Assorted links

1. The first robot?, and are real-life Transformers on the way?

2. There is no great stagnation: shelf life of the Twinkie extended from 26 to 45 days.  One email correspondent suggested to me that the de facto shelf life of Twinkies is already two to three years.

3. Which are the most abandoned books?

4. Good review of a good book on drones.

5. New and important critique of high-frequency trading (pdf), though I am not persuaded by their claim that batch auctions will lower the bid-ask spread.

6. New survey paper on trust and economic growth.


Regarding "high frequency" trading...why do we even need any kind of intraday trading? Would anything be lost if we could only trade once a day?

Intraday trading substantially lowers the cost to trade for uninformed investors (e.g. retail) relative to informed investors (e.g. insiders). Market makers use sophisticated strategies to segment out uninformed flow from informed flow. During periods of uninformed flow market makers tighten spreads and increase offer sizes, effectively decreasing transaction costs. And vice versa for periods of informed flow.

For example let's say some sort of news is released about a security at 14:00, and we see substantial trading between then and 14:30. Spreads will widen and price impacts will increase during the elevated trading period. Uninformed traders are as likely to trade at any time of the day, but informed traders are much more likely to have traded immediately following post-news. This means that averaged across the day informed traders will have paid an average higher transaction cost than uninformed traders. In contrast imagine if we only did one auction at 16:00. Both informed and uninformed trading that day will pay the news-premium on their transaction costs, because market makers have no way to segment the flow since it's all coming in at the same time.

This is a very simple example, but there's dozens of ways that market makers segment trade flow that simply doesn't work unless you have continuous intraday trading. Eliminating intraday trading is unequivocally bad for uninformed traders. It probably is even bad for informed traders. Since uninformed demand for liquidity has high price elasticity a small increase in transaction costs drives a lot of uninformed flow away from the market. This raises the average adverse selection that market makers face and raise the spreads that the entire market pay. So even though it lowers the gap informed pays relative to uninformed, driving uninformed trading away raises the costs for everyone. Very frequently the latter outweighs the former.

Not sure I followed all that, but what about the uninformed trader who happens to trade between 14:00 and 14:30? Looks like they pay the higher cost as well as suffering the disadvantage of being uninformed. Are their average costs truly lowered? Besides, you can say that being uninformed is their fault, and that makes sense when when we are talking about reasonable time intervals, even maybe minutes. But milliseconds?

I myself am unconvinced that HFT provides much if anything in the way of social benefit.

I'm not sure if I conveyed the logic properly, so let's consider the example a little more concretely. There are 13 half hour periods between 9:30 and 16:00. Uninformed traders by definition don't respond to news so they are as likely to trade in any of the intervals. Randomly 1/13 uninformed trades will occur between 14:00-14:30. In contrast informed traders will respond to news since its information that updates their models. So in the hypothetical it'd be reasonable to conclude that half of the day's informed trading occurs in the interval. Just to use round numbers assume informed trading makes up half the day's volume and that the average cost per trade is $0.01 in non-news periods.

Market makers set spreads based on the informational content of the flow, i.e. adverse selection. Inside the interval the ratio of informed to uninformed flow increases by a factor of 13, meaning the transaction costs increase to $0.13 per trade in the interval. Since half the informed trading takes place in the informed period the volume weighted average cost for the informed traders is $0.07. In contrast the day's volume weighted average cost for the uninformed traders is $0.0192 per trade.

If you move to a daily auction system market makers need to still be compensated for the same adverse selection on the same overall order flow. Meaning the transaction costs will stay the same. Now both informed and uninformed traders will both pay an auction cost of $0.0446 per trade. As you can see this substantially raises the cost of trading for the uninformed. Does every uninformed trade see a lower price? No, some will randomly pay higher costs, but stochastically flow segmentation heavily lowers the average cost of trading for the uninformed.

"we are talking about reasonable time intervals, even maybe minutes. But milliseconds?"

I intentionally picked a very simple example but the logic extends to very small time periods. Consider the period immediately after a large buy trade hits the market. Other traders will use sophisticated models to try to follow-behind market moving trades. Such a follow-on trader is substantially more likely to trade in the millisecond or two following a big trade. An uninformed trader is definitely not going to be timing his trades to coincide with the milliseconds after a big trade. The market maker faces substantially higher risk of being picked off by a sophisticated algorithm in that millisecond.

Naturally she will respond by widening her spread and offering less liquidity in that post-trade millisecond. In a few milliseconds she'll increase her liquidity and tighten spreads as the risk of informed trading decreases. The example works exactly the same as the original, but it's in the range milliseconds instead of hours. The segmentation of flow down to the millisecond works to lower the average cost to uninformed traders by raising the average cost to the informed algorithms.

You may think such an example sounds contorted and esoteric. It's a simplified but highly typical scenario that HFT strategists work with everyday. The ability to segment flows with such high precision is what enables market makers to offer the lowest financial transaction costs in human history. If you bluntly limit the ability of traders to act in millisecond time frames I guarantee that you will substantially raise the cost of trading and decrease market liquidity.

Thanks for the clarification. I see what you are getting at.

This seems to assume that the market-maker is also uninformed, as her decisions are based on the order flow coming in. I suppose that's reasonable in a very short time frame. I also suppose they have to adjust their prices in milliseconds, so there seems to be an arms race here.

I wonder about the assumption that uninformed trades occur uniformly over the day. Even an uninformed trader may see a price spike and react, one way or the other. Nor is it clear to me why spreads will widen, rather than both bid and ask prices moving up (in the case of a rush of buy orders).

Comments for this post are closed

I intentionally simplified the examples to convey the principles at play. Your points are quite accurate in regards to how the real world deviates from my spherical cows. In reality there isn't two separate sets of market makers and informed traders. Most HFT traders crucially rely on the combination of providing liquidity, predicting short term price changes and having access to low-latency infrastructure. Knock any of these three legs out and they'd fail to cover their own transaction costs. It's hard to disentangle and attribute overall HFT profit to each component.

A typical HFT strategy will make markets in some security using some model to determine when to provide liquidity aggressively. It will use another model that's making short-term prediction about price changes, and try to manage its inventory to be on the right side. If its bullish it will probably increase the size or price of its bid and fade its ask. If the signal's large enough it might even cross the spread and take liquidity. Market makers can create adverse selection for each other even if they don't trade against each other. Rival market makers are going to quote a side less aggressively when they think price will move in that direction. If other market makers have information that you don't, your quotes are going to be less crowded on the wrong side. More of your orders will fill that way, and your inventory will naturally build up on the wrong side of the information. Even if the flow crossing your quotes is itself uninformed.

It's important to note that there's no such thing as a perfectly informed trader. Every firm has different models and focuses on different aspects. Even the smallest firm still has mutual information against the most sophisticated firms. Everyone still worries about informed trading and adverse selection even if they themselves are highly informed. One way to deal with this is by making sure you can react faster than your HFT competitors, being able to place your orders faster than you can gives you an advantage to the extent that your models and information overlaps. Like you said this does create an arms race, and with it positional externalities. But there's no silver bullet that easily fixes this without imposing high costs in some other way.

Comments for this post are closed

Comments for this post are closed

Comments for this post are closed

So, the fees in HFT for trading the same embodied 30 year old labor have value because it is critical for investors in 30 year old embodied labor to get the absolute best price for that 30 year old effort instantly rather than averaging the price for the long ago paid labor over a decade of investing?

Or is trading shares a derivative of the hourly trade profit of Wal-Mart and Amazon and the HFT are needed to tune the market price to the instantaneous price of the revenue generation of the brand which is completely independent of the labor cost embodied in the tangible and intangible assets of these firms the stock is derived from?

Do you think Warren Buffett, or a Michael Dell doing a cash buyout, need to shave a penny of the share price to get the best deal for their buying of embodied labor which is generating a return on the capital value of the embodied labor?

If Warren Buffett who is buying a hundred million or more shares of a company is not concerned with a penny shaved off the share price, why should an individual investor buying a hundred shares?

I'm sorry I don't think I understand some of your terminology...

My point is that we want the lowest financial transaction costs possible. One because why spend money on something if you don't have to. Even if its a half a penny here, a tenth of a penny there, the volume of financial transactions is so huge that it easily adds up to billions. Electronic trading has lowered financial transaction costs by over 95% in the past two decades. That's an aggregate saving of tens of billions of dollars.

It enables strategies and structures that were previously impossible under historically high transaction costs. Reducing trading friction has made financial markets far more efficient, which in turns makes decreases the noise in the economic allocation of capital and savings. So far I think electronic trading represents one of the greatest economic and technological accomplishments of the 21st century.

Comments for this post are closed

What's "30 year old embodied labor"?

The only value anything has is future value - either as a hedge [to be traded off eventually], for direct consumption, or to be used to produce [perhaps through N layers of indirection] a directly consumed good.

Stocks certainly don't "embody labor" more meaningfully than any good. (Oh, we can say "all goods embody labor", quite truthfully, but they also embody land and material costs, and in some sense intellectual property. This means nothing at all in terms of critique, however. The labor and land rents and materials are all already paid for, once the object - or share - is sold.) Likewise, talk about "labor cost embodied in the ... assets of these firms" is irrelevant - stock prices model demand for shares and expected future revenue and value of firms. They simply do not and cannot model "labor cost" aggregated over the assets.

The idea that labor cost is primary seems to be a core misunderstanding here.

(See any Austrian for more on that - labor gets paid out of prices, rather than determining them. Same for land rents and commodities. Sale price of the finished good provides the pool from which the constituents get remunerated, rather than being the sum of the 'costs', at least at the macro level.

From the level of someone running a business, yes, the it looks like it works the other way, since market prices for all the above have been set by the previously mentioned process already. But the two are radically different.)

Also, don't forget - as HFT Trader seems to be trying to say - that the arbitrage opportunities HFTs allow, like all arbitrage transactions, quicker and better settling of the real market price.

Lastly, neither "Warren Buffet buying a huge chunk of a single company" nor "Joe Trader buying a hundred shares very infrequently" actually model most of the stock market ... nor is there any obvious reason why we should try to make that so, is there?

Is there anything here other than suspicion of "Finance" as such

Comments for this post are closed

Comments for this post are closed

Comments for this post are closed

If we traded once a day, or if we went with the idea of batch auctions, we would see substantially reduced liquidity and substantially higher volatility for two reasons

1) much higher uncertainty about where the market actually is would raise market makers' risk and reduce their desire to provide liquidity, all things being equal

2) if traders compete on price rather than time as the paper suggests, it would result in more prices being traded within a short amount of time - the very definition of increased volatility and again would raise market makers risk

Comments for this post are closed

Comments for this post are closed

The downside to restricting trading to once a day is that news changes many times over the course of the day, and so do fundamental expectations about value. Earnings reports, earnings rumors, newswire stories, all these things and more give people legitimate reasons to want to buy or sell a share of stock.

Comments for this post are closed

The "Transformers" pop culture reference is completely wrong. What you mean to ask is "are real-life mecha on the way?" Or "real-life Gundam Mobile Suits" (the pop culture reference actually used in the article).

Transformers are robots, not typically or necessarily operated by a human wearer, that can take on two or more useful shapes (e.g. large humanoid robot or wheeled car).

I fought the urge to write this post. But I'm happy someone didn't.

Also, some sort of reference to the Caterpillar P-5000 Power Loader seems in order...

Comments for this post are closed

Thank Primus for this.

What if I forwarded this link to someone else? I would have looked like a slagging idiot.

Comments for this post are closed

Further, that steam-pigeon is an automaton, but not a robot.

Robot being, etymologically and practically, an automated worker.

Comments for this post are closed

Comments for this post are closed

4. So, killing as a substitute for capture. So much for torture "working."

And to you out there who criticize people like me for criticizing this clown because we don't like his inconsequential healthcare boondoggle because we criticize him for exactly the same reasons we criticized the last clown, please, blow me.

Andrew' your recent fascination with fellatio-related posts is concerning. This is a family blog- kindly please get out more to fulfill your needs so you can bring your A game to the comment section.

No sex for us please, we're economists.

Comments for this post are closed

Comments for this post are closed

The alternative to "capture and interrogate" was never "let them live in peace and hope the TSA can stop 'em."

Who said the TSA shouldn't be abolished to?

"Obama parted company with Bush by abandoning hope of turning former state sponsors of terrorism into reliable American allies. Subsequent events have resoundingly confirmed the wisdom of restricting counterterrorism efforts to non-state actors. For instance, the alarming flow of weapons from Gaddafi’s arsenals into Mali and Syria has reminded US officials that disorderly regime change can risk proliferation. The collapse of a dictator in regions where smuggling is a highly developed art is bound to spill dangerous arms into the clandestine arms market, where they can be purchased at bargain-basement prices. Luckily for the nuclear-terrorism-obsessed neocons, Saddam Hussein did not have the weapons they had cited to justify their campaign to topple him."

Comments for this post are closed


This simply means that Obama is obviously worse than Bush. He's willing to kill kids solely to protect CIA criminals and make his legal headaches such as his renegged promise to close Guantanamo go away.

Are you reporting for blowing?

I suppose I agree that he is doing something less moral because it is more politically palatable. He's not been very public regarding what caused him to change his mind on the magnitude of the threat. But he sort of painted himself into a corner. His campaign was all about how this terrorism stuff was being blown out of proportion, and now we've got policy that's not that different from what Bush was doing.

For what it's worth, I don't think doing nothing was or is a realistic option. And frankly I much prefer an answer based on SEAL teams and laser guided bombs to one based on homeland security.

False dichotomy. Homeland security is just a total waste of money if we are lucky.

Comments for this post are closed

The idea that we could do this with, basically, good policing, is not a false dichotomy that I invented.

> Homeland security is just a total waste of money if we are lucky.

I'm not convinced of that. I'm not exactly convinced it's wrong, either. But it's important to keep in mind that the threat we face is not the 1979 KGB - they can't just brainpower and resource their way past all obstacles. We're a rich nation: our spending 100x dollars to make operations cost 10x for a poor enemy is not obviously wrong.

But yeah, offense seems to be a more effective strategy than defense in this particular conflict.

Comments for this post are closed

We never needed the torture if we can now just blow them away. If we don't need the torture, then we don't need to blow them away. If they are low-level, they are not actual threats.

They aren't coming over here, they never were, and if they do now it will only be because we slaughtered their kids.

Comments for this post are closed

It's just imperialism. But a much worse variety than Bush's. At least Bush was an idiot who believed his nonsense.

Comments for this post are closed

The correct model is the ZD30 dramatization version of getting bin Laden, minus the torture. I never understood why people thought the torture was critical. Even in the movie they simply "flipped states' witnesses." You can do this by threatening detention. Can't do it with drones.

Again, the facts of the drone program PROVES we don't need it. It's a fraud. There is noone in the hills of Pakistan who is a threat to the American people at home.

Comments for this post are closed

You seem semi-reasonable on this. Remember when people argued with me saying they weren't using Hellfire missiles?

I'm owed multiple blow jobs.

Comments for this post are closed

I think your reasoning is a bit circular.

There are a number of possibilities that aren't mutually exclusive:
1) The organization that was al Qaeda is just smashed; it doesn't really exist anymore as a significant threat, so we can declare victory and go to a lower level of operation. The threat is now mostly confined to radical Islam in other forms.
2) We're just running bigger risks like it's 1997 again. We lack the intelligence we used to get when we captured and interrogated, so now we let threats go that in 2007 we would have dealt with. Eventually this bites us.
3) We just shoot if there's any question now, so we engage more readily and accept that a higher fraction of the time we're blowing up uninvolved people because we don't have the intelligence to make the distinctions we used to. I know they say this isn't what's occurring, but they aren't particularly convincing.

2 and 3 are likely exacerbated by the tactic of using missiles rather than commandos. Without people on the ground it's hard (*) to make sure the right thing is done in tactical situations, and hard to tell after the fact how well you did. It will be easier for us to make systematic errors.

(*) Sometimes impossible. The commando raid on Bin Laden's compound left the kids alive. A B-2 strike would not have done that.

Comments for this post are closed

> You seem semi-reasonable on this.

I strive, often unsuccessfully, but I do strive, to be semi-reasonable. I find that on the internet, when I'm not striving to be semi-reasonable, I'm not. I suspect others are the same way.

Comments for this post are closed

My logic isn't circular, it is just a moderately complex system, which is why the bureaucracy can't handle it.

Our actual threat is not low-level peasants. The actual threat is actually addressed with the kind of effort dramatized in the hunt for bin Laden. (This is why I mocked Bush for "not thinking a lot about bin Laden at the time). The day after 9/11 everyone knew it was about human intelligence. Now we admit we don't know anything because we are just blowing away people. If they weren't low-level, we could use their intelligence.

You bring up another good point. We didn't even see fit to bring Osama in alive. We never needed the information from torture. There is no real organization threat and certainly not from low-level peasants and their children collecting firewood and attending weddings.

Obama is either actively pursuing the wrong strategy or being dragged into it by the CIA and Pentagon, both indicating he is worse than Bush- including the counterproductive ad hoc disorderly regime change strategy.

On every front, Obama has either accepted Bush policy or put his own spin on it to make it even worse.

Comments for this post are closed

> We didn’t even see fit to bring Osama in alive.

Surely it would have been very inconvenient for Obama to have had Osama captured alive... It's not necessary to believe that interrogation is useless or that there is no real threat to see the problem Osama in US hands would have presented to our fearless leader.

That said, I find it really hard to believe that Obama has had his arm twisted by the CIA. I just don't think he's that dumb and that manipulable. He did a complete about face when elected, going to being really half-hearted about closing Gitmo and raising the intensity of the drone program. I find this hard to square with the threat being low.

Comments for this post are closed

"He did a complete about face when elected, going to being really half-hearted about closing Gitmo and raising the intensity of the drone program. I find this hard to square with the threat being low."

We could have captured Osama. He was as unarmed as the children who were spared. If Al Qaeda were what we were told it was, then nothing would be more valuable than Osama alive. Of course, it's not, which is why he was killed.

Gitmo: It's way simpler than that. He won the election and didn't need the campaign promises anymore. He tossed them in the trash can along with the '08 stickers. As for the relationship of Gitmo to the drone program, did you read the book review article? They don't need the intel and don't want the headache of detainees. So, they are just killing people. Are they killing high-level operatives? Well, not important enough to capture for their intel. So, presumably, they are just being killed because they are somehow anti-American. So, the question is, the purely amoral utilitarian question that is, is 'can you assassinate enough people so that the world likes you?'

The explanations for the drone program and the regime change are geopolitical, because national defense explanations don't add up. You are giving Obama the benefit of the doubt that he must earnestly know something we don't despite the mounding evidence that he's just a loping opportunist.

Comments for this post are closed

Comments for this post are closed

Comments for this post are closed

Comments for this post are closed

Comments for this post are closed


The author's purported cure is far worse than the disease. Positional externalities from shaving latency are indeed real, but they're not really that large relative to market size. A good way to estimate their magnitude is by how much money has been spent on cutting down the Chicago-NYC messaging latency, the two most liquid and hence profitable trading centers. The cost recently spent on this infrastructure (largely microwave relay networks) is about $500 million. Assume that the infrastructure depreciates in about a year and generously assume that the spending on intra-market latency is about the same.

That's a total cost of about $1 billion/yr in market costs imposed by latency based positional externalities. American equity markets trade $24 quadrillion in value a year (and that's only counting shares, not derivatives). Which means the cost to the typical investor of the latency externalities comes out to an upper bound of $4.5e-05 per dollar traded, or for example to trade one share of MSFT: $0.0016. That's the upper bound of cost savings by perfectly eliminating latency externalities. The cost certainly isn't trivial, but it is much lower than the forced imposition of $0.005 in bid-ask trading costs because the SEC refuses to decrease the minimum $0.01 tick size. With an economical tick size the average bid-ask spread would easily go in half. (Plus it would reduce the latency externalities since market makers could price improve rather than rushing to jump first in line the order book queue). My point being is that if we're that worried about reducing costs to investors there's an alternative that we're already ignoring that both has a larger impact and poses much less risk than completely tearing up the foundations of the market structure.

Finally the authors assume that batch auctions don't come with any of there own structural costs. Not only do they indeed have substantial defects themselves, but they don't even eliminate the latency externalities. The market already uses batch auctions at market open and close. As any trader will tell you these are far more manipulated than continuous trading. During a batch auction an indicative price is published prior to crossing based on the currently resting buy and sell orders. A trader can easily change this indicative price or imbalance by entering a large order and canceling it before auction. Analogous strategies aren't impossible, but are much harder in continuous trading because a resting order can be crossed at any time, and hence poses real economic risks to the trader. To paraphrase Alex continuous trading acts as a tax on bullshit.

The flip side of a pre-cross indicative price is that traders will wait for as long as possible before the cross to enter their orders. No trader using proprietary signals is going to want other market participants to see his order for any longer than is absolutely necessary. The counter-strategy being shaving down your latency even further so you get to see others' orders first. Then modify yours accordingly by trading even closer to the cross time using your lower latency. So what frequently happens in opening and closing batch auctions is that the order book and indicative price is pretty much garbage until a few milliseconds before the cross, at which point the real price formation occurs. When I worked in a much larger HFT firm I was a continuous guy, but sat next to the batch auction guys. We certainly cared about our latency, but generally we focused much more on our signals and execution algorithms. The auction guys in contrast were always obsessed with their latency.

Switching to batch auctions will not reduce the cost of latency positional externalities, and is pretty likely to increase them. On top of all that it will give us a much lower-quality and less efficient market structure. There are certainly better ways to tackle the latency externality costs. But it's important to recognize the perfect's the enemy of the good here, I doubt we can ever fully eliminate them under any sane structure. It's better to think of moderate improvements that work on the margin, rather than centrally planned grand sweeping re-designs of the entire market structure.

Hi, thanks for this explanation, though I have one question. If the "indicative price" before the market-open batch auction is based on nonbinding bids (because they can be removed before auction clears), why don't people ignore it? A good rule of thumb in market design is that bids should be binding, and nonbinding bids should not be considered "indicative" of cost/value. Does this "indicative price" have any formal function in the market-open batch auctions, or is it just that market participants rely on them for submitting their own bids?

Although, even if the indicative price is just published informationally for market participants, I would hope a regulator is systematically screening for participants that routinely submit and withdraw orders within a fraction of a second, and other signs of disingenuous bidding. I'm pretty sure in most electricity markets, the market operators or regulators screen bids from generators and power traders for this sort of thing. FERC recently issued a steep penalty to JP Morgan for submitting disingenuous Day-Ahead market bids that were not a reasonable expectation of Real-Time costs for the sole purpose of manipulating a flaw in the "Day Ahead Margin Assurance Payment (DAMAP)" calculation, and so receive excess DAMAP payments.

Regulators don't do a particularly good job of monitoring this kind of thing. But if you put out a limit order at the spot you actually want to get filled at, there's no danger to you.

Comments for this post are closed

"A good rule of thumb in market design is that bids should be binding, and nonbinding bids should not be considered “indicative” of cost/value."

This is indeed a good rule of thumb, but its not an absolute truth. Non-binding bids do indeed have less informational content than indicative bids, but not zero. By 9:25 most of the orders in the open have already been entered. True some of these will be cancelled in the next minute, but not all of them. After all most of the net fundamental demand is coming from traders who aren't going to use colocated black boxes and wait until the last second. Auction liquidity providers need to use some measure of the liquidity imbalance to know how to enter their own orders. And a noisy estimate is better than no estimate. In addition there is still continuous trading occurring prior to the open (its pre-market hours trading, so its lighter is not subject to reg NMS). A system can look at both the continuous price and auction price to determine when the auction price smells funny.

To expand on this: if you deal in, say energy options, and you have storage numbers released at fixed times every week, you might want to put in limit orders (bids and asks) a few minutes before the release to deal with book hedging; because you will still need to handle option orders coming in. (the underlying futures is not your primary market segment).

Comments for this post are closed

Comments for this post are closed

Comments for this post are closed

- At the least, it's nice to see somebody arguing that there is a problem with HFT who isn't completely ignorant. The solution isn't quite there, but at least the authors have done some homework.

- I agree that it doesn't fix all the problems. It makes me think of Zeno's paradoxes: go with a continuum or discrete units and you wind up with one problem or another.

- Regarding the first argument, we've got de facto sub-penny pricing with the rebate system, so it's not really a half-penny built-in spread. In fact, I think that's why you're right that under a more frequent batch auction system people will try to find a way to fight over milliseconds. And since that $24 quadrillion is largely market-making etc., plus there is a lot of investment into market-making/technical-arb stuff beyond just HFT positional latency investment, the amount of money that goes into market efficiency is a bit bigger than you're saying.

They're right that, at some point, a milli-/nano-second doesn't benefit the market in any real way. But here's the thing: if the amount of money made on the arb isn't changing, it's not hurting the market any more, but instead it's all being re-invested into speed machines. Ultimately, then, it will cost so much to grab that arb that the margins disappear. Basically, then, HFTs are the big frontier in communication infrastructure innovation. And that is obviously a good thing! Just because they're not helping the markets any more (at some point, on humanly-meaningful time horizons) doesn't mean they're not useful.

- The more damning argument here depends on combining continuous orders with batch executions. I don't think they have to do that. To fix this you could disseminate order information simultaneously with executions. Additionally, a cancelled order can still be filled during the auction period following its cancellation, so that you can't have backing-away. (You wouldn't get rid of IOC orders, of course.) This would lead to all sorts of *trade-offs*, yes, but I think it's more true to the batch mentality and it circumvents the problem you're pointing out here, which is based on continuous quotes and order flow. A version of this already happens, as you know: with existing auctions, at some point prior to the official cut-off cancels and last-second orders are often too late to affect fills.

Of course, at best I think it's probably just a trade-off, not a real solution. Truth is that competition kills the margins in the absence of innovation, so other than the need to enforce good regulations that have been around since the 30s (no backing away, no front-running) I don't see why you need to do anything about HFT. But hey, it would move us into unknown territory, creating plenty of inefficiency for folks like you and me. ;)

Comments for this post are closed

Comments for this post are closed

I have abandoned two books in my life: Tess of the D'urbervilles and Gravity's Rainbow. Love the opening line in Gravity's Rainbow but have now tried to get over 100 pages into it twice and have failed. I will try again someday.

I renewed my checkout for Gravity's Rainbow and still only got halfway through. There are intensely brilliant passages, but its just too much work.

An attempt to list abandoned titles *per start* (or per shelf, in Goodreads terms, I suppose) would be appreciated. Gravity's Rainbow would have to be near the top. If you just look at the grand total with those tags you're just going to end up with a subset of the most popular books that happen to be somewhat controversial. That said, three of the top five classics are among my personal favourites. The the other two I detested, one of which I'd throw in the garbage before I'd give it to a friend.

Comments for this post are closed

Comments for this post are closed

I abandon books all the time, yet I still feel that I'm not aggressive enough in doing so. If I'm mildly enjoying a book I'll continue reading, but should I ditch it for another book and another until I find one with a much more high level of consistent enjoyment?

as a friend of several authors I would say they Want you to ditch their books if your enjoyment is at best mild. The ones I know would be happier with fewer passionate readers rather than more mild readers

to clarify, none of my friends make a living from writing .....

Comments for this post are closed

Comments for this post are closed

Comments for this post are closed

There is literally nothing AT ALL standing in the way of the creation of a frequent batch auction exchange. If someone thought they could make money with the concept, they would have created it by now. Capitalism, baby!

The fact of the matter is that for the vast majority of users, HFT are irrelevant. If you trade in the closing auction or with VWAP you are already immune to who would play in your batch auction sandbox?

Reg NMS might make it pretty tricky, at least if someone wanted to do it for US Equities.

Broker-dealers can use ISO orders which aren't subjected to Reg NMS. Pretty much every large trader, market maker, HFT trader, and internalization pool has broker-dealer status or piggy-backs on someone else's. So you could set up a batch auction exchange and only open it to these participants. Regular traders wouldn't have access, but internalization pools like Citadel or GETCO could fill orders at NBBO and then execute on the batch auction exchange.

They'd have to take the economic risk of filling at NBBO and holding until the auction crosses. But if auctions were done every second this would be pretty trivial. If the batch auction exchange really offered more economic execution, the profits should more than justify the cost. Since internalization pools pay for order flow and are pretty competitive with each other, most of the surplus from the batch auction exchange would accrue to the retail traders.

Reg NMS isn't the main impediment to a batch auction exchange. The reason no one's ever tried it is because almost every player knows that batch auctions are a far inferior market structure to continuous trading. (As I talk about in my above post). No one would invest the time or money in building something like that because it'd be a waste. Execution would almost certainly be far less economic, not more and no one would ever use the exchange.

ITG runs a batch auction type system once a day for US equities:

The dramatic growth of dark pools, mostly run by brokers, where traders can trade without displaying their orders to the public via the exchanges, and their potentially deleterious effects on price discovery is one that I think is far more important market structure problem than the arms race posed by continuous trading, for which, I agree with HFTrader, the costs are not that high relative to the amount of trading going on and the for which alternatives are worse. I also think we are likely to see the HF arms race come to an end. Competition, lower trading volumes, and lower volatility, have caused profits from HF Trading to fall dramatically since 2008 (the FT estimated an 80% decline), and the low hanging fruit - colocation, fast processors, fast lines, is mostly gone. Spending a lot of money on capital investments to capture a rather small amount of potential profit doesn't make sense - a rational investor would not spend 300 million to build a slightly faster data line between Chicago and NY.

Comments for this post are closed

Comments for this post are closed

Comments for this post are closed

Comments for this post are closed

#6: An ungated copy of the paper "Trust, Growth and Well-Being: New Evidence and Policy Implications" by Yann Algan and Pierre Cahuc is available here:

Comments for this post are closed

Tyler, what is this "abandoned" books stuff.

This is an image of "abandoned" books:

Here are some liberated books:

Comments for this post are closed

Has anyone worked on a way to avoid the need for trust? Like maybe some technology that would ensure transparency or something like that.

This is a big part of what cryptocurrencies are about: reduction in trust requirements. Open source software more generally is in part about trust reduction as well.

Comments for this post are closed

Comments for this post are closed

I suspect the stopping point is pretty much the same as the starting point: You read just far enough to manage a conversation with whomever knowledge of, e.g. Pushkin, is a cultural Rorschach inkblot. This is often someone you hope to see naked.

Cf. "Area Man Well-Versed In First Thirds Of Great Literature" (,1323/ ) and "Great Books Of Western Civilization Used To Accent Den" (,3776/ )

Comments for this post are closed

Selectively sampling stories of trivial or banal "innovations" is no substitute for comprehensive empirical evidence on the slowdown of real technological innovations, for which I have yet to see evidece. I am sure that for every story you quote there could be a counterargument, such as this one:

Comments for this post are closed

Footnote 2 in the HFT paper is one of the best footnotes ever.

Comments for this post are closed

Regarding HFT, how about changing exchanges so that, instead of trading immediately the full amount, a matched trade happens incrementally over a time period T (that is, after time T/4, 1/4 of the shares have been exchanged)?

This would allow market participants to post offers on the exchange without being vulnerable to the fastest guy who can react on sudden news releases.

The time T could possibly be specified by the market participants themselves: in this case a "fast" offer may simultaneously be matched and run against multiple "slow" offers, so that putting an "infinitely slow" offer doesn't have any effect on the market.

Comments for this post are closed

Of the top five abandoned classics, I finished and enjoyed Catch-22, Moby Dick (although it was assigned) and Atlas Shrugged (might have skimmed that speech). Put down LOTR after the first book and never cracked Ulysses.

Comments for this post are closed

Am I the only one shocked to discover the shelf life of Twinkies is measured in days, not years?

Comments for this post are closed

This excellent website definitely has all the information and facts I wanted about this
subject and didn't know who to ask.

Comments for this post are closed

Valuable info. Fortunate me I found your website unintentionally, and
I'm surprised why this coincidence didn't took place earlier!

I bookmarked it.

Comments for this post are closed

Its like you read my mind! You appear to know a lot about this, like you wrote the book in it or
something. I think that you could do with a few pics to drive the message
home a little bit, but instead of that, this is magnificent blog.

A great read. I'll definitely be back.

Comments for this post are closed

Comments for this post are closed