Agreed. ML is the "Wolf Hunter", but in this case, he was hunting sheep. I am astonished by the press's reaction to this. I think a lot of it has to do with the relatively small and obscure HFT industry - it is not well understood, and it is easier to believe it is rigged than to understand what it is.
To even call it "high frequency trading" is not fair - this implied large volumes throughout the day by continuously providing quotes on both sides of a given security. Their PnL comes from the bid/ask spread and rebates for providing liquidity. They are primarily market makers. This is 99.99% of the HFT industry.
Instead, ML and IEX are talking about "speed trading" or, as the press has coined it, "latency arbitrage" - they take advantage of their speed to reach exchanges (2 milliseconds vs. 20 milliseconds) to front-run orders and react quicker to new information. These types of trades only happen at points during the day when signals are met, and they do not provide liquidity but rather cross existing orders (often times before the exchange receives the cancel message from the participant) and TAKE OUT liquidity. I would say less than 1% of HFT firms engage in this type of unethical activity.
An example: Imagine it's 9:29am and the Dept of Labor Statistics is going to release the monthly unemployment numbers. There is a positive expectation, and so before market open there are a lot of buy SPY (S&P 500 ETF) orders queued. The report comes out at 9:30am as the markets open, and the numbers are bad. Now, a rational investor would immediately attempt to cancel his order for SPX as the market is going to move downward. Imagine at the same EXACT time, a speed trader see's this investors buy order on the book and decided to cross him and sell. Due to his speed advantage, the exchange receives his message to cross before the investors message to cancel. The investor loses out, SPX invariably moves down, and the speed trader then buys everything he just sold for an essentially risk-less profit.
A final point is this: for any buy-side market participant (anything from a mutual fund to a "Average joe"), transaction costs are much lower due to the much higher liquidity and tightened spreads that high frequency trading has brought to the markets. HFT is making the markets more efficient. Cliff Asness (Founder, AQR Capital Mgmt) wrote a piece in WSJ talking about this: http://online.wsj.com/news/articles/SB1000142405270230397830...
SPX is the index. SPY is the ETF. Also the idea that you can rest an order with an expectation that if the news is good you profit but at worst you lose nothing is ridiculous. That would be a free option. If you want that kind of optionality you have to pay for it. If you rest an order, especially before a big news event, you are taking a risk that you'll get hit and the market will move against you. That's how markets work and there isn't anything wrong with that.
I think you missed my point: if people are reacting to info at the same time (e.g. clicking at the same time) the exchange should receive the message at the same time. Similarly, market data should be transmitted at the same speed, and exchanges should not offer co-location to give one participant faster data than another - This is how they front-run.
I didn't miss your point and you don't have any idea what front running is. Front running is when your broker, your agent who has a legal responsibility to seek the best execution for your order, trades in front of your order. If someone who you have no relationship with manages to react to a news event faster than you do, that's not front running.
If you took away exchange colocation, you'd just have a boom in property values around the exchange as automated traders would still try to get on the shortest network path to the exchange. Exchange colocation is actually a way to level that playing field. Everyone who is colocated has exactly the same network delay to the matching engine as everyone else who is colocated.
Lewis is not saying that all investors should be given the exact same access to the exchanges. He is saying that some traders are extracting money from the market simply because they have a faster route to the exchange and they are buying/selling ahead of other investors. They are not buying/selling based on market fundamentals, speculation, research.
Millions of Americans invest a large percentage of their retirement money in 401ks because they believe that the US Stock Market is a relatively good place to make long term investments.
When we see examples of people skimming money off of the market it makes us lose faith in the market. That is why this practice should be banned.
The problem is that Lewis either through misunderstanding, narrative purposes, or due to a conflict of interest is implying "buying/selling ahead of other investors". But that is not what is actually happening in the latency arbitrage case. In the latency arbitrage case a HFT uses trade information from one exchange, to update their own prices on other exchanges.
I find it fascinating that once people that understand the details they feel that nothing is wrong with this, and the people that stand back and look at the big picture see a rigged market.
It is Heisenberg morality. Once you look at it, there is nothing wrong with it. They are just updating prices to reflect demand, but at the same time they are extracting money from the market just because their data center is physically closer to the exchange.
" at the same time they are extracting money from the market"
This is the sentiment I don't understand. They aren't extracting any money. HFT market makers provide a service in the form of liquidity. It is natural and expected for the price of liquidity to go up as demand for it rises. There are liquidity purchasers that want to minimize the price of liquidity and therefore obscure their demand. This is also natural and expected. These are the forces that drive price discovery, this is what we want the markets to do.
Computers have made this process extremely fast and efficient to the benefit of nearly every market participant.
Which practice, exactly? You described one party having a faster connection than another. That's not exactly something you can "ban". Lewis even acknowledged that relative inequality of this nature will always exist.
The same way insider trading is banned. It is impossible to prevent 100% of insider information being involved in some trades but you can prosecute people who can be proven to have made trades on insider information.
You don't ban faster connections the same way you don't ban insider information. Faster connections and insider information will always exist. What you ban is trading on specific information that puts you at a substantially unfair advantage.
If I'm the CFO of company X I have an unfair advantage when it comes to making trades minutes before the quarterly results are release for my company.
You ban trades that are made in an attempt to purchase shares that you believe someone else is also attempting to purchase based on knowledge that you have received because of your proximity to an exchange.
This would be very similar to insider trading in that there is going to be a grey area where it would be hard to prove that the intent was to undercut someone else but having the regulation there would at least curtail this practice somewhat.
As I understand it, George W. Bush deregulated the stock market and allowed multiple venues to compose the market. This resulted in stock trading becoming distributed among multiple venues. When a retail customer submits an order to an exchange they are required by law to give the customer the best price available (no front running). Since exchanges are now distributed they must communicate to ensure the best price is given to the customer. If a market participant can outrun this calculation by placing themselves on exchange A and exchange B simultaneously and transfer the information faster than the participating exchanges, then they are effectively able to front run the customer. There are no rules around this outrunning of the best price calculation, although most people agree there ought to be because it makes execution quality worse.
A technological solution to this problem is to either execute on alternative venues such as dark pools where the price and possibly quantity are hidden from all participants, or use an algorithm such as VWAP to hide your intentions. Smart money already does both of these. If you submit block orders to a public exchange now a days you're considered foolish and are getting a bad deal. The reason why this keeps coming up is because the rubes are finally starting to realize what wall street already knows: they are getting screwed.
You do make a cogent point about discussion of automated trading being laughably imprecise, but your argument doesn't really debunk the immorality of the behavior that is occurring.
And finally, your point about non-farm payroll is a non starter for me. What you are talking about is informed order flow (wall street), VS. uninformed order flow (you and me). When non-farm payroll is released designated market makers have been known withdraw from the market because the price is being corrected by the informed order flow. When volatility quiets down they re-engage their market making activities.
"As I understand it, George W. Bush deregulated the stock market and allowed multiple venues to compose the market. This resulted in stock trading becoming distributed among multiple venues. When a retail customer submits an order to an exchange they are required by law to give the customer the best price available (no front running). Since exchanges are now distributed they must communicate to ensure the best price is given to the customer. If a market participant can outrun this calculation by placing themselves on exchange A and exchange B simultaneously and transfer the information faster than the participating exchanges, then they are effectively able to front run the customer. There are no rules around this outrunning of the best price calculation, although most people agree there ought to be because it makes execution quality worse."
Your understanding of what is going on in latency arbitrage is incorrect and implies that a latency arbitrage trader can see your order before it executes on an exchange. This is not true.
Further, the only people who think independent price synchronization provides worse execution are large block liquidity takers. That is large institutional investors who have the intention to remove all the liquidity from a group of exchanges. They have always tried to hide their intentions so that the market cannot take those intentions into the account. They are now using scare tactics to make it seem like this is a problem when in fact it is the markets behaving as they should.
Market segmentation & correct price discovery help small retail investors not hurt them.
That is large institutional investors who have the intention to remove all the liquidity from a group of exchanges.
One of the central themes of Flash Boys is that this (sweeping multiple market centers) is somehow a hard thing to do. I think to most practitioners it seems like Brad Katsuyama and RBC were just exceptionally bad at it. Yes you have to invest a little bit in technology and network connectivity, but the big sell-side banks have the money to make those sorts of investments. Given that they got paid to competently execute these big trades, it seems almost inexcusable that their excuse was like "We didn't actually understand what was going on."
There are all manner of odd themes in the book. That Brad Katsuyama got paid exorbitant amounts of money to execute trades poorly for large institutions is presented as a virtue, even though his bonus came specifically from those large institutions profits we are so outraged that HFT is trading against.
My favourite one is a section where Lewis starts naming HFT algos and they invariably have names like Dagger, Viper, etc. yet he never seems to point out the HFT Brad Katsuyama created was named Thor.
If BATS has the best price or matches the best price available in the market, your order will execute at BATS. If there is a better price available at another market center displaying a protected quote, BATS has to route your order to that exchange.
Continuing with the example, lets say I'm trying to hit my price, and only 300 shares are available on BATS and another 700 are available on some other venue. Would BATS send an order for the remaining 700 to the other venue, or does it only route the order and not chop it up?
Depends on the routing instructions in the order, but generally speaking if BATS matches the best price then BATS will match as much as it can against the liquidity that is available at BATS and route the remainder. See page 2 of this PDF for BATS: https://www.batstrading.com/resources/features/bats_exchange...
Actually, GS will never be a "normal" bank holding company. Their commercial banking business is open to a very select number of clients, and will stay that way. GS became a BHC because they would have access to the Fed's emergency funds in case of another major crash -- a line of defense. GS hardly makes any money from their commercial banking arm.
GS's core business will remain their cash cows: Investment Banking, Securities, and Asset Management (in that order). GS will NEVER wind down their Investment Banking arm, in which GS is king above all other banks (closely followed by Morgan Stanley).
GS's core businesses actually did quite well relative to the rest of the industry. This loss is really just "on paper", coming primarily from the Private Equity portfolios. GS has a reputation for aggressively marking-to-market, and their PE portfolio lost a lot of value on paper because the equity markets shit the bed this quarter.
This guy is an independent trader because no one would hire him. He's misguided in his understanding of the markets. Goldman Sachs is an investment bank. When he says "anyone can make money from a crash", he's right: any INDEPENDENT investor/fund. Such as a hedge fund or himself, an "independent trader". These people are referred to as the "buy side". However, Goldman Sachs, as well as all the other banks he probably thinks "rules the world" is on the sell-side. The sell-side provides "prime" brokerage services to the buy-side clients -- that is they connect buyers and sellers via the exchanges. In fact, with the upcoming Volker rule, no investment banks will be allowed to engage in proprietary trading (trading for profit with the firms money), which is what the buy-side does.
Investment banks might actually lose money in recessions because they might take illiquid, toxic assets onto their books to service demand (point and case: the mortgage crisis). And securities is only a part of the investment bank business model. Advisory services, largely driven by M&A and IPO volume, provide a decent chunk of profits for banks. Capital markets dry up during recessions, which will completely stifle M&A and IPO activity and therefore revenue on that side of the bank.
This guy is full of shit. When asked what to invest in when the market goes down, his best advice is bonds and "hedging strategies". Bonds do indeed rise in value during bear markets, however hedging has almost nothing to do with profit or loss. Hedging is risk management: covering your ass in case of an unexpected move. For example, if I expect a downward market turn, as per his advice, I might buy up treasuries. But, to "hedge" the possibility that the market moves UPWARDS instead, I might buy an index tracking the Dow, which will increase in value as the market moves up. In this case, hedging is actually DECREASING my profits in the case of a downward movement in the markets. There are much more intuitive ways to play a downward market.
The basic thing is that while there are ways to play down-markets, down-markets and up-markets are not mirror image and are not simply "different ways to make money".
An up-market inherently creates - maybe-not-money - but the appearance of money, the availability of money, "liquidity". An healthy up-market inspires healthy production and makes the liquidity it generates really correspond to people having more wealth on average. An unhealthy up-market naturally involves mis-allocated resources and its liquidity thus becomes illusory and so it is followed by a down-market. A down-market eats liquidity and decreases production meaning the decreased-money people have really corresponds the people also having less stuff, on average. So given the downer that is a down market, profiting become harder on average. Some do great but the average person should assume that the law of averages to applies to them...
For any interested in Burry's story, pick up Michael Lewis's The Big Short. Great (if somewhat miscontrued) tale of the housing crisis, ripe with corrupt financiers and the "smartest men in the room".
Burry's lightbulb concerning the crumbling housing market was a product of a staggering amount of research on mortgages, contra to the research (mostly by rating agencies) already published. No average Joe is going to foresee a bubble about to explode.
I was thinking something more conventional. For example, contrary to what many may believe, history actually IS a good predictor of future. As an investor, I am not only limited to investing in individual companies -- I can also bet on entire markets/sectors (for example, Burry bet against the housing market). Also recall that the markets are cyclical (that is, recessions follow booms and vice versa).
With that in mind, I could, for instance, have a sector-based model hinging upon the business cycle. Certain sectors, historically, have tended to outperform during different segments of the cycle, and with well-timed bets I can always make money just by recognizing what state of the business cycle we are in.
For example, currently we are in a (if somewhat shaky) "recovery" phase. During recovery, financials and tech companies tend to outperform. I might use ETFs (IXG and IXN) to go long on these markets. I might even enhance my bet and short Consumer Staples, which are expected to underperform during recovery.
However, any kind of shorting strategy involves not just an expectation that the market will go down some time in the future but instead requires that you say exactly when.
Especially, if the stock that you are short begins rises, you may be forced to buy back the stock you've sold - the traditional "short squeeze".
http://wiki.fool.com/Short_squeeze
Basically, playing to a down market is inherently harder than playing to an up market. It can be done but it is harder. Just another way the video is full-of-shit as many folks have mentioned.
To even call it "high frequency trading" is not fair - this implied large volumes throughout the day by continuously providing quotes on both sides of a given security. Their PnL comes from the bid/ask spread and rebates for providing liquidity. They are primarily market makers. This is 99.99% of the HFT industry.
Instead, ML and IEX are talking about "speed trading" or, as the press has coined it, "latency arbitrage" - they take advantage of their speed to reach exchanges (2 milliseconds vs. 20 milliseconds) to front-run orders and react quicker to new information. These types of trades only happen at points during the day when signals are met, and they do not provide liquidity but rather cross existing orders (often times before the exchange receives the cancel message from the participant) and TAKE OUT liquidity. I would say less than 1% of HFT firms engage in this type of unethical activity.
An example: Imagine it's 9:29am and the Dept of Labor Statistics is going to release the monthly unemployment numbers. There is a positive expectation, and so before market open there are a lot of buy SPY (S&P 500 ETF) orders queued. The report comes out at 9:30am as the markets open, and the numbers are bad. Now, a rational investor would immediately attempt to cancel his order for SPX as the market is going to move downward. Imagine at the same EXACT time, a speed trader see's this investors buy order on the book and decided to cross him and sell. Due to his speed advantage, the exchange receives his message to cross before the investors message to cancel. The investor loses out, SPX invariably moves down, and the speed trader then buys everything he just sold for an essentially risk-less profit.
A final point is this: for any buy-side market participant (anything from a mutual fund to a "Average joe"), transaction costs are much lower due to the much higher liquidity and tightened spreads that high frequency trading has brought to the markets. HFT is making the markets more efficient. Cliff Asness (Founder, AQR Capital Mgmt) wrote a piece in WSJ talking about this: http://online.wsj.com/news/articles/SB1000142405270230397830...