in existence, including currencies, in relative proportion to the supply in existence. Right now, investors don’t have an easy $GLOBE option–the closest option would be a robo-advisor. The difference between $GLOBE and a robo-advisor, however, is that the robo-advisor uses a predefined allocation–e.g., 60% stocks, 40% bonds/cash–as opposed to a “market” allocation that is based on the outstanding supplies of the different asset classes.
Using a pre-defined allocation can create inefficiencies in cases where the supply of one of the asset classes is naturally tighter than the other. When everyone tries to allocate in such a configuration, the tighter asset class gets pushed up in relative price and valuation, resulting in a potential inefficiency. The dynamics of this process were discussed in more detail in a prior piece.
Returning to $GLOBE, in my view, a liquid fund in the mold of $GLOBE, if it caught on with the lower-skilled segment of the market, would make the market more efficient, not less. Unskilled investors would be able to use the vehicle to remove their unskilled timing and rotation decisions from the market fray, reducing the opportunities of skilled investors to outperform at their expense. In truth, the effect on broad market efficiency would probably be very small, if perceptible at all. Regardless, markets would not be made any less efficient by it, as opponents of passive approaches want to suggest.
As markets continue to develop and adapt going forward, the investment area that I suspect will be the most likely to continue to present inefficiencies is the area of asset allocation. Asset allocation is an area where a passive stance is the most difficult to fully embrace. The temptation to try to “time” asset classes and “rotate” through them in profitable ways can be quite difficult to resist, particularly when conditions turn south. Unlike the game of individual stock-picking, lower-skilled investors are not likely to want to voluntarily take themselves out of that broader game–in fact, of all market participants, they may be the least inclined to want to do so.
Decisions on how to allocate among broad asset classes–equities, fixed income, and especially cash–are among the least professionally-mediated investment decisions that investors make. If you want to invest in U.S. equities, there are tons of products available to help you. These products eliminate the need for uninformed participants to make relative bets inside that space.
But the decision of what asset classes to invest in more broadly, or whether and when to invest at all–whether to just keep the money in the bank, and for how long–is a decision that is much more difficult to offload onto someone else, or onto a systematic process. Even if an advisor is used, the client still has to make the decision to get started, and then the decision to stick with the plan. “The stock market is doing really well, at all time highs. But I’m not invested. Should I be?” “This isn’t working. Our stuff is really down a lot. Are we in the right setup here?” Naturally, we should expect a space in which unskilled investors are forced to make these decisions for themselves–without experts or indices to rely on–to exhibit greater inefficiency, and greater opportunity.