For those of you that don’t know what HFT means, it is High Frequency Trading. I used to write a little code way back in the day, but I am by no means a proficient programmer. An HFT is an algorithmic trading program designed to take advantage of extremely short term price discrepancies in a given stock or future. My real contention is against stock based systems and I’ll tell you why. It takes a pretty in depth understanding of how markets truly work and I will try to get into that below. Let start with the basics.
Stocks, like futures, are bought and sold on an exchange. Futures have one exchange, yet stocks can have multiple exchanges and Supplemental Liquidity Providers (SLPs) which are nonexistent in the futures market. But more on that later. When you go to buy a stock, the market finds a seller so you can buy it from them. This is called the bid and ask. You buy at the ask price and sell at the bid price. It’s a little counter intuitive at first but it all comes back to the fact that if you buy, you buy it from a seller of the same security. What HFT systems do is “make a market” for a security. This means they are in the middle of a millions of transactions every day. It has been suggested that HFT systems comprise almost 80% of stock market transactions these days. When you go to buy a stock, through your broker, trading site, whatever, you buy it from a seller. That seller might be an HFT system, who’s aim is to sell the stock to you then buy it back at a lower price, creating a profit on the trade. Except these systems run virtually at the speed of light, buying and selling more shares than most of us could dream of affording every second of the day. Because of the predominance of these systems in the market place, when you buy a stock you invariably buy it a higher price and sell it at a lower price. I guess it’s your fault for not having a low-latency system co-located at the exchange so you too can take advantage of bids and offers coming in every microsecond.
What these systems are actually doing, though, is not simply exploiting existing discrepancies, but creating them. These systems run predatory trade techniques where in fact they will send out orders to the market to gauge buying and/or selling interest only to cancel the vast majority of those orders instantly. When you place your order, the HFT systems sees it, along with millions of others, and finds the best way to buy and sell shares to turn a profit. Sometimes this means bidding up a price microseconds before your trade hits the system so you but the stock at a higher price then shutting off the buy orders and letting the market fall back down. To most people it will go unnoticed, but when this type of trading becomes so widespread as it is now, it can make it almost impossible for a normal investor to survive. It is no wonder everyone is leaving the stock market, yet every day I see job opening for programmers looking to get into the financial markets. Normal people are leaving and are being quickly replaced by computer systems that could care less about housing data or inflation numbers. It has led to a completely broken and irrational market.
Quote Stuffing
This is probably the ugliest part of the whole HFT dilemma and sadly the hardest to understand. It all comes down to off exchange quote systems and the SLPs I mentioned above. Stock markets run electronically on bids and offers coming from many different locations across the world. In order to promote better pricing exchanges have opened up trading to these SLPs. Yet the HFT algo’s have found a way to exploit this system. When a one of the exchanges get inundated with requests for information it creates a lag in price discovery, sometimes even shutting off the system temporarily. I think the number is like 20,000 request for quotes (RFQ) per millisecond is the cutoff to give you an idea of how fast this shit is moving. Say a system, or better put 1000s of systems slam the main exchange with RFQs and halt true pricing from being displayed. In the seconds that follow these systems will then go to the SLP systems and slam the price a stock with sell orders running prices lower. They can then buy shares at a cheaper price than anyone else knowing that once the main exchange server comes back online, prices will be instantly higher. This in effect create the opportunity for mini flash crashes in stocks.
Algo's Gone Wild
Sometimes, a system or two, or millions, go off either due to bad coding or some unforeseen event like no more orders to buy at lower prices. This has happened many times and is in fact what happened to the stock market on May 6th but on a smaller scale. The last one I saw reported was in Nucor Stock on September 14th. The algos took out the entire bid side order book running share prices from $39.58 to $0.01 in a second. Now the SEC stepped in a cancelled the vast majority of the trades in that stock and reopened trading at $39.90 about 5 minutes later but you can see what kind of effects this has on a normal stock trader/investor. These wild swings are occurring more and more often and have led to an entirely broker stock market. It’s the same reason that a major report can have seriously negative implications while the market soars higher. The HFT systems see exactly what will cause the most amount of pain for investors and run the market to turn a profit.
In this recent stock market upswing volume has been abysmal. Meaning there are now real investors in the stocks, only systems designed steal money from retail traders. It is making it harder for traditional investments like mutual funds or seasoned traders to survive. It is no wonder we are seeing all investors leave the market. Everyone is deleveraging exposure right now, yet the only one increasing exposure is the US gov’t. I leave it to you to figure out what the end result will be.
No comments:
Post a Comment