IT spending has hollowed out labour markets, to the detriment of middle-income workers
Source: Automatic reaction | The Economist
Not surprising by itself, but it is nice to have some numbers to confirm what I (and I suspect most other people) had suspected about the effects of IT on the job market.
I wonder where IT workers fall in the low-middle-high income scale.
Companies spend billions on rent, offices, and office equipment so their employees will have a great place to work.
However, when you ask people where they go when they really need to get something done, you’ll rarely hear them say it’s the office.
If you ask, you’ll usually get one of three kinds of responses: A place, a moving object, or a time.
Source: Why the office is the worst place to work – CNN.com
Obviously, as with all generalizations, there are exceptions. However, for “creative” jobs (programming, music, graphics, art, writing, designing, etc.) I think this is largely true.
Peter Gibbons: [Explaining the plan] Alright so when the sub routine compounds the interest it uses all these extra decimal places that just get rounded off. So we simplified the whole thing, we rounded them all down, drop the remainder into an account we opened.
Joanna: [Confused] So you’re stealing?
Peter Gibbons: Ah no, you don’t understand. It’s very complicated. It’s uh it’s aggregate, so I’m talking about fractions of a penny here. And over time they add up to a lot.
Joanna: Oh okay. So you’re gonna be making a lot of money, right?
Peter Gibbons: Yeah.
Joanna: Right. It’s not yours?
Peter Gibbons: Well it becomes ours.
Joanna: How is that not stealing?
Peter Gibbons: [pauses] I don’t think I’m explaining this very well.
Joanna: Okay.
Peter Gibbons: Um… the 7-11. You take a penny from the tray, right?
Joanna: From the cripple children?
Peter Gibbons: No that’s the jar. I’m talking about the tray. You know the pennies that are for everybody?
Joanna: Oh for everybody. Okay.
Peter Gibbons: Well those are whole pennies, right? I’m just talking about fractions of a penny here. But we do it from a much bigger tray and we do it a couple a million times.
High-frequency trading (HFT), also known as algorithmic trading, is hardly a new story. Many people, however, still do not quite know what it is, how it works, or why it is still legal. I am going to try to illustrate the point by providing an analogy that might make things clearer.
High-frequency trading is the trading of securities (e.g. stocks, bonds, derivatives) using powerful computing systems with special software, complex mathematics, and trade-secret algorithms in order to frequently execute extremely large numbers of trades. HFT is frequently mentioned alongside, and sometimes confused with, flash trading, co-location, and naked access. By 2006, 1/3 of all stock trades were executed by algorithmic actors [2] and HFT accounted for 70%-73% of total daily trading volume by 2009 [3][4]. Average daily volume on the New York Stock Exchange (NYSE) increased by 164% from 2005-2009 [5]. TABB Group and FIXProtocol both estimated annual HFT profits of around $20 billion as of 2008-2009 [6][7]. The SEC was investigating HFT by the summer of 2009 to determine if new controls or regulation were required, and subsequently proposed to ban flash trading [8][9] and naked access [10][11][12]. This potential regulation has been stalled for months [13] even after the SEC’s final date for the public to submit commentary.
Co-location is where entities or objects are located in the same place. In the context of networked computing systems, co-location describes the situation where servers are physically located in close proximity. Securities market centers which offer co-location services might put the customer’s computer in the same building, on the same floor, in the same room, or even in the same rack as the servers that actually create and manage the exchange by running its software.
There is a physical limitation to how many servers can be placed closest to the “action”. However, since trading speed is not absolute but only relative to other market participants, someone will be physically located closest to market center servers.
There are many places where one may attempt to buy or sell financial securities; some have floors, are entirely electronic, or have other features, but there is not one single entity known as “the market”. When orders are routed between securities market centers, the forwarding market center, exchange, brokerage, or trading firm uses a unique market participant identifier. For a fee, some of these entities allow customers to trade using the service provider’s market participant identifier in order to shave microseconds off of cross-exchange trade execution times – this is termed naked access, or sponsored naked access. Naked access accounted for 9% of trade volume in 2005 and grew to 38% of U.S. stock trading at the end of 2009 [14][15]. Overall sponsored access, including both naked and filtered access, accounted for half of all U.S. equity volume at the end of 2009 [14].
1) Reduced trade execution time is attained by skipping pre-trade risk checks.
2) Some naked access service providers permit the use of their market participant identifier without proper (or any) supervision.
3) There are no uniform rules enforced by the SEC covering naked access services.
First, one must understand the difference between at-market orders (buy or sell orders to be executed at the best available current price; if you click “sell” on your E*TRADE account, this is probably what happens) and limit orders (trade orders placed to sell at no less than or buy at no more than some specified price limit).
If someone submits a buy or sell limit order for a trade to a market center, then that marketplace first tries to fill the order locally (where it profits from the completed trade). If the marketplace cannot execute the order locally but a different location can, then the order must be routed out to the remote location as prescribed by Regulation National Market System (Reg NMS) of the United States Securities and Exchange Commission (SEC). This routing involves a route-out fee charged to whoever submitted the order. Additionally, in every trade there is a poster and a taker; the poster is “providing liquidity” by sitting around willing to buy or sell a security at some price whereas the taker is “taking liquidity”. The poster receives a rebate per share and the taker (or responder) pays a fee per share for conducting the trade on the exchange. The fee is slightly higher than the rebate and this margin is part of how an exchange makes its profit; take note that because of this pricing structure, the exchange profits from handling a larger volume of traded shares. I also want to emphasize that flash trading is voluntary, although it may be opt-out rather than opt-in, depending on the ECN.
Example: Now suppose that a stock is trading at $99.98-$100.05 (that is, the bid-ask spread is 7 cents where the most someone is willing to pay for a share is $99.98 and the least someone is willing to sell for is $100.05). Also suppose the per-share liquidity provision rebate is $0.0020, the per-share liquidity take fee is $0.0026, and the per-share routing fee is $0.0035. Now pretend you want to sell this stock and make as much as you can from the sale. You have 3 options:
The example liquidity and routing fees and rebate were drawn from the April 1, 2008 NYSE Arca Equities fee schedule [16].
The Securities Exchange Act of 1934 excepted flash orders from quoting requirements on the basis that they were ephemeral. Since 1978, paragraph (a)(1)(i)(A) of Rule 602 of Regulation NMS has contained language which “excludes bids and offers communicated on an exchange that either are executed immediately after communication or cancelled or withdrawn if not executed immediately after communication” [9]. The orders which qualify for this exemption are now referred to as “flash orders”. The exemption was intended to ease manual trading on exchange floors by not requiring quotations that were impractically difficult to include in the consolidated quotation data disseminated to the public. Some HFT has repurposed these kinds of trades to permit trade execution ahead of anticipated market movements.
1) Legal Front-Running?: For a nice graphical example, see “The Thirty-Millisecond Advantage” [17] by the New York Times for their article “Stock Traders Find Speed Pays, in Milliseconds” [5]. The article itself is not very good, simply mashing the outrage button by painting an unfair picture of the situation by confusing arbitrage as the only form/method/strategy of HFT, treating naked access and co-location as endemic to HFT, and failing to distinguish between the different kinds of trading orders. The graphic is good though for clarifying one way in which these market access tools can be abused – legalized front-running. If you submit a trade order to the market and your broker trades for their own account ahead of the order you placed through them in order to take advantage of their knowledge of your pending trade, then that is front-running and it is illegal. However, since anyone can pay a market center providing flash order service for permission to access the flashed trade information, that information is considered public knowledge and can be used freely. Some market participants with extremely fast computers and connections can thus read flashed data and act on it before the flash trade itself is fully acted upon.
2) Artificial Market Movement: If an HFT outfit sees a trend in the orders being flashed then it can try to nudge the market along in the same detected direction. For example, if it believes that Google stock is on its way up for the day then the HFT outfit could buy a huge number of Google shares at market+$0.01 and immediately turn around and sell those same shares at market+$0.02, pocketing the penny-per-share price increase as well as the liquidity rebate on both trades. Some might argue that the HFT outfit is simply moving the market price to a more accurate assessment of a security’s value at a given time with present information, but others feel the practice is closer to bidding by a colluding shill in an auction.
3) Restricted Access by Expense: Charles Duhigg quotes a high-frequency trader at the end of his article “S.E.C. Starts Crackdown on ‘Flash’ Trading” for The New York Times: “As long as everyone is subject to the same rules, I’m not concerned. Profits have always flowed to whoever dominates the marketplace, and we have a technological advantage that it costs millions to match.” [18] I think this quote provides an incredible insight into the difference between the perceptions of financial insiders and the general public. Is it fair that only people with millions of dollars can participate in high-frequency trading, or any other financial system? Industrial-scale manufacturing has such high barriers to market entry; are the barriers in finance naturally, appropriately, or artificially high or low?
High-frequency trading works by using present and historical price and news information as inputs to a virtual market model (a theoretical version of the market, intended to mimic and predict reality) and then taking action in the market using the output from the model and a trading strategy. Some potential strategies for high-frequency traders include: synthetic market making, provision/take rebate capture, liquidity detection, long-short statistical arbitrage, cross-asset arbitrage, short-term statistical arbitrage, and volatility arbitrage [19][20]. For a short, simplified example, consider a long-short statistical arbitrage (or StatArb) trading strategy. Prices for a specific stock have just jumped up (perhaps over the past ½ second). Your model predicts that the price of another stock is statistically positively correlated with the stock whose price is increasing (e.g a car company and its parts suppliers). The price of the other stock has not changed recently (perhaps the past couple minutes). The automated trader then buys the unchanged stock long and sells short the stock that increased. The short sale hedges against the possibility that the first stock’s price jump was a fluke and will momentarily drop to its previous price spread. As soon as the trading algorithm (or Algo) detects it has an unfavorable chance either for further profit or further loss, the automated trader will exit both positions in order to maximize profit or minimize loss. As long as the algorithm is statistically profitable in its trades (say up a penny 60% and down a penny 40%), then over the course of millions of trades the statistics will flatten out and the algorithm’s owner will walk away a very rich man.
It is possible to purchase special news-feed access from services such as Reuters, parse the news using a computer to determine whether the news is good or bad and which companies ticker symbols will be affected, and be the first to act accordingly. Some HFT outfit probably short-sold British Petroleum stock on April 20th, 2010 [Reuters] [WWL] before you even checked your email. Similar actions can be taken using sophisticated algorithms written in low-level computer languages and run on powerful computers to analyze the market for trends and statistical correlations.
HFT is able to cancel its orders if it detects a change in the market that disadvantages its previous positions. This capability is based on the same “ephemeral” exemption as flash trading. Some also claim that this has been used to unfairly detect the limit price of slower traders [21].
HFT has produced serious profits for financial firms [6][7] and of course financial markets do provide important services for which they deserve to be paid. Some of these services are: 1) supplying savings in exchange for debt plus interest, 2) mitigating risk, 3) supplying market liquidity by supplying and demanding securities in exchange for a trading margin.
1 A) Traders of any flavor, high-frequency or otherwise, must start with some initial capital. HFT actors generally do not maintain a financial position long enough to explicitly accrue interest, but they have added some amount of savings to the market. Even if they begin from a leveraged position (i.e. they took on debt to begin operation), no one smart would have lent them starting funds without significant collateral of some kind. Still, the savings value of HFT is probably marginal at best.
2 A) Since HFT neither sells whatever information it might have about risks directly, nor creates securities which might contain an assessment of such information indirectly, I see no reason to think that HFT mitigates risk for anyone. I also know none who would claim it does.
3 A) Market liquidity is how quickly and easily an asset, financial or otherwise, is to buy and sell without moving the price, and it has 4 components: immediacy (the speed with which the trade can be completed), breadth (the cost of providing liquidity often measured by the bid-ask spread), depth (the volume available for trade at a given bid-ask spread), and resiliency (how much the price changes due to large transactions). [22]
Immediacy: Competing HFT outfits have reduced the execution time for electronic trades to mere milliseconds. How many milliseconds would you wait before you were ready to pay to have the trade executed faster? I personally do not think there is substantial value added to investors by cutting transaction time from seconds to milliseconds.
Breadth: No one can make anyone pay more or accept less than the limit on their trade order. The spread in securities markets is the thinnest it has ever been, likely with a significant part played by competing HFT outfits and strategies frequently trading in volume. A thinner spread means that on average the seller is receiving more and the buyer is paying less than they would have otherwise.
Depth: HFT does approximately double the number of trades, since HFT takes some financial position only to exit the position perhaps a half-second later. However, since an HFT agent frequently does not itself hold onto a reserve of securities to sell nor buys without believing it can readily find a buyer, HFT as a system does not seem to increase the overall effective volume of supply or demand.
Resiliency: This component is trickier. At its current size, HFT is probably not substantially moving the market on its own and by adding willingness to trade is probably increasing resiliency. However, when HFT detects a shadow attempt to move a large block of an asset in chunks (see iceberging) the algorithm may try to take advantage of it by trading enough to affect the price, thus potentially reducing resilience. As HFT grows in size, the amount of effect it has on the market will also increase, and its automated nature may in the future lead to larger losses of resiliency and less stable pricing.
HFT increases market liquidity by objective, quantitative measures, and HFT outfits are clearly extracting significant financial value from the system for providing this liquidity. Does HFT need more regulatory oversight and should some of its tools, such as flash trading, be restricted or banned? Probably. Are we going to see meaningful reform [23]? Doubtful. Does the value to society of the increased liquidity outweigh the costs of HFT and does HFT increase the liquidity of markets in ways meaningful to those who must pay for it? These are very difficult questions.
[1] : Memorable quotes from “Office Space”, 1999
http://www.imdb.com/title/tt0151804/quotes?qt0996772
[2] : “Algorithmic Trading: Hype or Reality?” by Aite Group, 2005/03/28
http://www.aitegroup.com/reports/20050328.php?PHPSESSID=c6f095758583be80e5689ec5c06678fb
[3] : “New World Order: The High Frequency Trading Community and Its Impact on Market Structure” by Aite Group, 2009/02/25
http://www.aitegroup.com/reports/200902251.php?PHPSESSID=c6f095758583be80e5689ec5c06678fb
[4] : “The Real Story of Trading Software Espionage” by Rob Iati, Partner, The TABB Group, 2009/07/10
http://advancedtrading.com/algorithms/showArticle.jhtml?articleID=218401501
[5] : “Stock Traders Find Speed Pays, in Milliseconds” by Charles Duhigg, The New York Times, 2009/07/24
http://www.nytimes.com/2009/07/24/business/24trading.html
[6] : “Risk Management and Electronic Trading” by Bernard S. Donefer, FIXProtocol, 2008/05/29
http://www.zerohedge.com/sites/default/files/CUNY_algo_2008.pdf
[7] : “How big is high-frequency trading?” by Felix Salmon, blogs.reuters.com, 2009/07/30
http://blogs.reuters.com/felix-salmon/2009/07/30/how-big-is-high-frequency-trading/
[8] : “SEC Proposes Flash Order Ban” by the Securities and Exchange Commission, 2009/09/17
http://www.sec.gov/news/press/2009/2009-201.htm
[9] : “Elimination of Flash Order Exception from Rule 602 of Regulation NMS” by the Securities and Exchange Commission, 2009/09/18
http://www.sec.gov/rules/proposed/2009/34-60684.pdf
[10] : “SEC proposes “effective” ban on naked access” by Rachelle Younglai and Jonathan Spicer, Reuters, 2010/01/13
http://www.reuters.com/article/idUSTRE60C4DP20100113
[11] : “SEC Proposes New Rule to Effectively Prohibit Unfiltered Access and Maintain Market Access Controls” by the Securities and Exchange Commission, 2010/01/13
http://www.sec.gov/news/press/2010/2010-7.htm
[12] : “Risk Management Controls for Brokers or Dealers with Market Access” by the Securities and Exchange Commission, 2010/01/13
http://www.sec.gov/rules/proposed/2010/34-61379.pdf
[13] : “Ankle Bracelets For High-Frequency Traders” by Liz Moyer, Forbes, 2010/04/14
http://www.forbes.com/2010/04/14/exchanges-flash-trading-business-wall-street-sec.html?boxes=businesschannelsections
[14] : “ “Naked access” now 38 percent of U.S. trading: report” by Jonathan Spicer, editing by Leslie Gevirtz, Reuters, 2009/12/14
http://www.reuters.com/article/idUSTRE5BD0HT20091214
[15] : “SEC: No More Naked Access” by Alexandra Zendrian, Forbes, 2010/01/14
http://blogs.forbes.com/streettalk/2010/01/14/sec-drive-responsibly-in-the-markets/
[16] : “NYSE Arca Reduces Fees for Equities and Options Trading, Effective April 1, 2008” by NYSE Euronext, 2008/03/31
http://www.nyse.com/press/1206959049219.html
[17] : “The Thirty-Millisecond Advantage” by The New York Times Company, 2006/07/24
http://graphics8.nytimes.com/images/2009/07/24/business/0724-biz-web2TRADING.jpg
[18] : “S.E.C. Starts Crackdown on ‘Flash’ Trading” by Charles Duhigg, The New York Times, 2009/08/04
http://www.nytimes.com/2009/08/05/business/05flash.html?fta=y
[19] : “TABB Group Pulls Back the High-Frequency Trading Curtain” by Adam Sussman, Director of Research, The TABB Group, 2009/10/07
http://www.advancedtrading.com/algorithms/showArticle.jhtml?articleID=220301041
[20] : “We Fear What We Don’t Understand” by Kid Dynamite, Kid Dynamite’s World, 2009/07/26
http://fridayinvegas.blogspot.com/2009/07/we-fear-what-we-dont-understand.html
[21] : “High Frequency Trading Is A Scam” by Karl Denninger, The Market Ticker, 2009/07/24
http://market-ticker.denninger.net/archives/1259-High-Frequency-Trading-Is-A-Scam.html
[22] : “Liquidity, liquidity, liquidity” by David Longworth, Deputy Governor of the Bank of Canada, to the Investment Industry Association of Canada, Toronto, 2007/10/3
http://www.bis.org/review/r071005e.pdf
[23] : “SEC looks to rein in trading battlebots, maybe” by Jon Stokes, Ars Technica, 2010/04/08
http://arstechnica.com/tech-policy/news/2010/04/sec-looks-to-rein-in-trading-battlebots-maybe.ars
“Reports: New Research from Aite Group”, by Aite Group
http://www.aitegroup.com/reports.php
“SEC Moving on Flash Trading, High Frequency Trading” by Manuel Arreaza, The MTTLR Blog, 2009/11/23
http://www.mttlrblog.org/2009/11/23/sec-moving-on-flash-trading-high-frequency-trading/
“Cash Cow – High-Frequency Trading” by Samantha Bee, The Daily Show, 2009/09/30
http://www.thedailyshow.com/watch/wed-september-30-2009/cash-cow—high-frequency-trading
“SEC May Curb ‘Flash Trading.’ Goldman Says It’s Not In The Game” by Laura Conaway, NPR.org/blogs, 2009/08/04
http://www.npr.org/blogs/money/2009/08/sec_may_curb_flash_trading_as.html
“Goldman’s $4 Billion High Frequency Trading Wildcard” by Tyler Durden, Zero Hedge, 2009/07/17
http://zerohedge.blogspot.com/2009/07/goldmans-4-billion-high-frequency.html
“< ‘The Quants’: It Pays To Know Your Wall Street Math” by Terry Gross, NPR, 2010/02/01
http://www.npr.org/templates/transcript/transcript.php?storyId=123209339
“More on High Frequency Trading” by Kid Dynamite, Kid Dynamite’s World, 2009/12/08
http://fridayinvegas.blogspot.com/2009/12/more-on-high-frequency-trading.html
“You Really Shouldn’t Care So Much About Flash Trading” by , Kid Dynamite’s World, 2009/09/20
http://fridayinvegas.blogspot.com/2009/09/you-really-shouldnt-care-so-much-about.html
“Rewarding Bad Actors” by Paul Krugman, The New York Times, 2009/08/02
http://www.nytimes.com/2009/08/03/opinion/03krugman.html?_r=4&adxnnl=1&ref=opinion&adxnnlx=1249304676-mDiSQ0yDhjzLlz1BgJSWKg
“Flash Point: Equities industry clashes over flash and step-up orders” by Nina Mehta, Traders Magazine, 2009/07/–
http://www.tradersmagazine.com/issues/20_296/-103978-1.html
“What’s Behind High-Frequency Trading” by Scott Patterson and Geoffrey Rogow, The Wall Street Journal, 2009/08/01
http://online.wsj.com/article/SB124908601669298293.html
“Concept Release on Equity Market Structure” by the Securities and Exchange Commission, 2010/01/14
http://www.sec.gov/rules/concept/2010/34-61358.pdf
“The Matrix, but with money: the world of high-speed trading” by Jon Stokes, Ars Technica, 2009/07/28
http://arstechnica.com/tech-policy/news/2009/07/-it-sounds-like-something.ars/
“US Equity High Frequency Trading: Strategies, Sizing and Market Structure” by Adam Sussman, Larry Tabb, and Robert Iati, The TABB Group, 2009/09/02
http://www.tabbgroup.com/PublicationDetail.aspx?PublicationID=505&MenuID=13&ParentMenuID=2&PageID=8
“Unintended Consequences of Market Structure Regulation” by Larry Tabb, The TABB Group, 2009/12/08
http://www.sec.gov/comments/s7-27-09/s72709-21.pdf
“Focus Topic: High-Frequency Trading” by Thomson Reuters
http://hft.thomsonreuters.com/
“Why Do High-Frequency Traders Cancel So Many Orders?” by Matt Levine on Bloomberg View, 2015/10/08
http://www.bloombergview.com/articles/2015-10-08/why-do-high-frequency-traders-cancel-so-many-orders-
The authors made the claim that GDP is a poor measure of a country’s wealth, because it does not include externalized environmental costs. This may be true, but the authors went a step further. They said “….using GDP to measure a country’s true wealth is remarkably poor at best, and highly damaging at worst.”
Source: Wonky Thoughts: In Praise of GDP, by Doug Robbins
RE: Greening of the Gross Domestic Product | EARTH Magazine, by Garrett C. Groves and Michael E. Webber
Data visualization on Gapminder shows that increasing wealth, measured by GDP, is correlated with nearly doubling human lifespan in almost every country on earth:
The data show stunning (>95%) reductions in child mortality as GDP rises:
(Note the logarithmic scale for both child mortality and income.)
And perhaps most importantly, how fertility has fallen from 5 children per woman to about 2 children per woman (stable population!) as prosperity has risen, as measured by GDP:
A remarkable aspect of these relationships is that they hold for every nation in the data. As shown in an earlier post, per capita GDP even shows a strong positive correlation to integrity, as measured by the corruption index published by Transparency International.
…
Imagine for a moment, that there was an economic indicator that showed the reverse correlation: that as the indicator rose, human life expectancy was cut by half; child mortality increased twenty-fold; population growth rose from zero to doubling in every generation; and corruption flourished. Would the authors dismiss such an indicator as remarkably poor and highly damaging?
Does there exist any other indicator, other than GDP, that is so useful in measuring the progress of a nation in improving the lives of its people?
This is the point where the agency throws up its hands and turns the question on the public, but mostly in a manner that weakly attempts to disguise rhetorical questions as real ones. … And this is why, instead of issuing a public request for comments, the SEC would have done better to issue a public request for better questions first.
Source: SEC looks to rein in trading battlebots, maybe | Ars Technica
