Even if you don't care about the index fund "front-running" "scandal", the section starting at "The value of market-making is hard to see and easy to criticize" is critically important to understanding why the markets work the way they do.
You know what's easier to criticize? Manipulating LIBOR and nobody going to jail. The downstream effects of instruments pegged to LIBOR is staggering, well, would be if the industry / reglatory agencies actually did anything of merit.
Remember, the main rationalization for Bernie Madoff's unbelievably consistent returns was that he was front-running, and fund after fund after fund after fund lined up to give him money. Besides, front-running isn't really where the big money is anyway. Insider trading is way, way more profitable from an individual standpoint.
From what I can see, this article is on point, but is missing an important factor: the risk these "front runners" take. As soon as the announcement is made that a company is joining the index, it's public knowledge. In theory, the expected increase, minus a risk premium, should be priced in immediately. There will likely still be money to be made over the following days until the addition is complete, but it's far from guaranteed, and comes at the expense of reduced diversification. (Which I suppose is another way to say that you're getting paid for providing liquidity, as the article says.) Just because AA went up X% over the 4 days, or whatever, before it joined the index, doesn't mean the next stock will. Perhaps its jump will be overestimated by the HFTs, and retail investors trying to get in in the days following the announcement will end up losing money. Probably not, but it's certainly a significant possibility. So if a person wanted to pursue this active strategy, they would need to manage their risk appropriately. It's not necessarily a bad idea if you enjoy spending your time on that kind of thing, although personally I'd rather index (with a moderate small/value tilt).
Likely true. Also not without risk of course, since the chance of a stock getting added should also be priced in. If you're better than the "market" at predicting these things, you'll likely do well. I don't expect I am.
Order-handling companies pay for "dumb"
flow. Vanguard can reduce their outright trading costs to negative by being as dumb about it as possible, and then use these negative costs to artificially lower their reported fees.
Just because Vanguard claims to be smart about it, doesn't mean necessarily they actually are incentivized to be smart about it or actually are in practice. People can still judge them by how close they track the index, but that is reported separately from fees, which are all a lot of current and future retirees look at after having fees fees fees drilled into their heads. And the indexes themselves take a hit, so you need to adjust for that with a much more complicated measure.
They can effectively launder bad (or even good) tracking of the index into lower reported fees, by letting the order handlers profit on the inanity (and kickback via order-flow payments), allowing the fund managers to give themselves higher compensation without commiserate alarming fees.
To what extent, if any, do they actually do this? Do they report their income from paid order-flow in the fund prospectuses? Do they break it out by the managed funds, vs retail flow from their clients? Do they get major concessions to their retail trading costs in tacit exchange for being dumb with their etfs?
I haven't offered any, and it would most likely be illegal if it was explicitly going on. And there is likely a wall between the different trading desks (though often the physical embodiment portion of this is literally a cubicle wall the employees can hear each other over). But the orderflow compensation doesn't need to cross over into the retail desk if the ETF itself has enough trading volume to mask some kickback without appearing too egregious.
But price fixing is also illegal--nevertheless, two gas stations across the street at a profitable intersection can engage in it solely through price signal tit-for-tat[1]. This effectively masks intentionality.
Much more fantastical and speculative: machine learning algorithms at both firms could now, or in the future, arrive at this cooperative strategy, even with the only communication being through price signals. Without any human ever even knowingly giving the explicit go-ahead.
Short answer is that Vanguard Brokerage Services did not receive compensation for equities order flow in 1QTR 2015, but did receive compensation for options order flow.
I have no idea what the laws are surrounding their index funds or how those index funds are allowed to/actually do interact with Vanguard Brokerage Services.
There is a principal agent problem between managers and shareholders. Managers get paid out of shareholders' pockets. If the shareholders are over-focused on fees as a metric, which they often are, managers have some potential tools (which I went into) to launder part of the fees into poor fund performance. I'm just throwing it out there, I'm not saying it is actually going on.
---
As for proxy voters choosing the management, proxy vote research usually doesn't make sense. You see strategies amongst large hedgefunds like this: keep their valuable proxy vote research private, make a vote based on the most predictable-to-them-but-not-to-you outcome based on the research (positive or negative for the company, doesn't matter unless it will be picked up by others immediately, even then they can do things like exiting their position through an obscure hedge), and then hold or sell their shares based on the overall vote outcome and its implications in the research.
Researching and making proxy votes out of naive benign interest of the company is often just doing altruistic work for a greater collective, something markets frown on and usually punish.
Sure, but it still makes it harder for managers to screw the shareholders, especially when compared to funds like ishares where there is no accountability.
It is basically order flow that often blindly takes liquidity. Traders pay to get it, execute it prop at the best price (though they also can dump straight to the market without taking it prop if it isn't behaving dumb enough). Then they resell the position on the market at a more deliberate pace, providing liquidity (which lets them capture the spread, and lets them earn kickbacks from exchanges that pay for liquidity). Or they match the positions against future dumb flow coming in on the other side.
Shareholders of the ETFs. Again, I'm not saying this specifically actually happens in the case of Vanguard, but all kinds of other conflicts of interest like this happen all the time, and reporters should be ever skeptical, and not just report he-said she-said, "oh Vanguard says they got this covered."
On Monday an index is made up of 100% MSFT, and it announces that on tuesday the index will be rebalanced to 100% APPL. The index has ETFs tracking it that are large enough to exhaust the availability of liquidity in both MSFT and APPL during a quick rebalance. Eventually over time the liquidity arrives to correct the mispricing, valuing things based again on underlying fundamentals. Both the index and the shareholders of the ETFs take a hit as the price of MSFT rises (no longer part of the index) and APPL (part of the index) lowers back down to normal, assuming no changes in fundamentals in the meantime.
The ETFs can be smart about it and try and make the trades over time instead of all at once, and Vanguard likely does. This then feeds back in and lowers the amount that the actual index takes a hit. The equilibrium in reality is that both the indexes and the shareholders in the ETFs take a bit of hit.
This is called the "index rebalancing" trading strategy. Prop desks and hedge funds have known about it for decades. A lot of money is passively benchmarked to many popular indices provided by the likes of S&P, DJ, Nasdaq, etc... One reason people invest in funds that track these indices is because they believe the index provider is a good benchmark for whatever its tracking. For example, the S&P 500 tracks the 500 largest US names. The Nasdaq 100 tracks the 100 biggest (mostly tech-related) names that are Nasdaq-listed. etc... In addition to being a good benchmark, a set of rules (here are S&Ps: https://us.spindices.com/documents/methodologies/methodology...) are published by the index provider that govern how stocks are added and removed to the index. Understanding these rules allows arbitrageurs (aka market-makers) to predict when names are moving before they are announced by the index provider. Since a fair amount of capital is already tracking these indices, the passive indexer will be required to buy/sell the names in the index in the right proportion so as to be properly benchmarked.
Another interesting point is that the Volcker Rule has more or less caused a massive shift of this type of strategy away from US investment banks and into hedge funds. I don't have real data on this - just my observations.
Matt Levine stands out amongst journalists and commentators as someone who actually knows what he's writing about because he's been there, done it, and now wears the t-shirt when he's changing the oil on his car.
There should be a bot that posts everything that Matt Levine writes to HN -- It's such great quality writing from someone deeply knowledgeable in finance.
Another good article where he discusses the Goldman software developer who 'stole' company code:
Ehh, not quite. The author is indeed correct about the market-makers providing liquidity to everyone who wants to purchase on the day a company is added to an index.
But saying "index funds free-ride on the work done by active investors" and then following with "no one thinks that active managers should be able to charge for their services, is a world that will spend too little time and effort on allocating capital to the right businesses" is FUD.
The value of the market represents the sum total opinion of everyone in it (plus noise), not just the managers of mutual funds losing business to index funds. Frankly, it sounds like the griping of someone telling fund managers that they deserve their fees, but the supposed loss from using index funds described in the original article (~.2%) is still dwarfed by the increased fees of actively managed funds.
Most index fund expenses are around .1-.2%, while most active funds come in at a whole 1-2%. To justify the cost of an actively managed fund, a manager has to not just beat the market, but trounce it. Very, very few can do so for any length of time, and they know it, which is why articles trying to convince people of the virtue of active fund management are constantly written. Unless your manager is as good as Buffett, buy an index fund.
The math is simple, but there's many fund managers out there trying to convince you otherwise.
Levine does not think you should invest in actively-managed funds.
The little coda about active management makes more sense if you read him religiously, because this is a schtick of his. Passive management helps most investors. But the market as an entity benefits from active management, because active management makes prices more accurate. This despite the fact that for the most part, contributing to the accuracy of prices comes at the expense of the actively-managed funds.
So without active management, the passive funds would perform more poorly, because their prices wouldn't benefit from the corrections of people trading into them to profit from mispricing.
Seconded. This is called the "Grossman-Stiglitz paradox".
There's also a kind of second-order version of market efficiency that says that active fund managers that can actually beat the market will increase their fees until their post-fee returns are the same as everyone else. So even if active _fund managers_ get compensated for making prices more efficient, there's no reason to believe that _fund investors_ will be.
Many years ago I did a comprehensive analysis of Canadian mutual fund returns over about 20 years, and found that the average return per year was dead on the market. The distribuiton of returns was Guassian and had a width of about 1%. I concluded from this that in fact fund managers can beat the market... by precisely amount they pay themselves.
This is evidence for the "second order version of market efficiency" your mention: it was uncanny, and put me into index funds for life (that, and the fact that there was no way of predicting from year-to-year which funds would beat the market the following year.)
I think too many people are erroneously conflating active fund managers with active traders, and my major objection to the article stems from how the author blurs the line between the two. But I admit, I'm not a regular reader of the author.
The market needs active traders for stock prices to accurately reflect investor opinion, but actively managed funds are not the only source of active trading.
I don't read him regularly, but I read it as him downplaying index funds as some sort of arbitrary, socially derived benchmark, which just isn't the case. He glosses over - he surely knows this given his background - all the efficient market theory that created the index funds in the first place.
There is a good reason that index funds are very difficult to beat consistently, and it's not because they are copying all the hard work everyone else does for fees. It's because you don't get paid for specific risk.
I didn't necessarily take him to be saying that. He is absolutely right that a market that is, quite literally, 100% passive would just sit there and not do anything -- it would just grow as money comes in, but there wouldn't be any relative movement of one share against another.
However, we are in no danger of running out of active traders, so there's no need for anyone to run out and sell their indexed investments to save the free market ...
Hmm, well, that wasn't the impression I got. Given the venue and audience, that part of the article felt more like it was giving fund managers the talking points they need to lure in unsavvy investors.
If it were 20+ basis points a year it should show up in the returns and as a failure to track the index. I am not an expert, but that isn't what I see eyeballing a chart of VFIAX over 35 years. It doesn't track perfectly by an amount that does matter, but not .20 basis points a year.
Also by this logic total market funds should outperform other indexes by a healthy amount over time. Also maybe not what we are seeing. Granted total market funds invest in something that is different from what other indexes track.
No, the 20+ basis points is considering how the index itself underperforms, because stock prices get bid up just before they get added to the index and then drop back down as the liquidity crisis settles. The index, not just the etfs, take a hit. Vanguard claims to soften this by trading more deliberately and not buying or selling it all at the opening auction on the day a stock gets added or removed, respectively. So you would expect them to be able to do better than the actual index (until you add in all the other management costs/trading costs).
To an extent the amount the ETFs are effective at this lowers the amount the index underperforms, because they (the etfs) are themselves the driver of the liquidity crisis the index is getting subject to. So you would expect some sort of equilibrium, and the claim is therefore to be taken that this equilibrium settles down at of 20+ basis points.
And yet, the index itself (not even the funds) reliably beats the overwhelming majority of active traders over almost any time window you care to look at.
That is typically due to low fees though, and not something that is inherently due to it being an index.
Hint: The people constructing the index are doing as much of an active choice of stocks as the active traders are. Nothing stops a trader from selecting stocks following the same strategies as how an index is selected.
I'll admit it's looking for a needle in haystack though. In all my looking I've only found these three. Really just VHT and POGRX, since the success of BRK (Warren Buffett's company) is common knowledge.
> Are there any index funds that track the entire market? Wouldn’t investing in all companies solve the problem of index tracking?
Yes, there are funds that try to match the entire market. Vanguard runs a couple "total market" indices (VT and VTI, for example), and there are more.
But that's not to say that they actually invest in the entire market. They don't; they use sampling to try to match the performance of the thing they index. (Vanguard, at least, gets very very close to this...)
As always, Levine is fucking fantastic.