They Still Haven't Told You
arxiv.orgThis is a really poor use of the arXiv.
Here's a discussion of the same phenomenon that provides some explanations that aren't "a shadowy trading firm is propping up prices by painting the tape at open":
https://systematicindividualinvestor.com/2021/01/15/the-magi...
To be honest, the author's strategy could really be happening: some market player (or players) may be aggressively buying up stocks at open and selling them throughout the day at a loss so that the overnight gains positively impact their much larger buy-and-hold tranche of the same stocks. So what? Not only would they be taking on a risk premium by holding that larger slice of stocks overnight, but they're also opening themselves up to massive tail risk. A strategy like this works by taking advantage of the change in order book depth throughout the day to pump up P/Es. P/Es will eventually come back down. When that happens, who knows whether the crash'll start during a trading session or overnight. It reduces into a market timing strategy. This "paper" is ridiculous.
But that's illegal. It's market making.
It is not ridiculous if the data is longitudinally consistent.
Seems to me like an easy explanation is that a whole ton of firms wouldn't want to hold anything overnight because you can't respond to it until the next morning? So they pile in in the morning, and exit in the afternoon.
You clearly didn't read the abstract. There's obviously a conspiracy and if any of this is news to you, it is because the people you trust to alert you to such problems still haven’t told you.
Only Bruce knows the truth!!!
/s
Bruce should make a newsletter so they can sell some branded mugs
Please properly open your sarcasm with a tag before closing it because my AI now thinks everything is a conspiracy.
This would make tons of sense with regard to actual HFT firms. If you're whole schtick is doing stuff over timescales that are (far) shorter than a minute, being locked into a position for hours is risk you really don't want to take.
Exactly, the randomness of overnight would completely erase the consequences of a day full of miniscule trades.
Yeah this is so obvious I would hope it could be controlled for by the Analyst. There are simply a lot of funds and traders who, by policy, do not hold positions overnight.
Shades of Slartibartfast: "... Anyway, the recession came, so we decided to sleep through it."
No, since expected intraday gains are negative, this would be a losing strategy. Closing price seems to be lower than opening price on average.
The suggested explanation is: buy orders are placed before open, raising opening price. The gains on the holdings are then larger than the cost of buying in the opening, and selling during the day.
The market goes down during the day, that isn't the same thing as everyone losing money during the day. On net, all of the market participants lose money, but this is still consistent with sophisticated financial firms making money off of trading with big actors who are less price sensitive and retail traders.
but this article is explaining that unless you are selling short in the morning and covering at night, there are mostly negative returns for intraday trading, i.e. "pil(ing) in in the morning, and exit(ing) in the afternoon".
There are lots of firms and funds and floors that never hold overnight, but this research demonstrates that that is basically a statistically losing strategy. If you follow markets it's almost impossible to not notice that almost all the real action happens after hours, and the day's trading tends to "erase" whatever happened overnight. Bruce's research is hard to contradict, unless the data is wrong or his formulae or off.
Or does it just feel that way because overnight as much is happening in the real world as during the day, but nothing is happening in the stock-market? Then when morning comes a lot happens in the stock-market to sync it with the state-change overnight.
I think the issue with this is that one person's loss is another person's gain. Lots of people lose money during the day, but it's not generally the very sophisticated hedge funds and HFT people.
And if you were to hold overnight, you'd have to sell in the morning, and the volume of overnight trading isn't high enough to support a bunch of big firms buying into overnight trading and selling at open, they'd all sell into a disaster, which would make overnight returns go away?
This makes sense to me as given a constant rate if information per hour from world markets and the world in general, there will be more information outside of trading hours than within them.
I'm not sure there is "more information" about the value of a stock. There is only newer information.
The new information during the night can tell us "Hey the price should be higher" but then more information coming in later can cancel the previous good news.
Am I blind or does this paper spend a huge amount of time lamenting the failures to notice the issue, without ever once actually describing the issue. I don't claim to be very knowledgable here, so can someone fill in the gaps for those of us who want to know exactly why Fig 2. is so damning?
"They still haven't told you and I'm not going to either" should have been the correct name for the paper.
I don't think there is anything new here. The overnight trading anomaly has been observed for years
I work on wall st and, trust me, even the greenest traders know this. I think the author is trying overly hard to be dramatic in order to achieve his PhD certificate.
I don't think that these papers will earn the author a PhD, and I also don't think that this is a goal. Looking at the references, the author seems to like to publish such analyses on the arXiv, which is arguably not really what it is meant for; but still, it's a pre-print server so who cares.
I don't understand: if this is a persistent effect why can't you short in the morning and cover in the afternoon?
The author explains what he thinks is happening on page two of this paper (which he cites in the OP, but doesn't actually explain): https://arxiv.org/pdf/1912.01708.pdf
TL;DR: Stock prices in the US go up overnight and come down during the trading session. The only explanation must be that some shadowy trading firm with a lot of money is buying a bunch of $stock in the morning and selling it back later in the day (at a loss), because it inflates the value of their much larger stockpile of $stock that they hold onto overnight.
If they sell later in the day that should bring the price down. How does that increase the value of the stock overnight?
The paper asserts that, if you buy an amount of stock in the morning, that moves the price up more than selling the same amount of stock in the afternoon moves the price down.
To back up this assertion, the paper cites one of the author's previous publications from 2018, which in turn cites this WSJ article: https://www.wsj.com/articles/early-birds-suffer-in-market-14...
Thanks for the explanation. I looked at the WSJ link and it seems to be saying that volatility is higher in the morning than in the afternoon. That makes sense since in the morning it is less certain how the markets will go today whereas before the closing bell we have seen the market behave for 6 hours or so.
And if you trade when volatility is high you could think it has a bigger effect on the market than if you trade when volatility is low, because in a volatile environment other traders are more likely to react abruptly to any change.
But this only works if your trades are big enough to be noticed by other traders.
It all makes sense to me now. And it makes sense that big traders who think this is the case would try to take advantage of it (I don't think it can be made illegal, or is it already?).
Therefore we can assume that prices are artificially high in the morning, and so it makes sense for us small guys to sell in the morning, buy before the close -- if you are a small investor. The big guys will of course do the opposite, to accomplish their market-manipulation.
I thought it was a volume thing? So overnight there are less shares to buy so they fetch higher prices. That then normalizes over the course of the day as more shares become available?
Clickbait paper titles on arxiv, good grief. Out of curiously I looked the guy up -- he must have gotten access via some technical publications ~a decade ago. There ought to be away to cut off access to people who are no longer publishing in their field of expertise.
I thought anyone can publish to arxiv?
no. from what I know, you need two endorsements or have a certain academic credentials.
> Fortunately, the other sentences in footnote 78 contain no words that start with v, so we should be able to take them at face volume.
Savage.
I saw lots of charts & graphs & flashy wordsmithing, but I didn't actually see any evidence or examples of firms doing unscrupulous trades.
I'm not an expert, but I know better than to dish it out better than I can take it. My opinion is that these "exemplary" market returns are simply the result of markets being open only part of the day: between 0930h and 1600h there's liquidity to buy/sell your position at any time, for the prevailing price. Markets are open only 7h of the day but 24h worth of events takes place each day.
The other elephant in the room is that all market participants know the trading hours. Much news, releases, events, etc. happen outside of the liquid trading hours, resulting in discrete jumps between the close of one day and the open of another.
These are also cumulative returns over a huge timespan: everybody knows the fed can crash the markets mid-day with the wrong jawboning. the reverse price effect can also be true, resulting in huge open-to-close changes.
> between 0930h and 1600h there's liquidity to buy/sell your position at any time, for the prevailing price. Markets are open only 7h of the day but 24h worth of events takes place each day.
Yes indeed. The author claims that there is less risk in overnight positions than in intra-day positions. I think there's more. Firstly more time passes overnight and secondly the lack of liquidity means you can't unwind overnight positions if you need to.
The author's earlier paper explains more clearly exactly what the firms in question are doing and how they profit from it: https://arxiv.org/abs/1811.04994
The first page suffices to get the idea.
But spreads aren't larger in the morning are they? If anything they are smaller, due to the action of the opening auction. Without actual data to back up that assertion, the rest of it doesn't really need reading.
Assuming his assertion is correct, the reverse is also a strategy. Selling in the morning, causing prices to drop, then buying back when they are cheap in the afternoon, so ending flat but having made money.
So anyone that buys and then sells or sells and then buys makes money. This is easy! What could possibly go wrong?
My knowledge of this is maybe limited, i've not worked as a quant, but i've stared at a fair bit of market data having written feed handlers for a fund.
Expand doesn't necessarily mean buys. It's a leveraged market neutral portfolio, so most likely built from a combination of the future and hedged options books. So would be a combination of buys and sells of derivative products. Even if the net position is larger.
> Figure 2 shows plots of overnight and intraday returns for twenty-one major stock market indices around the world. Turn the page and compare Figure 1 with Figure 2. See if you can tell a difference.
These images... do not render well on my machine, to put it lightly. So they look remarkably similar to me. Perhaps the author could spell out what this difference is? There is eventually mention of "striking similarity in the overnight and in- traday return patterns in the indices around the globe," but I don't think I'm ready to conclude that strong correlation of phenomena across markets in a global economy must be caused by a collection of manipulators acting on all of those markets.
I'm curious to see what others have to say, since I lack the hardware and background to properly read this document.
important news is often released b4 market open or after close, or pundits will hype the stock over the close.
Manipulation, such as gapping the price higher or lower to make profit from options or increased liquidity of regular trading hours. So you spend $10 million in the pre-market hours to make a stock open 5% higher and then use the extra liquidity to unload a $100 million position at the open while also selling calls.
Tangential question: why isn't the stock market open 24/7/365 (minus periodic maintenance of the computer systems)?
Because once upon a time people had to go in person to a building to trade and no one wanted to do that 24/7.
Now, we still maintain that for some markets. I'm guessing its a combination of laws, inertia, and the ability to do outside of RTH news/sys updates.
Some markets are open much more than the 9-430 stock market - for example futures markets open sunday night.
probably because there are still 2 days needed for settlement
crypto markets don't have this issue and are open 24/7
Because professionals don’t want to monitor stock prices 24/7, and because the end-of-day price is significant a lot of things.
A lot of trading takes place in the last 30 minutes of the day for that reason.
> Because professionals don’t want to monitor stock prices 24/7
I'm sure you could pay professionals to work in shifts to monitor stock prices 24/7.
> because the end-of-day price is significant a lot of things.
Isn't that circular reasoning? "The stock market needs to close during the day because the end-of-day price is significant to a lot of things because the stock market needs to close during the day". Take the spot price at 00:00:00 AM as your end-of-day price and call it a day.
>> Because professionals don’t want to monitor stock prices 24/7
> I'm sure you could pay professionals to work in shifts to monitor stock prices 24/7.
I guess you'd have to pay somebody extra to monitor those prices overnight. And the companies themselves by-and-large are doing less in the middle of the night in their local time. I mean, I know it is pretty detached from reality, but the stock at least tries to pretend that it is in some way grounded in the fundamentals of the companies involved, right? No news, no new info to make trades on.
Two words: shift work
Another two words: algorithmic trading
FTX has tokenized stocks that trade 24-7
This is such a horribly clickbait title. Please edit or remove it
This title is peak click bait. I don’t think you could’ve made it more clickbaity if you tried.
Lots of condescension here, but supposing that overnight returns are in fact on average substantially greater than intraday returns, what is the layman-friendly, non-conspiracy-theory explanation of this phenomenon?
There are so many questions waiting to be answered. He's plotted intraday versus overnight returns and showed that the former are larger. He claims this is evidence of market manipulation because overnight positions should be less risky. So demonstrate that by plotting the volatilities! There's no mention of observed volatility in either paper. I'd be willing to bet the overnight vols are correspondingly higher.
An interesting result -- but not worth the hot air.
> one or more large, long-lived quant firms tending to expand its portfolio early in the day (when its trading moves prices more) and contract its portfolio later in the day (when its trading moves prices less), losing money on its daily round-trip trades to create mark-to-market gains on its large existing book.
Renaissance Technologies' Medallion Fund?
Simons is a genius.
I am surprised that by now, decades later, no one has the goods on Renaissance . What is to stop someone who works there or former employee from uploading to the dark web the "Renaissance strategy", for a price tag of $10-100 million btc, monero or something. Who would know. Although no one would beleive him.
Well one of the big factors was tax avoidance. Basically mislabel your long term strategy as short term for a lower tax bill. Then just wait, and know that the final penalty fine (while still large) will still be less than the total you made over time by not paying the tax.
https://www.reuters.com/business/finance/renaissance-executi...
You don't geta 60% cagr with tax loopholes. all hedge funds try to minimize their taxes by whatever means possible.
One good reason is that the employee retirement fund is the Medallion fund, so employees profit from the firms continued success.
Here's a web version if you don't want to read a PDF: https://www.arxiv-vanity.com/papers/2201.00223/
So he is saying large firms use money to pump up prices early morning to promote FOMO and chaos and they would trade the predictable chaos and even after they sell their initial pump, they still make money?
No.
He is saying that some firm own lots of $stock that they buy-and-hold. They then buy smaller amount of $stock early in morning to cause a swing up in price. Over course of day, the ability to influence price declines, so they can sell the amount they just bought without influencing price as much. Sell for profit or loss, doesn't matter.
The root goal (how they make money) is that they should have influenced the price enough that the large buy-and-hold stock they own has increased in value. Specifically, they want the "overnight" price change to be more positive than the decline across the day (again, by pumping up the morning price). They don't have to buy/sell, its more the value of their holdings are higher.
People tend to point out how much energy cryptos consume, is there an estimate for how much high frequency trading consumes world wide?
People have certainly tried, but I think that these comparisons don't make any sense. What makes the "waste" of crypto stand out is a qualitative concern:
HFT uses energy as a means to an end (i.e. computation), while proof-of-work mining has an incentive structure that directly rewards energy expenditure. In other words, one is energy- (and hardware-)bound, the other isn't.
As a thought experiment: If fusion energy and self-replicating nanobots were to become viable tomorrow, how would that impact HFTs and proof-of-work mining, respectively?
You left it unsaid but wow, PoW is a horrifying grey-goo type scenario that ends in dyson spheres and intersellar war. I had never explicitly thought that through before.
HFT firms use a tiny fraction of the energy of proof of work crypto schemes. Even if you take into account energy requirements they used to build their high speed networks.
The main reason being is that they are all co-located with the exchange, and the byzantine fault tolerance is completely side stepped by the exchange having a single point of serialization between the HFT participants and the matching engine. And then just ensuring everyone has the same length cables to that single point of serialization.
So the policy of everyone connects through this single point, solves the fairness problem.
And then the fault tolerance is often just solved by having an active standby architecture, so if the primary matching engine goes down, you fail over to the backup.
You would be surprised at the low number of servers actually in the primary path for financial exchanges.
Well for the actual trading they will be operating out of the same datacentres as the exchanges which are not particularly massive and will have relatively small limits on the heat (and therefore power). There will be other computers in other bigger cheaper datacentres for analysis and suchlike. And office buildings of course. But not really comparable to e.g. the energy usage of Argentina. I would guess the total is less than one of the massive internet companies like Google or fb but that might be a bit low.
Not sure, but I understand crypto uses > 0.5% of global electricity production. So you can pull up any chart that accounts for the 99.5% of top electricity users and verify HFT is not one of them if that's what you're seeking to show.
Honestly? I blame the reptilians for this.
Can someone explain for a total layman?
No, I don't think that can be done. No insult to the layman, whatsoever.
TL;DR, a lot of hedge funds don't hold positions overnight.
This paper reads more like some click bait article on BuzzFeed or Business Insider than an academic piece.
it reads like the sort of thing they would publish
Does this mean I should sell more often in the morning than in the afternoon?
What are the implications if this is true?