BOOK REVIEW Flash boys can't match central banks Flash Boys - Revolt on Wall Street by Michael Lewis
Reviewed by Chan Akya
I pre-ordered Michael Lewis's new book, Flash Boys - Revolt on Wall Street on my Kindle, and proceeded to read it within the next two days. As is typical for Lewis, a combination of oddball characters, wonderful prose and the gentle tugging on the strings of an underlying mystery all featured, making the book an easy read. It is also a fairly short book compared with previous efforts by the same author.
In telling the story of Brad Katsuyama and his motley crew of telecom engineers, disenchanted traders and clueless bosses in Canada, Lewis uncovers the "scandal" around high-frequency
trading, or HFT, wherein customer orders on one exchange are quickly front-run on other exchanges and lead finally to poor price execution, costing investors in the millions as profits go into the pockets of the HFTs, who in turn pay various US exchanges and Wall Street banks for privileged access to trading platforms.
The basic construct of the scandal is as follows:
Wall Street banks need to ensure "best" execution for all customer orders. They therefore only trade at the lowest price/highest price (for buy and sell orders respectively). There are multiple exchanges across which to do the above trades.
HFTs seed various exchanges with minimum order trades at market prices, and whenever they get "hit", that is, either shares are bought or sold, an algorithm goes to work trying to figure out the size of the order. The job of the algorithms to figure out such information is made easier by the predictability of traders on desks of Wall Street banks and their customers, for example in terms of timing large purchases and sales.
Once the algorithm figures out what to do, it instructs the HFTs' other computers, which then race to other exchanges in the US and front run the order - that is, place their own orders before that of the customer.
The customer's order meanwhile takes longer to propagate across the system and so arrives later to any potential (seller) for a (buy) order - in effect, the HFTs have already mopped up all the available stock by the time the customer order arrives.
The net result, for example for a buy order, is that prices jump up quickly and in keeping with the rules of large orders, are often executed at the highest prices.
"Dark pools" which are operated by Wall Street banks to try and match sellers and buyers without communicating orders to the wider market (and thereby pushing prices) are corrupt constructs, which, thanks to the role of HFTs, leak information all over the place.
The math, algorithm and technology involved is for nerds only, but Lewis provides an accessible cover for the same with his simple language and lucid prose. Many a time, I laughed out loud while reading the book.
While the story is well constructed and narrated, I couldn't help but feel that Lewis has missed a very important element, that is, the role of HFTs in pushing up the market, and what that means for any down leg.
Let's consider that carefully: in any secular uptrend in markets, ie as macro-economic numbers improve, earnings are better and liquidity is stronger, the above scenario would inevitably cause prices to move up more than they would have otherwise, and in turn cause a virtuous cycle of stock market increases. The general construct of that would be:
1. For a given index level of say 1,000, assume one of the stocks is at 100. An investor places an order to purchase 1,000 shares at say 100, which is the current market price.
2. There is an existing seller of 1,000 shares at 100. In a "normal" market without intervention and commissions, the seller and buyer would cross their positions, 1,000 shares would trade at 100, and there would be no change in the index.
3. However, in the HFT world, the information of the buyer placing an order for 1,000 shares would have leaked, so the HFTs would race to the seller and mop up his 1,000 shares, thereby leaving no shares available for the buyer at 100.
4. The HFTs can then mark up the shares, say to 100.25, and sell to the buyer, costing him $250 in the process (compared to the "normal" scenario).
As the HFTs have taken no risk (being already aware of the buyer wanting to purchase 1,000 shares), their profit is in effect a tax.
However, there is another tax here: as the transacted price is now 100.25, the index will be pushed up, and the good old boys at CNBC will announce that 2,000 shares traded and there is "a lot of interest".
We already know, from the crash of 1987 onwards, that various buy-side firms have their own algorithm-driven trade computers, which place orders automatically. So when the price is pushed to 100.25, they would go in and buy say another 1,000 shares at 100.25 - go back to steps 3 and 4, and the execution is at 100.50.
In this feeding frenzy, its easy to bump up the stock prices to 105, with a strong effect on the broader index as well.
Now step back and think of the above steps when the first order is a sell, not a buy. It is possible, thanks to step 3, that the downward pressure on the stock may see it quickly crash to 95 in the same scenario. Thus, it is easy to figure out that the HFTs could well be the "amplification mechanism" that we have been failing to understand in the stock market for the past five to six years.
But is it all so simple? We also have to consider another factoid: that investors tend to be more active (viz buying and selling more) in a rising market than in a falling market. In other words, prices change more slowly on the way up than on the way down when fewer trades can push them down.
Thus the question begs, which Lewis doesn't answer: what is the role of the Wall Street strategy departments to inflate the prospects of the market, get more investors involved, and drive profits to the HFTs because the banks also benefit from higher volumes?
Are the banks guiltier than what Lewis makes them out to be? In other words, has Lewis missed a bigger story here?
And what about the central banks?
Then there is the whole other side of the "scandal" that is the quantum. Quite frankly, the few billions that Michael Lewis accuses the HFTs of pilfering is mere bagatelle compared to the very real tax on global financial assets that has been imposed on all savers by the central banks since 2009.
Lets consider a few examples:
a. The US Federal Reserve has been buying US Treasuries but not in auctions, rather in the "secondary" market from broker-dealers who are authorized to participate in the auction. Assuming that Wall Street banks are able to pass on 0.1% on each trade, the $2.5 trillion or so purchased by the Fed would alone have put $2.5 billion into the pockets of various Wall Street banks.
b. In addition to US Treasuries, the Fed has also purchased $1 trillion of mortgage-backed securities (MBS). This has in turn put another $2.5 billion into the pockets of banks.
c. The biggest impact though is on prices of MBS in general thanks to the Fed mopping up supply; conservatively, this has added 10% to the prices of various securities (or put differently, reduced yields on such assets by 1% per annum, assuming duration of 10 years). This is a good $100-$500 billion in terms of subsidy to people who owned the assets at the beginning of the financial crisis - or, put differently, a tax of the same amount on the mutual funds, insurance companies and other investors (including foreign central banks such as China's State Administration of Foreign Exchange and the Bank of Japan).
d. The "whatever it takes" policy of the European Central Bank (ECB) has been a boon for holders of various peripheral securities. For example, before the ECB intervened, a five-year Greek bond was paying a 5% coupon, but where investors demanded 20% yield to maturity due to default risks would have been priced at 53% of par value. After the intervention, when the demanded yield dropped to 10% after one year (ie. four years remaining) and the coupon remained at 5% (as the EU continued to expand its subsidies to the Greeks), the price of this bond would be an eye-popping 83% of par value.
That's a $0.30 increase for every one euro of bonds held, an indirect subsidy running into at least $50 billion per year on all the European peripheral sovereigns. Once again, this has been grabbed by private enterprises, namely the European banks.
There are other examples of central banks engaging in quantitative easing, such as the Bank of England and the Bank of Japan, both of which have easily put a few billion into the pockets of their banks every year since the subsidies began.
Then there is currency intervention, such as practiced by China. By keeping the yuan from appreciating too quickly, the People's Bank of China effectively funded a carry trade for various investors, thought to have given them some $10 billion in profits over the past five years alone. To be sure, this hasn't been a one-way street as the volatility in the yuan over the past month erased a lot of these gains. Still, it was there for a while.
A quick total of the above shows a number well north of $150 billion. In that context, the "Flash Boys" are rank amateurs, hardly deserving a mention in front of their true masters, the global central banks. Lewis missed this completely in his book.