Tuesday, June 12, 2012

You're a Mean One, Mr. Grinch - Or Are You?

An Inside Look at High-Frequency Trading

High-frequency Trading has been in the news a lot in the past few years.  I’d like to say it has been in the sort of way the recent Queen’s Jubilee has, a highly covered, well-reported, highly-esteemed occasion that at its heart is nonetheless an ultimately boring and uneventful affair of pomp and ceremony.  Unfortunately, a better analogy is with the photos of a drunk and underdressed celebrity, let’s call her Britney, exiting a car with a little less modesty and decorum than usually befitting a grimy alleyway at 2am.

Why the animosity, why the disdain? The most visceral responses are to general headlines about executive compensation, insider trading, market manipulation, and collusion.  It’s not as pervasive as it seems, however.  The recent rash of such headlines have less to do with any real increase in such occurrences and more to do with public interest in them, sharply risen with outrage over the banks’ involvement in the housing market, outrage over government bailouts, and outrage over the growing disparity between the upper echelon and the general public.  The abundance of passion and ideals, as well as ire, has fueled the Occupy Wall Street movement and continues to ricochet at dining room cocktail chatter.

The irony is that high-frequency trading is almost completely absent in these problems.  The genesis of high-frequency trading comes from the “high-frequency” part, which requires lots of participants who are confident in the rules, understand the simple nature, and find the process easy and affordable, and so frequently interact.  This ubiquity and confidence is primarily found in the stock market.  Such cannot be claimed by the OTC derivative markets, credit default swaps, mortgage backed securities, convertible bonds, and numerous other deals and contracts heavily customized on a one-off basis and then sold and resold with little oversight and hazy mathematical mechanics.  By contrast, from hollywood to grandma’s house, nearly everyone understands how to invest in stocks, with the protection and oversight provided by the SEC, FINRA, numerous legislation and an abundance of case-law around each.  This means regulators and politicians understand, for the most part, how the stock market works and have as a result built an extensive system of checks and balances with vigorous monitoring and strict enforcement.  Far from subverting these rules, high-frequency trading thrives adhering to these rules, because these rules ensure fairness, ease, and ubiquity, the exact ingredients that lend to the creation of the “high-frequency” part of the name.

Moreover, due to the frequency of the trades, all profits are considered “short-term capital gains” by the IRS.  While the lens of public scrutiny over capital gains being taxed at a flat 15% is well deserved, it requires mentioning that only “long-term capital gains” and some interests and dividends are taxed at this rate.  This means private-equity firms like Bain Capital and their former members like Mitt Romney paid this highly coveted, effective rate.  It also means that high-frequency trading firms pay ordinary income tax, which is what short-term capital gains are treated as.  A high-frequency trading firm in New York City pays a combined federal, state, city and borough tax rate north of 45%.  If there were any hard evidence that high-frequency trading is not a privileged member of the establishment or bosom buddies with political elites, the harsh tax treatment is it.

Why, then, all the focus on high-frequency trading?  The ease of understanding stock markets that makes it accessible to the multitude of participants is the same ease of understanding that makes it a convenient target for the political microscope.  It also fits the luddite narrative of evil machines taking over the world, a very evocative and frightening image.  It also plays towards the sentiment that there’s some inherent unfairness and difficulty in the stock market which resonates at a time of plummeting net worths, a prolonged recession, and postponed retirements.  This narrative is convenient in that it garners readers easily, sounds extremely plausible because it does contain a kernel of truth, and helps move the spotlight away from the many real problems with Wall Street and inequality that desperately need reform.

Let’s demystify high-frequency trading.  A lot of the misconceptions persist because while fanciful prose about machines, the speed of light, and number-crunching algorithms have been penned and repenned, no one actually knows what it is.  The most common misconception is that, somehow, really-fast machines buy before someone else can buy and then sells back to that initial person for a higher price.  The first problem with such a notion is that it means high-frequency trades cannot lose money, which as any auditor can assure the public, is not true.  Like most financial transactions, nearly half of all high-frequency trades lose money.  The second problem with the notion is over how it implies the high-frequency firm has advanced knowledge of the initial buyer’s intention.  This is called “front-running” and not only is it already illegal, but is nearly impossible to execute since high-frequency firms are not retail brokers and are therefore not privy to a buyer’s intent.  The easiest way to tell whether the writer has no idea how high-frequency works is to see the word “somehow” in the description of the trading strategies.  It might as well read “magically” as a tell-tale sign of fantasy not fact.

Now that we know what it isn’t, what is it?  Well, we first need to understand “two-sided quoting”, and a delicious way to do so is with cake!  It’s an old puzzle:  two individuals both desire cake and it needs to be divided in a way both deem fair.  Even with rules like “cut down the middle”, what prevents the cutter from cheating and giving himself a bigger piece?  One clever solution is to have one individual cut the cake, and the second individual choose which piece to take.  If the first person cuts unfairly, the second will pick the larger piece and only the cake-cutter winds up with the smaller portion.  This system ensures the first person will try to cut the cake in the fairest way.  In financial lingo, the first person is considered the “market-maker”, a fancy word for the same concept.  In the stock market, a market-maker thinks of a fair price for, say, shares of IBM, and offers to buy and offers to sell around that price.  The passerby can then choose whether to buy, whether to sell, or keep going without interacting.  Accusations that the price is unfair, however, would be silly since that only hurts the market-maker’s interest in the same way it hurts the cake-cutter.

Okay, so market-makers seem to actually have some value, but how do computers and fast speed come into play?  First we need to understand what a “spread” is.  It’s jargon for the subtle difference between the highest price someone is willing to buy a stock and the lowest price someone is willing to sell a stock.  If you travel overseas and try to exchange currency, you’ll also notice a spread; converting from dollars to euros back to dollars will leave you with fewer dollars.  Having smaller spreads benefits investors in the same way being able to convert from dollars to euros back to dollars and winding up as closely to the initial amount benefits travelers.

Now, suppose an investor has the view that Facebook should have a market value somewhere between 40 and 70 billion dollars.  This investor will then have a willingness to buy stock when the market value of the company dips below this range and a willingness to sell the stock when the market value rises above this range.  In between the range, the investor is neutral and will likely neither buy nor sell.  The investor, like a market-maker, has a two-sided quote, except the spread is $30 billion of market value.  A company like Facebook does not change in market value by $30 billion very often or very quickly, so the frequency of such trades are low, even though the magnitude of the trading opportunity may be large.  This would be low-frequency trading.

By contrast, if an investor has a view that Facebook should have a market value somewhere between $58.01 and $58.03 billion, then the investor will find himself frequently buying and selling the stock.  This is high-frequency trading.  Unlike the investor who may be very confident about the $40 to $70 billion range, and therefore be willing to make a massive investment, an investor is likely to not be very confident about the $58.01 to $58.03 billion range and thus take a tiny investment.  These two investors co-exist within the same ecosystem, one attempting to make big payoffs on sizable investments over opportunities that occur once in a while, the other attempting to make modest payoffs on small investments over opportunities that occur multiple times a day.

Hmm, again, how do high-speed computers come into play?  Well, in order to be semi-confident about the tight spread around the market value of the company from the last example, the investor needs to process a lot of information that is rapidly updating.  The information that directly affects the company does not update frequently, such as quarterly earnings, customer surveys, the health of the CEO, pending court cases, employee retention and so on.  However, information that hints about the company, such as whether a related company has suddenly surged or fallen, does frequently update.  These relevant but not directly related data points are considered correlating factors.  It’s a bit like seeing the neighbor’s kid getting bit by a dog and then worrying about your own kid.  While information that directly affects the company can send its value up or down by several billions of dollars, these indirect but correlating factors only affect the company subtly, often by less than a tenth of a percent in value.  However, the interval of $58.01 to $58.03 billion dollars in market value for Facebook translates to only 1 cent per share, so subtlety is key.  If the two-sided quote is a penny per share too high or too low, the high-frequency investor loses 100% of his spread.  This means a lot of subtle, correlating factors that frequently update need to be processed and the forecast for the stock updated in a timely fashion, a feat only made possible by well-designed computer software.

The comparison between high-frequency and low-frequency trading is very analogous to the comparison between weather and climate.  Weather is about predicting whether it will rain this afternoon in your city.  It requires watching minute-by-minute the clouds over nearby cities and communicating that back to your city and having a computer quickly process the data.  Climate is about predicting how sea levels will rise over the next ten years.  It uses monthly or annual data, also uses computational methods, but may use less efficient software.  Both are essential endeavors.

What about news over how these computers destabilize markets, and caused the ‘Flash Crash’?  The Flash Crash was a widely reported incident wherein the U.S. broad market fell by more than 5% in the span of minutes, and regained to its original level over the course of a few more minutes.  The popular blame lay squarely on “robots going berserk”, a diminutive way of accusing high-frequency traders for the volatility.  Most of such news reports are sadly written by people who have no clue what they’re talking about.

The SEC has scrutinized the incident and found the origin of the Flash Crash to have likely emanated from a futures contract trader who had misentered the quantity of contracts to sell by two orders of magnitude.  The massive selling spree of $4.1 billion, manually instigated, would have sent the futures market to depths far lower than what was witnessed if high-frequency trading firms hadn’t formed a virtual mattress to soften the plunge.  However, this cushioning effect is achieved only because the load is distributed to a variety of other related stocks and products.  When high-frequency firms see the futures contract go down while the highly related underlying stocks do not go down, they see a mismatch.  It brings into doubt whether the stocks that didn’t go down perhaps ought to have, and also into doubt whether the futures contract that did go down perhaps ought not to have.  This downward doubt on stocks and upward doubt on the futures contract resulted in high-frequency firms taking small, measured trades that sent the stocks slightly lower toward the related futures contract and sent the futures contract slightly upward toward the related underlying stocks.  In essence, high-frequency firms acted as a type of elastic glue that slowed down the falling object and tugged related stationary objects down a bit.  Blaming high-frequency firms for the Flash Crash is a bit like someone running a car into a brick wall at 60mph and blaming the airbag for breaking their nose.  After the firm that entered the manual sell orders detected their mistake and stopped, the futures market quickly rebounded, and the glue-like mechanism provided by high-frequency trading firms brought back up the related, underlying stocks.

Okay, so perhaps high-frequency trading isn’t harmful, but how is it fair for investors to make money while not contributing to society in any material way?  Well, this is a bit of a loaded question in that it presumes no societal benefit by investing.  Investing is a broad area, and high-frequency trading is only one specific type of investing; nonetheless, it’s fair to question whether investing helps society.  We’ll take a quick aside before answering.

Companies can raise money in basically one of two ways, raising money through borrowing from a bank or bond holders, or raising money through selling an investment stake.  That is, instead of borrowing $20 million and having to repay it with interest, they may give an investor 10% of the company in exchange for $20 million.  They never have to repay an investor, since the money isn’t borrowed -- the company sells 10% of itself for that money.  In both examples, the company raises $20 million and may hire employees, expand its facilities, or otherwise spend the cash however it sees fit.  In the investment example, the company has the added benefit of never being on the hook to repay -- it doesn’t have a gun to its head to constantly pay interest or risk bankruptcy.  For this reason, many companies choose raising money through investors.

It’s a little less understood how the stock market contributes to the ecosystem.   The investment example from earlier is called “private equity investment”.  As the word ‘private’ may denote, it is not open to the public and unless you have a chat with the CEO over dinner, such opportunities are not available to you.  This is problematic not only for the general public but also for the CEO who may not have that many dinner buddies interested in investing.  This is where stock markets come into play.  The company may decide to do a “public offering” where they sell a large chunk of stock to an investment bank who then facilitates providing this stock to the general public.  This helps the CEO raise the cash he may not have been able to with dinner buddies alone, and it also allows anyone in the general public to become an investor -- no yacht-club memberships required!

So, how do high-frequency trading firms play a part in this mechanism of helping companies raise capital?  This is through a somewhat indirect effect.  Take the traveller’s dilemma from before, where converting from dollars to euros back to dollars results in a slight loss.  When converting is more efficient, it encourages people to make the conversion.  Similarly, because high-frequency trading firms make trading extremely efficient for everyone by promoting cheaper conversion from cash to stock and back again, it encourages people to make the conversion.  It’s important to keep in mind how a crucial factor dictating whether a company can raise $15 million, $20 million, or $25 million in a public offering is the appetite of investors.   Some of those investors may be enamored with the company and invest regardless of any other facilitation.  However, some investors are skittish and need a nudge.  Imagine a company selling a new gizmo.  It costs $299 to buy one of these gizmos, and returns are not allowed.  Suppose a thousand people are eager to take that deal.  Now imagine the company allows returns with a 35% restocking fee.  Perhaps an extra hundred people who were on the fence now will take the deal.  Now imagine the company allows returns with a 0% restocking fee.  Perhaps an extra five hundred who were on the fence now participate.  In the case of a public offering, if the company can raise $25 million instead of $20 million because a few more thousand investors who were on the fence now decide to buy on the confidence that if they have buyer’s remorse they can punt their investment back to the stock market quickly and efficiently, it means $5 million more to hire employees, expand facilities, et cetera.

The impact of not having high-frequency trading is apparent by looking at investing opportunities that lack their participation.  These are “low-liquidity” stocks which seldom trade, “pink sheets”, and private investments.  Without high-frequency trading firms providing clarity over what the stock price is, accurate to less than a tenth of a percent of the market value, and the relevant glue to make them ebb and flow with their related peers, these low-liquidity stocks exist in a sort of neverland where they may stagnate in price while their peers all go up.  The most disadvantaged in this example are not wealthy investors who can patiently wait for their investment to turn a profit but investors in dire financial straits who need cash.  A wide swath of hollywood movies have the iconic image of a lad down on his luck pawning off precious items for fractions of their true worth.  If pawn-shops had the spreads offered by high-frequency trading firms, the lad down his luck would get the same price for his autographed Mickey Mantle baseball as a Christie’s auctioneer, and be able to buy it back at-cost whenever he’s ready.  It all comes down to spread and being able to revolve cash to an item back to cash and barely lose anything in the process.  It encourages risk-taking, and this in turn puts money back into the greatest risk-takers of all, the American middle-class.

Copyright 2012  thoreaulylazy.  All Rights Reserved.

Monday, February 20, 2012


Internet Poster's Question:
I am not a economist, but just a simple question: What does India really make to warrant its economic growth? Only thing that I have experienced is the frustrating call center service from India. Don't think India really export any agriculture, natural resources, or anything else high tech. So really, what is India doing to generate wealth?

My Answer:
GDP does not equal the trade imbalance, or even exports. It is a measure of goods & services provided. Keep in mind that Earth does not export anything to Mars, Pluto or other planets, and yet you would agree that us Earthlings have become wealthier from 4000BC til present day. Earthlings enjoy televisions, internet, cellphones, and all the comforts we know despite never exporting to Mars and Pluto. That is because wealth doesn't merely come from exporting, it comes from producing. India has a large domestic industry, producing a multitude of goods, as well as various services like tailors, spas, restaurants, etc. to a population that equals 1.2 Billion people, the same as Earth's entire population in 1850.

Is India self-sufficient? No. It lacks one key resource: oil, which is by and far India's largest import. The largest export industries for India are textiles, agriculture, steel, automobiles, pharmaceuticals, and the smallest of its exports are its IT related services like call centers and software. Will India improve its exports? Yes, slowly. Why does India have tiny exports? Part of it is that India did not open its economy until the early 1990s; China, by contrast, opened its economy in the late 1970s / early 1980s as part of Nixon's ping-pong diplomacy. A closed economy does not trade very much with outside economies, usually achieved with high tariffs to discourage trade, and restrictive rules regarding foreigners opening shop in the country.

Why were China and India once closed economies? Security. A closed economy is immune from sanctions and other trading decisions of a foreign market. An open economy can become (dangerously) reliant on some key component, say, rubber, that can be turned off by the foreign supplier at any time, or used to extort the fledgling nation. During WW2, the allies stopped trading with Germany, and Germany suddenly lost key items like rubber and helium. This led to them trying to use hydrogen instead of helium on their airships, and that led to the Hindenburg fire. Also, foreign companies can engage in hostile takeovers of domestic companies, or product dumping, and this also leads to a security problem. For example, Domino's Pizza in the U.S. was known to sell a large pizza for $5 in a market that ordinarily charged $12, a ridiculously low price that lost them money; however, this ensured all competing pizzerias in the area went out of business, after which Domino's Pizza would raise their prices to $20. For a fledgling nation with small companies, the potential for a larger foreign company to use this practice was a threat.

Why did they open their economy? Trade imbalance. Despite being a closed economy, certain items, like oil, needed to be imported. These imports need to be paid in a currency the seller desires. Oil is sold in U.S. dollars, and the only way countries like India and China can acquire U.S. dollars is by (i) exporting, (ii) foreign direct investment (FDIs), or (iii) borrowing. Borrowing only delays the problem, since if they borrow U.S. dollars, they owe more U.S. dollars at a later date. FDIs are when, say, an American company wants to create an Indian subsidiary. All the Indian construction workers to erect the offices, employees, and local computers and other goods, need to be paid for in rupees, the local Indian currency. The American company exchanges some of its U.S. dollars for Indian rupees to achieve this in order to invest rupees into their Indian subsidiary. This currency exchange leaves India with some U.S. dollars and the American company with Indian rupees. India can then spend those U.S. dollars to buy oil. Exports are the simplest to explain: Indian companies sell some item to the consumers in the U.S., and are paid U.S. dollars. The U.S. is actually a really bad exporter, as our trade deficit shows, but the U.S. dollar can be spent to buy oil, which is by and far the most valuable product sold for U.S. dollars.

China actually has a problem in that it has too much U.S. currency. How on earth can you have too much? Because U.S. currency can only be spent in certain ways, such as buying oil, and China acquires more U.S. dollars each year than its need to buy oil. The excess money it then re-invests into U.S. treasuries, earning interest, since it can't spend it. The U.S. treasury interest is only 5% whereas China's domestic market is growing at 14%, so it would much rather invest that money for higher growth than 5%, but can't.

This is also why the world watches with bated breath as several oil-producing nations have been grumbling about the weakening U.S. dollar and contemplate selling oil for other items, like a basket of currencies. Iran, under recent trade sanctions by the U.S. and Europe, is rumored to be in negotiations to sell oil for Chinese yuan. If oil is no longer U.S. dollar denominated, it radically changes the global trade dynamics, since much of the value of the U.S. dollar is in buying this key commodity. It is also why the U.S. has such huge vested interest in the middle-east.

I hope that helps give an introduction to world economics and currencies.