Why haven’t their gains been arbitraged away? Conceptually what they do seems simple enough; and presumably you just need capital to do it. Hell, their own former employees could theoretically compete against them - as could many traders who would pay to learn those strategies.
So why are they still making so much? I don’t understand why their “advantage” hasn’t been arbitrated away into a commodity business.
I worked in High Frequency Trading over a decade ago (not Jane Street) so I have some insight into this. In general, markets are always evolving. I've seen a team of four traders that make $500k/day go down to $0/day over the course of 6 months.
So you're actually right more than you think. Plenty of teams that make money do stop making money over time (or make less). I'm sure there were teams at Jane Street over the years that have folded or dissolved because they stopped making money. But as a firm, they may be making more money.
In the HFT firm I used to work for, that one team went from making $500k/day to $0/day. But another team went from making $30k/day to $2M/day. That's right, a team of 12 people making $2M/day. They probably ate most of the former team's lunch. That's how it goes, survival of the smartest and fastest.
So why hasn't Jane Street's edge been arbitraged away? Some of it has, probably by people within Jane Street who found even more edge. When you have hundreds of the brightest minds in the world, who's going to eat your lunch but yourself (and a few other high quality trading firms)?
But ultimately your assumption is wrong. What they do is not simple at all. It is not like building a simple CRUD application. Throwing money at the problem does not mean you will succeed. You will most likely fail. Strategies that used to work may no longer work when markets change, and markets are always changing. So some traders may try to take their ideas and apply them elsewhere, and they may have some limited success (not to mention the high quality speed and infrastructure you need to win), but when the market changes in 6 months, are you smart enough to keep up?
> They probably ate most of the former team's lunch.
Are you able to elaborate on this?
I’m assuming that teams wouldn’t be sharing strategies with other teams. Is that accurate?
Would these other teams be independently finding some strategy that is either directly better than another internal team or is having some indirect impacts on other teams?
> I’m assuming that teams wouldn’t be sharing strategies with other teams. Is that accurate?
Correct. At least at this firm I was at (different firms have different org setups and therefore incentive structures), each team was a silo. All the teams shared the common core infrastructure to talk to exchanges, but that was it.
Teams have negative incentives to help each other.
> Would these other teams be independently finding some strategy that is either directly better than another internal team or is having some indirect impacts on other teams?
Exactly. You have no idea who you're trading with. It could be against other teams in the company or other players in the market (probably a mix of both).
Imagine someone outside of the tech community thinking along this line...
"Making high performance CPUs that are also highly power efficient should make a ton of money. Why isn't everyone doing it?"
Well, turns out that isn't exactly something that a small group of engineers can whip up in a garage anymore. Same goes for highly efficient market making systems.
CPUs operate due to quantified phenomenon. They're well understood. They've been refined over nearly 100 years.
HFTs came into their own over the past decade or so -- during a time of falling interest rates, unprecedented growth, and notable lack of regulation in financial markets.
One of these things is not like the other. I'd be entirely unsurprised to see most HFTs turn out like Lehman Brothers, Enron, or AIG. They all lasted more than a decade or so. But their gains were fraudulent and they failed spectacularly.
A recounting of the recent history of US financial markets suggests, at least to me, that these firms have the burden of proof. If they haven't proven legitimacy and societal benefit, assuming fraud is a pretty safe bet. I honestly can't name any investment firm with double digit returns YoY for more than a decade or two that doesn't have bodies in the closet. Even Berkshire Hathaway pretty much tracks the S&P500 these days. And as for hand wavy platitudes about price discovery, I don't understand them in the least.
Occam's razor is all I'm saying: What's the simplest answer to the question, why aren't large HFTs with high overheads being eaten alive as technology decentralizes access to trading? Wouldn't we expect types like Burry -- self-driven, confident financial geniuses -- to be equally decentralized? Wouldn't we expect returns to become equally decentralized?
Fraud is the simplest answer. Maybe that comes in the form of market coercion, regulatory capture, negligence, or any other plain old market manipulation. Look back at Enron: The Smartest Guys in the Room. It's all much too similar for my tastes. Time will tell.
> If they haven't proven legitimacy and societal benefit, assuming fraud is a pretty safe bet.
This is absurd reasoning. It's like saying Apple has such large profit margins on their iPhones that they must be either cooking their books or in cahoots with someone somewhere. It's just a phone! How hard is it for a competitor to make a comparable phone?! They've had 15 years to copy them!
> I honestly can't name any investment firm with double digit returns YoY for more than a decade or two that doesn't have bodies in the closet.
It's clear you have literally zero idea what HFT actually does, yet you don't hesitate to call them frauds. HFT firms do not "invest" like traditional investment firms or hedge funds. They provide liquidity and sometimes take liquidity but only tend to hold those positions for seconds or minutes. At the end of every day, most HFT firms have zero position (some might hold some spreads or hedged positions overnight but those are generally less risky).
> why aren't large HFTs with high overheads being eaten alive as technology decentralizes access to trading?
HFT firms don't compete against each other on pure "technology", but more so on mathematical models or what you could call intelligence. Intelligence is not simply arbitraged away over time, although it does happen to some extent. My comment earlier discusses some of this [0]. Technology has little to do with their success. By the same reasoning, why hasn't Apple's margins been eaten over time?
> Fraud is the simplest answer.
The ancient Greeks thought that Zeus was the simplest answer for lightning, but clearly we know that not to be the case.
> Time will tell.
We do not need time. We already know. That you personally don't know doesn't change the fact that nothing illegal or wrong is going on.
As mentioned earlier: phones and financial services are not similar. The market for consumer electronics hasn't fundamentally threatened the stability of the US government every decade for the past 5 decades. I can pick up an iPhone and use its features. It's clearly a complicated device with years of widely publicized, widely understood iterations. Renaissance Technologies Medallion fund may as well be an insider trading scheme for all the public can tell. We have no ability to introspect it. It's a complete black box. And the other financial services provided by Renaissance are garbage.
> They provide liquidity and sometimes take liquidity but only tend to hold those positions for seconds or minutes.
Do HFTs issue or purchase billions of dollars of debt with the express goal of market liquidity? No? So why do you use the term "providing liquidity?" You know "providing liquidity" isn't the same thing as "frenetically purchasing and selling derivatives," right? If these funds didn't exist, I've seen no compelling evidence markets would so much as hiccup. I've seen plenty of compelling evidence that much of financial services in this country are a net drain on its productive capacity.
> We do not need time. We already know.
We'll have to agree to disagree. I think far too much activity happens in financial services, that they aren't nearly boring enough. I harken back to the 70s/80s back before massive deregulation in financial markets, when financiers had to get two jobs just to tread water. I suggest we might revisit such a time. Time will tell!
It's hard to continue this conversation in good faith without pointing out that you literally have no idea what you're talking about. You don't know what HFT firms or what market making actually is, yet you don't mind throwing out outrageous accusations of fraud and illegal activities with zero evidence to back it up.
Your entire argument is "I don't know what they do, and they make money in finance, so they must be doing it illegally." The burden is on you to provide evidence to back up your claim. That's how things work. I could claim that HN user rgifford is the Zodiac Killer, and my evidence is "I don't know him but if you rearrange some subset of the letters of words he's typed in the past, they spell out 'I am the Zodiac Killer'". Whether you realize it or not, your reasoning is that specious at best.
> The market for consumer electronics hasn't fundamentally threatened the stability of the US government every decade for the past 5 decades.
Nothing HFT has ever done has come within even 0.000001% to threatening the stability of the US government. Period.
> Renaissance Technologies Medallion fund may as well be an insider trading scheme for all the public can tell.
Again, claims require evidence. You need to provide that evidence. But more importantly, the fact that you compare RenTech's Medallion Fund to HFT again demonstrates your complete ignorance on the topic. The Medallion Fund isn't powered by high frequency trading. It started decades before HFT ever existed.
> Do HFTs issue or purchase billions of dollars of debt with the express goal of market liquidity? No? So why do you use the term "providing liquidity?" You know "providing liquidity" isn't the same thing as "frenetically purchasing and selling derivatives," right?
You don't know what providing liquidity means. When you place an order to buy or sell 100 shares of MSFT in the market, who is taking the other side of your order? The vast majority of the time, it's an HFT firm.
> If these funds didn't exist, I've seen no compelling evidence markets would so much as hiccup.
I'm sure if you took all of the grease out of your car engine, it would continue to operate without a hiccup.
> I've seen plenty of compelling evidence that much of financial services in this country are a net drain on its productive capacity.
Another false comparison. HFT is not a standard "financial service" that is provided to consumers.
> We'll have to agree to disagree [...] Time will tell!
Flat earthers also ignore reality and disagree, but that doesn't make their perspective equally as valid as those who look at pictures of the earth and conclude it's spherical. Time has already told us, but whether or not you want to look at facts or educate yourself on even the basics of what you think you're talking about is a different question.
Almost every sentence you said in your post contained some flat out false or wrong statements. Your unfounded arrogance is matched only by your ignorance on this topic. You have not made a single concrete or factual point about how HFT might be fraudulent, yet you seem to hold this worldview with the same conviction that the sun rises in the east.
Literally anyone. Literally any other market participant with interest in longterm ownership.
> Your entire argument is "I don't know what they do, and they make money in finance, so they must be doing it illegally." The burden is on you to provide evidence to back up your claim.
That's not my argument. Snakes bite. Highly profitable, speculative, and complicated US financial firms have consistently grown to threaten the stability of US financial systems before collapsing. Precedent matters. When an assertion violates a precedent, it's what must be proven. Your assertion has the burden of proof, and it's an incredibly high one at that.
> ...the fact that you compare RenTech's Medallion Fund to HFT again demonstrates your complete ignorance on the topic
Medallion is one of the most well regarded and profitable hedge funds in modern American history. It's a grandfather to modern HFTs. Even its sterling record is highly suspect.
I sincerely hope the American public doubles short term capital gains taxes and regulates financial markets back to the stone age. Smart people need to spend their time on engineering, medicine, and research. The amount of human capital wasted on silly little financial machinations these days sickens me. I go back and read about financiers jumping from building back in the 20s and I understand completely. Our best and brightest, so confident of what they had to gain, traded their potential in pursuit of Alchemy only to be left with nothing. And who can blame them? Newton fell for the same.
> Literally anyone. Literally any other market participant with interest in longterm ownership.
You don't know what you're talking about. If there were no market makers, your orders probably wouldn't execute within any reasonable time frame.
> That's not my argument. Snakes bite. Highly profitable, speculative, and complicated US financial firms have consistently grown to threaten the stability of US financial systems before collapsing. Precedent matters. When an assertion violates a precedent, it's what must be proven. Your assertion has the burden of proof, and it's an incredibly high one at that.
That is your argument. Snakes bite? So, therefore, what? Financial companies commit fraud as their primary form of operation? What kind of argument is that? It's ludicrous.
> complicated US financial firms have consistently grown to threaten the stability of US financial systems before collapsing
I cannot emphasize this enough. You literally have zero idea what HFT firms actually do day-to-day. Many people use the word literally when they mean figuratively. I actually mean literally in this case. Comparing HFT firms to banks selling mortgage-backed securities is like comparing your local florist to a drug cartel. In fact, it's probably more like comparing a piece of wood to drug cartel. They're entirely different things. HFT firms don't even sell anything.
> Precedent matters. When an assertion violates a precedent, it's what must be proven.
The precedent you name is totally irrelevant. It's like saying since drug cartels do illegal things (they do), then this piece of wood lying on the ground in the forest is also breaking the law by existing. That's how disconnected the two things you're comparing are, but, again, since you literally have no clue what HFT firms do, you don't even understand that.
> Your assertion has the burden of proof, and it's an incredibly high one at that.
Wrong. You can claim that, but like all of your other claims, it's entirely false.
> Medallion is one of the most well regarded and profitable hedge funds in modern American history. It's a grandfather to modern HFTs. Even its sterling record is highly suspect.
I truly don't think you're even reading what I'm writing, or if you do, you keep moving the goalposts (not that it helps your case--you're still 100% wrong regardless of how much further you try to move it). HFT firms are not hedge funds. Stop comparing them to hedge funds, unless you think comparing a piece of a wood to a drug cartel is a useful comparison, in which case there's no point in talking anymore.
Also, "The Medallion fund [...] is a grandfather to modern HFTs" is such a vague and meaningless statement. Sure they have some algorithms, but lots of investment firms do. So what?
But if we follow this absurd line of reasoning, if your grandfather committed a crime (which even in this analogy, they didn't), does that mean that their grandchildren do also? Of course not.
> Smart people need to spend their time on engineering, medicine, and research. The amount of human capital wasted on silly little financial machinations these days sickens me. I go back and read about financiers jumping from building back in the 20s and I understand it completely.
This is a totally reasonable opinion, but entirely divorced from the question of whether or not HFT commits fraud. Again you have provided literally (not figuratively, but literally) zero evidence to back up your claim.
Your entire argument seems to be:
- Some financial firms have done bad stuff in the past
- HFTs are financial firms (sort of, but whatever)
- Therefore HFTs must be committing fraud
That's literally your argument as far as I can parse it. If you read that, that's just so patently absurd that I don't think I can do anything more to highlight how absurd it is. Its absurdity stands on its own. Maybe something like:
How exactly would this fraud work? Most HFT firms only trade their own capital and distribute gains internally, there’s no one to defraud. Also it’s been going on a lot longer than a decade.
Market coercion, regulatory capture, negligence, or any other plain old market manipulation like pump and dump or insider trading or bear raiding, etc.
Enron straight up lied to regulators, many of their employees were also plain negligent. HFTs will probably find their own flavor of fraud given a few more years, if they haven't already.
> HFTs will probably find their own flavor of fraud given a few more years, if they haven't already.
This is what psychics call a "cold reading" - a statement that is bound to be true eventually! At some point in the future HFTs will "find" (?) something approximating fraud. That almost can't not be true. But I don't see how it relates to your statement that Jane Street's reported profits are fraudulent.
Sure, that statement is a bit of a non sequitur. Here's one that isn't:
Highly profitable, speculative, and complicated US financial firms have consistently grown to threaten the stability of US financial systems before collapsing. As such, these types of firms have a high burden of proof for legitimacy. If that hasn't been met, betting on fraudulence is pretty safe given historical context.
Fair. I disagree with "consistent". Some highly profitable, speculative, and complicated US financial firms have done this, but others haven't. E.g. insurance firms.
{x} came into their own over the past decade or so, during a time of falling interest rates, unprecedented growth, and notable lack of regulation in {x's field}.
Context matters. We were speaking in the the context of financial services. In that context, the past decade has been shooting fish in a barrel. You had to be an idiot to lose money with how index funds performed.
Point me to three funds that have maintained greater than 20% YoY profits for more than 20 years. I would be floored if you could do it. Apple, arguably the best and most profitable business in the world, manages between 20-30% YoY profit. They're the largest contributor to world financial markets rather than operating only on derivatives. I can not imagine a world in which the largest trading firms can outperform that without fraud of some kind. In my mind, it's like gravity. Little rocks rotate around bigger rocks.
Devil in the details. MMing is conceptually simplistic but the operational costs are huge and are generally getting worse. Making while fighting these costs against competition playing the same game as you turns non trivial real quick.
That said MMs have mostly consolidated heavily over the last decade (due to many firms collapsing against competitors) so in some ways the business has been commoditized. Not sure if true of MMing ETFs as an authorized participant (JS bread and butter) though, idk much about the logistics there.
It's not just like you can just have a few hundred lines of secret sauce implemented in Python and then go and fill your boots.
The costs of maintaining a trading platform are very high. You have to be colocated with brokers/exchanges, and have a full market data and trading platform optimized down to microseconds. You need a large data and processing farm to backtest your algos. You need the legal structure to dodge all the tax. You need access to lots of GPUs and FPGAs to train your models and execute fast.
At least on the quant side, I think the typical sentiment is that most researchers aren't interested in ops / developing infrastructure / curating datasets.
I think he's saying that a quant can work for a few years and retire unreasonably rich, or keep working at a place where they are well rewarded and everything works.
Or they can take a strategy built on advantageous relationships with banks providing credit for leverage, an accurate and clean history of the markets and prior data to feed models, all run by teams who know what they're doing and who are constantly working to improve the edge the entire firm has, and try to do it all themselves after only really working in one small area.
Like other investment options ... we tend to look at the successful few and assume that they know something special, or have something special. But over time, a few are just lucky and most are not, and we have no way of predicting luck :)
Because they are the ones arbitraging the gains away. That's why there's no opportunity left for you average punter to just start his own little script that take orders from one market and put them on another.
All the players that are left are highly sophisticated technologically, but also in terms of ecological position. For instance you have Citadel doing PFOF with Robinhood. Once you lock in a deal like that, you have a special position in the market. Having access to lower fees is also an important part of the game, and it only happens for players who are already in the game.
Lone wolf traders are satisfied in income in hundreds of thousands of dollars per year, while a large company wants millions or billions. There are way, way more opportunities on the market to make just a couple hundred k than to make a billion. In other words, a lot of the opportunities do not scale well.
Why haven’t their gains been arbitraged away? Conceptually what they do seems simple enough; and presumably you just need capital to do it. Hell, their own former employees could theoretically compete against them - as could many traders who would pay to learn those strategies.
So why are they still making so much? I don’t understand why their “advantage” hasn’t been arbitrated away into a commodity business.