> If there are n people in the market moving money around in a closed system then the combined wealth of those n people is constant.
> Admittedly this relies on money not being created or destroyed, which may cause the model to be a poor approximation of reality.
What investment markets provide is a way for firms to increase their productivity. Think of the farmer who has a bucket, a stream, and an acre of land. The economic pie gets a lot bigger if the farmer can secure an irrigation system.
The farmer can solicit the banker in town for a loan, or he can enter a debt offering into large market or exchange. Investors are more likely to invest if they know that they can exit their investment freely. i.e. there is liquidity, so they can sell their investment as their own needs dictate.
You can argue that getting equity pricing right helps companies with their access to capital, but once you have the pricing right at 100ms I don't understand what value HFT firms are adding by pouring money and talent into getting the pricing right at 10ns.
It seems like such an obvious win for society to mitigate the winner-take-all incentive of being first to market on a pricing disparity.
That's slippery slope logic. If the pricing is right at 1s, why pour money and talent into getting it right at 100ms?
What's the win to society to "mitigate" an "incentive"? Is the problem volatility? Other forces create huge volatility. Should we penalize anything that creates volatility? Maybe we should end all program trading?
Meanwhile, you're effectively vouching for a comment that models the markets as a closed system of people dividing up a single pot of money. Isn't it plain that such a model is wrongheaded?
What's the win to society to "mitigate" an "incentive"?
The main incentive is reducing time and energy devoted to a zero sum game.
A hypothetical: imagine a sunken pirate ship is discovered. Now suppose 10 crews of divers get into a race to retrieve the pirate gold. It's useful to society to bring up the gold. It might be useful to society to have a race between 2 crews to bring up the gold, to make sure the first crew doesn't dilly dally. The gain to society is $GOLD - 2 x $DIVER_COST, or perhaps $GOLD_AFTER_LONG_DELAY - 1 x $DIVER_COST. On the other hand, having 10 crews of divers all competing for the gold is pointless - the gain to society is $GOLD - 10 x $DIVER_COST, which is 8 x $DIVER_COST less than if 2 diver crews chased the gold.
HFT is basically the same situation as the race for pirate gold - a lot of smart people in a race to create a fixed amount of alpha. We might be better off if they were creating new alpha elsewhere instead of all simultaneously chasing after the same alpha.
(That's not to say I'm advocating a ban on HFT on this ground. A certain amount of effort devoted to HFT is certainly a good thing, and I doubt the government would get things right. I just don't think the market is getting things perfect either.)
So this argument makes a lot of sense, but you see that it's not the argument that's being employed against HFT in general, right?
What I see are a lot of people arguing that the HFTs are getting an unfair edge on other traders, as if some main street stock picker was actually in competition with an HFT prop trading shop.
My sense of it is that many of the people making this arguments believe that were it not for HFT's, people would have frictionless access to a real efficient price for any instrument they wanted to buy, when in fact they'd just be dealing with a much clunkier and less reliable set of middlemen.
I know this is an uncommon argument against HFT - I've only heard Tyler Cowen pushing this argument, but it's the one I find most plausible.
As I said, a certain amount of HFT is a good thing. If I thought it was harmful, I'd quit my job as an HFT programmer and find something else [1]. I'm just pointing out that there are costs, which don't necessarily outweigh the benefits after a certain point.
[1] This was a major reason why I quit my job as a postdoc, rather than trying to become a professor. I believe college is mostly rent seeking and I don't feel it's right to participate in that.
So you left academia to become an HFT programmer because you thought academia was too unproductive in the context of larger society? That's one hell of a scathing critique of academia and, er, um, I can't help but wonder if the money might have had something to do with it?
Actually, the money wasn't a major factor. It's piling up, but I literally have no idea what to spend it on.
Here are my opinions from when I was an academic (I've only been working in HFT since March 2010, full time since May 2010):
My opinion from 347 days ago: universities have vastly more problems than that. They are huge bloated organizations structured around funneling money to employees (from both students and the government) rather than educating students. The main reason people still go is for status signaling purposes, otherwise they would have been replaced long ago.
From 406 days ago: a $10 million grant; my university will take about $5 million off the top in "overhead" (to be spent on overpaid administrators, student stress counselors, the latino student center, and maybe even education).
(This was roughly the period when I decided to leave academia.)
From 469 days ago: The job description of "professor" is certainly a strange beast: teacher/scientist. It makes about as much sense as actor/programmer...The perverse incentives this creates are massive. Universities hire scientists rather than teachers in order to get their hands on half the scientist's grants. Scientists waste their time masquerading as teachers... This is harmful both to science (I'm not doing research in class) and students...Actual teachers are squeezed out, since there is no room for them.
From 999 days ago: Another part of the problem is that there is no incentive for cost control in the university...I'm currently teaching a "Quantitative Reasoning" class right now. Basically, take Weeks 1-2 of Prob&Stat and expand it to fill a whole semester (half a semester, due to poor planning and miscommunication). Some of this is my fault, some of it not...Plus, my students are all art/history/literature majors, and just don't need it. Everyone in the room would be better off keeping their $4,000 and not sitting through my class.
As a fellow HFT programmer I sort of think of HFT as another asset class: eventually (as more divers go for the gold) the returns will be on par with alternative asset classes. This is the beauty of capital investment and the free market--if an industry produces outsized returns, then capital/resources/effort will be poured into that industry until the returns diminish.
That's slippery slope logic. If the pricing is right at
1s, why pour money and talent into getting it right at
100ms?
I'm saying there are negative externalities[1], which are known to cause market inefficiency. Negative externalities typically look like "slippery slope" arguments, when in reality there is an appropriate level of penalty/tax/etc that restores efficiency.
Meanwhile, you're effectively vouching for a comment
that models the markets as a closed system of people
dividing up a single pot of money. Isn't it plain that
such a model is wrongheaded?
I'll concede it's not zero-sum, but surely you also will concede diminishing returns to liquidity. I'll also concede that my original "money sink" comment was hyperbolic & inflammatory.
Actually, trading can and does increase the total value of the system. If I own $100 of Intel stock and you own $100 of AMD stock, exchanging half of our stock (so that we each own $50 of Intel and $50 of AMD) leaves us both better off, since we each have less risk.
I'll agree that 100ms to 10ms difference probably isn't a social good, but if a certain amount of money is due to a group of people for their role in providing liquidity and they choose to collectively wast a fraction of it on cutthroat zero-sum competition with each other I'll say thats sort of a shame, but I don't really think its anything I need to worry about.
Your argument is silly for the following reason. Imagine that a firm was able to profit from discrepancies between values of two stocks that last 100 microseconds. Say this firm makes 1 billion dollars a year - it pays 350 million in corporate taxes to society, which is one direct way to measure the value it is adding. Looking at it differently, what value does a pizza cutter provide to society? why not ban pizza cutter production? Why can't people cut pizza with a knife? instead of having all those smart engineers work on making pizza cutters, let them make something worthwhile. Follow this line of reasoning to its logical conclusion and you end up with Soviet Union circa 1975 - i.e. system that does not work
If there are n people in the market moving money around in a closed system then the combined wealth of those n people is constant.
If another person joins in and is making a net profit then it must be the case that the combined wealth of the original n is decreasing.
Admittedly this relies on money not being created or destroyed, which may cause the model to be a poor approximation of reality.