Saturday, May 08, 2010

Aftermath: Remedies - Don't Lose Faith in Markets - Lose Faith in Market Orders!

First off, if you've missed my prior posts on the Crash of May 6th, please take the time to check them out now.  There is also ample valuable information from knowledgeable readers in the comment threads.  They are well worth your time to read through.

Now, where do we go from here?  Senators are already calling for investigations and studies, and there will be talk about increased legislation and regulations.  Here are my suggestions.

1) IF the NASDAQ wants to cancel trades that are at "crazy prices" - to use the subjective term - and I think it's beyond debate that they do want to cancel such trades, since they already demonstrated such, THEN steps should be taken to prevent these trades from taking place in the first place.  There is no reason to allow these trades only to cancel them after the fact.  If you want to protect traders from themselves, then PROACTIVELY make sure that the trades can't get printed unless people really want to trade at those prices.  Institute market wide individual stock based trading curbs, perhaps similar to the NYSE Liquidity Replenishment Points, but make sure that they apply to price increases as well as decreases, and make sure they aren't structured to attempt to prevent stock prices from falling.  We should not try to prevent lower stock prices via legislation!  We SHOULD try to prevent "abnormal" trades - if we want to protect traders from themselves.

2) Who are we protecting anyway?  I think it's widely agreed upon that no one wants to see professional investors bailed out of computer trading algorithms run amuck.  If a crazy computer algo takes a stock from $30 to 1c and then back to $30, who cares?  Investors shouldn't care - if you didn't read the news or watch TV intraday, you might not even know that the stock moved!  Your portfolio value is unaffected.   Traders should absolutely LOVE it!  If there are crazy computer algorithms in the market doing crazy things, guess what, that makes it much easier to make money!  We rational humans can outsmart algo's gone wild with ease, right?   The first thing I did Friday morning was enter a bunch of lowball bids in a number of stocks, PRAYING for a repeat of whatever happened on Thursday.

Of course, retail investors who lacked a true understanding of what they were doing got plugged on stop orders, which turned into MARKET orders when their price level was triggered.   One claim is that trading activity like this makes people "lose faith" in the market.  Don't lose faith in the market - lose faith in market orders! I like that credo.  Please join me here at Kid Dynamite's World in declaring war on market orders.   One commenter on Ritholtz's blog sarcastically noted, only semi-sarcastically, I think:

"I would always advise people against entering stop losses that become market orders. Perhaps it’s those type of orders that should be banned. Should I call my Congressional representative to try and enact a law to stop those people who use them from hurting themselves? Do you advocate that?"

I strongly agree with that sentiment, and noted that there's really no reason why we couldn't ban the "market" order type.  Really.  There's no reason why anyone should ever use a market order. None.  In fact, BATS exchange automatically adjusts all incoming market orders to limit orders, where the limit is the greater of 50c or 5% of the stock price:  higher for buy orders, and lower for sell orders.  Just a little bit of built in protection.  So, related to that point:

3) Brokers, especially broker catering to retail investors, should step up their client education platform IMMEDIATELY.  It's in their own best interest, obviously.  I'm sure Etrade doesn't want to have to explain to a client why their sell stop order in ACN got filled at 1c.  They make you demonstrate experience or investment competence to trade options, and they should do the same with market orders and stop orders.  If we want to protect investors from themselves, and again, I think it's been demonstrated that we do, we should educate them and make sure that they have all the information they need to make an intelligent, responsible decision.  If someone wants to enter a stop order, they need to be certified to do so  - this really isn't a big deal - the SEC could design a 5 question quiz demonstrating that when you enter a sell stop order at a price of $30, you understand that you can get filled at $30.25, $30, $29.99, $15, or 1c.  Investors need to understand this, and we need to continue to remove "excuses" from our markets.

4) Should we ban "high frequency trading" ??   I certainly don't think so.  As I've noted in my previous posts: of course the increased speed and technology of our modern markets contributed to the severity of the crash - but aberrations due to technological quirks are quickly self correcting (as we saw!) - stocks bounced right back.  When fast cars crash, they do more damage than horse & buggy crashes - but we don't ban automobiles.

One reason I included the tremendous audio from the S&P Futures Pit in Chicago is that it's an old school market with no computers.  The same people who blame high frequency trading for dominating markets also blame them for walking away! (the Wall Street Journal wrote an article stating as much, and others have expressed similarly confused views)  Well, non-HFT liquidity providers had ample opportunity to step in and show how much better they were when the HFT guys turned their systems off right before the big downdraft (please read PeterPeter's comment about why HFT guys did this - it's due to problems they were seeing in being able to accurately process information on time).  The S&P futures pit audio shows that there was massive illiquidity there also - in a non-computerized, old school market with traditional liquidity providers.  No one wants to step in and get run over when markets crash.  Computers/no computers/stock/bonds/commodities - it doesn't matter.  When markets crash, they crash hard and fast.  You can't regulate or legislate that behavior away!

5) Let's not forget something - prices got crushed because there were more (or more aggressive) sellers than buyers.  It's not because machines wanted to destroy the world, or because the markets are manipulated.  I think the decline was long overdue, as the massive previous rally in the markets was based largely on forced investing, goosed and fueled by the Fed in the way of zero interest rates.  Investors demand return - they need return, so they sought out risk assets in a way very similar to the bubble we're trying to recover from - buying across asset classes in a manner I thought was reckless.  When you buy because you HAVE to be a part of the party, you have a quick trigger on the way out - you don't want to be the last one holding the hot potato.   Throw in a dose of pure economic disaster and contagion brewing in Europe, and you have all the ingredients for the perfect storm.

I use the term Ponzi scheme a lot on this blog, and markets designed to go higher based on the need for more buyers participating rather than fundamental improvements are also destined to crash hard when the buyers disappear.  On Thursday May 6th, the buyers disappeared and the sellers wanted out.

To summarize - it wouldn't be hard to institute reasonable individual stock trading curbs to slow down markets a bit in times of extreme stress.  It would also be a noble goal to try to continue to educate retail investors who don't know what they are doing, so that they can avoid trades that result differently than they had intended.  I think trying to legislate away market movements is impossible, though, and that minimal intervention is needed:  after all, this crash was a long time in the making.

There's a saying:  stocks take the stairs up, and the elevator down...



txchick57 said...

Ha. I did the same thing Friday a.m. - entered GTC orders at crazy prices on a bunch of stocks I'd like to have.

EBIX, a company I really like who had blowout earnings traded from 15 to 75 cents and then to 15 again. The NASD broke trades under 6 so some people still doubled their money in minutes.

Transor Z said...

Hopefully an audit trail of transactions will be released before too long that will allow a data-driven post mortem on what happened on 5/6.

Kid, I'm pretty ignorant about the mechanics of market fragmentation, dark pools, etc. Can you (or any other readers) recommend any good sources to start with for those of us who want to come up to speed on that topic?

Kid Dynamite said...

transor - on dark pools, read my piece:

as far as market structure... i'm not sure where to start - maybe the NYSE's website?

basically, think of it like this: in the old days, there were a bunch of traders standing around a post at the NYSE trading IBM. Now, there are a bunch of different exchanges, each trading IBM, and each with their own liquidity - but the bottom line is that RegNMS requires all of them to execute at the NBBO or better. this then leads to flash trading, so you can read this:

flash trading's goal is to internalize order flow within a market center, so they can avoid the expense of routing the order out - and thus the person whose order is being flashed avoids the expense...

anyway - think of all the different exchanges as spokes coming out of the old NYSE post - only the NYSE isn't the boss anymore, and there are still rules to make sure that the best bids and offers are protected - which is to say that they are not traded through. as i wrote in an earlier post though, the problem this week was that the NYSE's quotes lost their "protected" status, and basically didn't count - when the NYSE's LRPs were triggered and they went into "slow" status. that was here:

Onlooker said...


I like you point 5 most of all. For months there's been all the chatter about why the market's going up, whether it made sense fundamentally, etc. And the answering mantra is that you have to just get in the game, follow the trend, and keep a close eye on the door.

Very few out there truly believe that the markets are anything but overvalued and going up on excess liquidity, or whatever. And so they're in the market but have a hair trigger on the sell button.

So go figure that once the momentum swung to the downside, and after bullishness reached extremes after a relentless rally from the Feb lows, that there would be a stampede for the exits at the first sign of real ugly action.

I mean really. And now all these geniuses are baffled as to what could have possibly gone wrong. We never learn, eh?

I'm only amazed that the intense selling pressure didn't hit gold as well (for more than the few brief minutes that it did). Sure gives credence to the bull thesis there, doesn't it? Glad I was holding miners and not much else.

And I'm also glad to have a bunch of money in the wise and safe hands of John Hussman. He will be vindicated; again.

Transor Z said...

Thanks, KD. I think "market structure" was the big piece I didn't get last year when bashing flash trading was all the rage.

q said...

a couple questions after reading the comments here.

-- why would fractured liquidity cause a big price drop? i could see it exacerbating a problem, but not causing one. the market was going down in a pretty steady and orderly fashion prior to the crash.

-- if you look at the list of securities with broken trades, you see a lot of sector ETFs. is there a possible reason why they would be more effected? i would think the opposite, that people wouldn't have stop losses in place for these rather than single names.

-- probably most importantly, why did the market recover as quickly as it crashed? what changed the dynamics? what suddenly changed the minds of momentum players?

Kid Dynamite said...

Q - i like to say "markets aren't totally efficient, but they are efficient ENOUGH" - that answers the third part of your question. they weren't efficient enough to prevent ACN from trading to 1c, but they were efficient enough to rapidly remedy the situation.

as for your first part: fractured liquidity didn't start the sell off - stocks were going down no matter what on Thursday, as i mentioned near the end of the article - they were ripe for a bloodbath.. the fractured liquidity offers an explanation for the EXTREME moves.

as for the sector etfs - many of them are less liquid in general, and thus probably more susceptible to the scenario we saw. i can't speculate as to if they have more or less stop orders in their order books.

q said...

basically what you are saying is 'there was momentum and suddenly fragmenting liquidity concentrated the momentum' and that makes some sense, but as an explanation it's pretty general and vague and certainly doesn't preclude the possibility of a single statarb firm going mad.

also, i don't see how the market sell-off that day was naturally a big cause of this. markets go up and down all the time, but HFT timescales are much shorter than human ones. so sure, the market was down that day. was the market down drastically in HFT timescales (ie what happened in the seconds and minutes before the crash)?

so in short i don't doubt that your explanation is plausible, or part of what happened, but i'm not convinced it was sufficient to cause the crash.

The Fundamentalist said...

the nyse shld do what Globex is doing:

Market orders on Globex are implemented using a “Market with Protection” approach. Unlike a conventional market order where customers are at risk of having their orders filled at extreme prices, market with protection orders are filled within a predefined range of prices (the protected range).

The protected range is typically the current best bid or offer, plus or minus 50 percent of the product’s No Bust Range (see below).

If the entire order cannot be filled within the protected range, the unfilled quantity remains on the book as a Limit order at the limit of the protected range.

Stop orders with protection are intended to avoid cascading stop orders being filled at extreme prices. Stop with protection orders are filled within a predefined range of prices (the protected range).

A stop with protection order is triggered when the designated price is traded on the market.

The order then enters the order book as a limit order with the protection price limit equal to the trigger price, plus or minus the predefined protected range.

The protected range is typically the trigger price, plus or minus 50 percent of the No Bust Range for that product (see below).

The order is executed at all price levels between the trigger and protection limit price.

If the order is not completely filled, the remaining quantity is placed in the order book at the protection price limit.

A buy stop order must have a trigger price greater than the last traded price for the instrument. A sell stop order must have a trigger price lower than the last traded price.

nick said...

I came across your blog for 1st time this weekend. Great posts on Thursday's crash. I couldn't believe it when i saw you're in concord, nh. I'm in henniker, 20 min away. Never thought it would be someone in nh providing the info to help better understand this event. keep it up

Anonymous said...

Hi Kid. You've said in a couple of these articles things akin to 'HFT guys turned their systems off'.

The Wall Street Journal article described a couple who switched off, but that doesn't mean all.

Where's the press for traders who used technology to trade through the chaos?

Anonymous said...

You were defending HFT last year:

do you still defend it?

EconomicDisconnect said...

This issue will be with us for a while. Next week should be interesting.

scharfy said...


Your posts on this have been some of your best stuff IMO.

My instincts tells me that the NYSE's system, whereby a stock or group of stocks is "slowed" for 90 seconds, and "real" bids are allowed to materialize and come into the market is the best way to do things.

The weak point in the market structure seems to be at NASDAQ, where ETF's traded 50% of their value, a penny, and other values which later had to be busted.

These erroneous prints triggered selling that wouldn't otherwise have occurred.

In times of panic, the size bids from institutions and real stock buyers that are needed to stabilize things just need that little 90 sec break to get in there.

All in all I don't think the market performed horribly, but NASDAQ should be fined for allowing prints to occur which contributed to the sell-off and which later had to be busted.

It was interesting to see the "hybrid" S&P futures model work admirably as the futures market didn't really miss a beat. Yes the markets were wide - but they were liquid, real and tradable. (remember S&P futures were 10 handles wide on an index basis of 1100. This is only 1% market width which is like a 100 dollar stock trading a buck wide. Not great but certainly acceptable given the circumstances.)

But whatever methodology each exchange uses, allowing false trades to occur is far more destructive than halting or slowing trading - which is why I favor the NYSE system in this instance.

Allowing bogus prints( which amount to false info) compromises the whole system and really should be penalized. The SEC should be heavy handed on this, as in my eyes is the root of the problem.

EconomicDisconnect said...

I wish they would stop BIDU every time it ramps up $50 and ask the traders why, lol.

Douglas Knight said...

What do you think of the suggestion in the comments of not busting trades below a threshold, but instead retroactively changing the price to be at that threshold? It seems like a pretty small change that's only an improvement.

EconomicDisconnect said...

EU is going for their own TARP? Word is about 560 Billion Euros for a bailout fund. Almost the TARP 700 Billion on the nose. It never ends.

Charlie said...

" When fast cars crash, they do more damage than horse & buggy crashes - but we don't ban automobiles." Even without banning automobiles, we still have a large set of rules governing where various types of traffic can go. For instance, I can't take my bicycle on the highway, 65mph traffic doesn't flow down my residential street, and formula 1 cars are restricted to tracks where the only other cars are also F1 race cars. In a crash between an F1 car and a family sedan, I'd actually bet on the F1 driver surviving. I guess what I'm getting at is that your analogy seems to support the idea that certain markets should be restricted to only high frequency trading, while other markets would be justified in banning it. I know that's not the point you were trying to make, but it's hard not to read that from your example.

Mencius Moldbug said...


Pardon my ignorance. But every time an ignorant person like me looks at this system, I ask: (a) why, in 2010, don't we have a a central limit-order book? (b) how many of these "HFTers" are actually injecting genuine information into the market, and how many are profiting from inefficiencies that would not exist given a CLOB?

Kid Dynamite said...

ok folks - sorry - i was out all day. let me try to briefly respond to the questions in the comments:

nick in Henniker - i'm closer to you than you think. i'm in Hopkinton. I'm in Henniker a few times a week to go to Sonny's or the Feed Store.

anon @ 11:23: sure - congrats to the guys who traded through the chaos using technology and made money... the public hates that though! they're the enemy! ;-)

anon@11:46 re: defending HFT: absolutely, the entire point here is that the mainstream media will tell you that the problem is "high frequency trading" when in reality that's probably very far from the truth. part of the problem is DEFINITELY electronic, fragmented markets. Another way of saying this is that 20 years ago on the NYSE, ACN almost certainly never could have traded at a penny. HOWEVER, on the NASDAQ, or ARCA, it could have... IT's the difference between having intermediated markets, and not...

douglas knight: re: adjusting trade prices instead of busting them - i'm not a big fan of that idea. just because you bought ACN at 1c doesn't mean you want to buy it at $10. I know it's easy to say in hindsight that those who bought at $10 would still make money, but that kinda skirts this issue that the buy at $10 may not have offered them the risk-reward they want. Maybe it's better than busting trades totally - if you adjust them to the low of the day - but i think preventing trades from being busted or adjusted proactively is even better.

charlie - I wasn't crazy about the automobile analogy, but I thought it was good enough. and to your point: we absolutely do have strict rules in place regarding all trading in our markets - despite what some would like you to believe... I certainly wasn't implying that come markets should be restricted to only HFT.

menicus - we used to have a kinda central order book - it was called the NYSE! competition changed that... as for your second question, that's not something i can answer in a comment thread, but you're probably on the right track. there is certainly a lot of exploiting of inefficiencies, which is a good thing, by the way. I've said in the past that i think the ideal market would be one central order book that's DARK... all you'd see is the inside market - best bid and offer.

to me, that's the purest market.

clancy said...
This comment has been removed by a blog administrator.
clancy said...
This comment has been removed by a blog administrator.
The Fundamentalist said...


you know the guys from Terra Nova? Press reports says that the P&G sell order originated from them

Transor Z said...


Terrific post and comments. Only footnote I would add to your statement that there is certainly a lot of exploiting of inefficiencies, which is a good thing, by the way is that, IMO, the kind of beneficial arbitrage I think you're alluding to completely breaks down in a thinly traded environment, the most spectacular example being the risk market c. 2007-08.

And again, ratings and financial transparency play key roles in identifying "true" inefficiencies!

Anonymous said...


Thanks for the post. Very useful. By way of background, I am a the sole employee of my own stat-arb fund (relatively small, but big enough to be considered a large trader under the new SEC proposal).

I shutdown at 2:41PM, and stayed down for the balance of the day.

Question for you. Did you guys measure the latency on quotes hitting the multicast feeds relative to their internal timestamps?

By the time I shutdown, I saw around 100ms delays on ARCA packets (I use the non-compacted multicast feeds about 2ms away from ARCA's servers).

Unfortunately, I didn't have my code running with a high enough debug level on other exchanges to log the latency, so I didn't measure BATS/Nasdaq/OMX BX, but I am really interested in knowing if the other major electronic exchanges performed better than ARCA.

Any input?


Mencius Moldbug said...


That's an answer but not an answer! The question of whether a CLOB is better is distinct from the question of whether the NYSE was better. Ie, did "competition change that" because NYSE sucked, or because CLOBs suck?

Eg: to sharpen the question, assuming an efficient electronic CLOB, would there be an incentive to trade off-CLOB? Difficult to imagine one.

Alas, if my suspicions are confirmed, HFTing is just a bunch of parasitic makework, like so much else in our economy. The fact that there were no standing limit orders in PG, ACN, etc, down to a penny, is particularly remarkable and reflects the general lack of information input into this market - a phenomenon I know you've remarked on.

What concerns me most about the general dominance of stat arb strategies is this lack of information input - the quants are not feeding off patterns created by a market of fundamental analysts, but simply off the patterns created by other quants. Unsurprising (in retrospect) that a failure case should exist in which all these noise traders shut down at once.

Mencius Moldbug said...

BTW, I like your flat dark market idea. I was thinking the same thing just this morning. Transparency is a source of inefficiency - by publishing open orders off the best bid and offer, information which those placing the orders would rather not publish, you are causing the publishers of those orders to effectively lose money. Ie, you are sucking their blood like a vampire, and distributing it in a fine rain over Connecticut.

Note that in the "purest market" (which publishes only the best bid and offer price) there is really only one knob you can turn if you want to sell a gazillion shares into the market. You can rip the band-aid off fast or slow. Technology can presumably be applied, but how much technology can you apply to one knob?

So, as a complete outsider, I really think that much if not all this high-speed finance is entirely wasteful and inefficient. Nice that it gets people paying attention to the latency of their TCP stacks, though!

Kid Dynamite said...

funamdentalist - i do NOT know the TerraNova guys, but they already put out a rebuttal saying that they didn't trade or clear ANY PG below a price of $60 on the day, so they are saying it absolutely was NOT them...

Kid Dynamite said...

menicus - first off, i think you're using "HFT" and "electronic trading exchanges" synonymously, while they are not the same thing.

also, make sure you read Clancy's comments.

as for a CLOB - my point was that we basically tried that already... and we didn't like the way the guy running the CLOB (NYSE specialist!) was printing money by scalping our trades... so we democratized it through competition, and now we have other players scalping teeny tiny amounts in greater numbers. It's unavoidable, IMO. You can't outlaw competition or profit.

great quote about distributing the blood in a fine mist over connecticut. lol.

Kid Dynamite said...

clancy - thanks for the comments.

Mencius Moldbug said...

I really appreciate Clancy's explanation. Information of this quality is not widely available.

It's not clear to me that "HFT" really means anything. In the 21st century, why shouldn't everything be high-frequency? The question is simply what kind of algorithm you're using to drive your trades. "Fundamental analysis" being one such algorithm.

What Thursday revealed, it seems to me, is that most of the bids in the market are placed based on transient and obscure correlation patterns, and relatively few are placed as a result of fundamental analysis. This does not prove that the broad price level (which did recover on Thursday) is not set fundamentally, but it is surely not evidence for the proposition!

As for NYSE, a CLOB that's scalping your trades is anything but the "purest" market! Arguing against CLOBs by using the NYSE as an example is like arguing against flush toilets by saying that, when you flush a clogged toilet, it overflows.

Kid Dynamite said...

sorry menicus - you lost me. i never called the NYSE the "purest" market, and I already said that i think one big dark pool, which i think is pretty much the same as your CLOB, would be the purest market.

all i was trying to say is that the old NYSE model was a different kind of CLOB that was inferior to what we have today.

Mencius Moldbug said...

Ie: in your "purest market", it is a clear conflict of interest for anyone to trade on the non-public book. The pure market simply matches orders and publishes best bid and offer prices. (In fact, one could imagine an even purer market that was perfectly dark - the only way to get a quote would be to trade.)

I don't know exactly what NYSE specialists do. But it sure smells like a corrupt legacy operation.

The fact that no market without such a conflict of interest exists is not a sign of healthy capitalism. It's a sign of the breakdown of the rule of law. In a healthy regulatory environment, clean and pure outcompetes dirty and conflicted.

Kid Dynamite said...

yeah - totally dark (no bid/ask published) would be beautifully philosophically pure - I think i've commented as much in the past on other blogs, but it would NEVER fly.

a market maker has to have a conflict of interest - he is trying to provide liquidity without trying to go bankrupt while doing it... what's the remedy? to make market making a government run money losing operation? oy vey - please no.

incidentally, and without any desire to get off on this tangent, we discussed this with respect to GS's role as market maker in CDOs... they CANNOT look out for their clients' best interest, because their clients' best interests conflict with their own! there's nothing inherently wrong with that..

and that's PART of the reason why we moved away from specialist driven systems in the first place - so that the specialists couldn't monopolize that market making profit.

Mencius Moldbug said...


No, you didn't. No disagreement here - just a matter of emphasis.

Mencius Moldbug said...

Balls. OK, now we disagree. Why should a market maker have to trade for his own benefit? What are his expenses? He's running a server room. Maybe he has $10K a month in Amazon EC2 costs. Maybe it's $100K. But it's sure not watering all the lawns in Connecticut!

For instance, suppose I'm IBM. Why shouldn't IBM run its own CLOB, in IBM shares? Simply keep a database in which each IBM share is marked with an owner and a limit price, and a list of bids. Buy IBM - submit a bid, see if it gets filled. Sell IBM - set a limit, watch it get filled.

This would absorb one gazillionth of IBM's IT capacity. Why shouldn't IBM provide this service for free, so as to maintain a clean market in its own shares?

Now, it is a little silly for each issuer of a security to have to run a book for itself. IBM can do it. Med Patent can't do it. So: outsource. One big database for everyone.

But no. The confusion between trading infrastructure and market participation is endemic on Wall Street and always has been. Perhaps it once needed to be. It strikes me that there's still a lot of legacy assumptions from the paper age.

Kid Dynamite said...

menicus - don't get confused. Market makers take risk. they need to be compensated for this risk. they won't provide orderly markets for you if they are losing money doing it.

again, it seems to me that you're confusing "trading exchange" with "market maker".

your IBM example is not about market makers at all, and it wouldn't maintain a clean market in its own shares - it would maintain a horrendously thin, illiquid marketplace in which to trade its own shares.

if you expect someone to provide liquidity, you should expect them to (try to) make money doing it. full stop. i can't argue this point - it's fundamental.

Mencius Moldbug said...


Sorry - mea culpa. I am confusing you, by using the word "market maker" where I meant "exchange." In my example, IBM runs its own exchange. So just s///.

I am not in any way arguing that creating liquidity is not and should not a profitable enterprise.

I actually greatly dislike the term "liquidity." I feel it is used to mean at least two categorically separate things, if not three.

What I believe you mean here by "liquidity" is: the coefficient by which an order moves the market. Ie, a market is "liquid" if a big order moves the price only a little, and "illiquid" if it moves it a lot. A perfectly fine terminology if the word was not otherwise employed.

If everyone who holds and/or trades IBM trades it on IBM's private exchange, then this exchange (CLOB) contains all the IBM liquidity in the world. That's what makes it "central," n'est ce pas?

So there remains a disconnect here. Are you asserting that, in order to act as a "market maker" and provide liquidity, ie, price resilience against irregular order flow, an agent has to be able to see into the book? (As, I presume, an NYSE specialist can.) I would strongly disagree with that assertion - or would at least like to see it backed up.

But if the assertion is not true, what exactly constitutes a liquidity provider? I guess what I'm asking is: what are the regulatory and competitive mechanisms that prevent the market from converging on your preferred solution, the "one big dark pool" with the same interface for everyone - just publishing a log of trades, perhaps, with no visible bid or ask?

Kid Dynamite said...

menicus - no - i definitely don't think that one needs to see into the order book to be a market maker, although if the market maker CAN see such things, it will result in his ability to take MORE risk.. (ie, if IBM is trading $100, and there is a $95 bid for size, the market maker may be more likely to provide liquidity at $97, thinking he has less risk if things move against him)

but to answer your question - i see no reason why we couldn't have my one big dark pool solution - except for the fact that politicians, regulators, and retail individuals aren't philosophically savvy enough to understand that it would work.

Mencius Moldbug said...
This comment has been removed by the author.
Mencius Moldbug said...

Right. But of course, if he can see into the order book, he can also devise almost infinitely complex schemes by which to scalp his (so-called) customers.

BTW, I like your coverage of EuroTARP. You might enjoy Carlyle on bankruptcy:

"Great is Bankruptcy: the great bottomless gulf into which all Falsehoods, public and private, do sink, disappearing; whither, from the first origin of them, they were doomed..."

Kid Dynamite said...

yes - what i described -looking at a $95 bid and bidding $97 as a result - IS scalping customers, in a way, and yet at the same time, it's helping the seller, and preventing further price declines...

Doc Merlin said...

This is going to sound conspiratorial, but this seems like someone took advantage of the reflexivity that is always going on about.

Baltimore Financial Planner said...

I think it all boils down to education, like you noted. I would bet most people reading your blog don't need a primer on the finer points of trading, but the problem really lies with the reactions of the uninformed to bumps and "glitches" in the markets.

Dave Pinsen said...


You make good points here about the virtues of limit orders. Another tool I think too few individual investors are taking advantage of is hedging, and I think that's mainly because for most of them, when it comes to options, it's Greek to them.

We developed a site called Portfolio Armor to sweep away the Greeks and simplify the hedging process for investors (it's a tool I use myself). Contact me if you would like to demo the site.

If, after trying it out, you think it might be a useful tool for your readers, you can become an affiliate and earn 50% recurring referral commissions for every reader of yours that becomes a subscriber.

slackful said...


A market maker is by definition an entity that doesn't want to own a stock. He need have no more understanding of the fundamentals of a stock than a bookmaker does of the two competing teams. He simply places at risk offers to buy and sell and adjusts them based on what he learns about where the equilibrium price is located. In so doing he acquires unwanted temporary long and short positions that he disposes of by further adjusting the prices he offers. He exists because he is demanded, and can make money so doing (though not easily). Knowledgeable investors don't curse his scalping, they are grateful that, due to him, they can pick up the phone or log on to a computer and instantly buy or sell stock without going through the pain and suffering of finding a counterparty themselves. It's just that painfully simple. Getco's guy on the SEC HFT panel did a wonderful (and down to earth, as a former floor trader would be expected to do) job of explaining just this. A flash recording is available on the SEC site.

Why all the fuss about technical information being used to adjust prices? Large size has always moved markets, probably for about 30,000 years. I loved the quote by someone: to not adjust prices based on technical information is like playing hide-and-seek with a 4 year old, and pretending you don't see his leg sticking out behind the couch. The technical information is material and important for fairness. Of course everyone complains about the market makers, but it's in the sense that they complain about the price of roast beef.