How Investment Banks Turn Pennies into Billions
Next time you log onto your online brokerage account to buy a stock, don't think there's a human from your investment bank on the other end selling to you. More likely than not, you're buying the stock from a computer that trades in and out of stocks millions of times a day.
Wall Street has been taken over by an army of computers that buy and sell stocks. Some sources estimate that nearly 70 percent of the trading on the major stock market exchanges is being done by computer programs. These programs, often referred to as algorithmic trading, program trading, or automated trading, are a big area of interest for many investment banking operations.
Computerized trading can be used for a number of reasons, including the following:
Serving needs of clients: Sometimes computers are employed to serve the customers of investment banks, helping them sell large positions of stock. Selling for big customers takes finesse — if all the stock is dumped at one time, the stock price can be pushed lower and cause the seller to reduce the proceeds. Investment bankers have systems in place to help them sell more gradually to avoid these problems.
Part of market-making responsibilities: Computerized trading may also be part of investment banks’ role as market makers. Investment banks certainly trade to try to make a profit, looking for chances to buy and sell stocks for a gain. But in some cases investment banks can also serve a secondary role.
When making a market, investment banks aren't necessarily trading to make a profit (although they probably don't complain when they do). They're the buyer of last resort, standing on the market looking to buy and sell when there are people looking to sell and buy. They're providing liquidity.
To make speculative bets: Investment banking operations may look to conduct proprietary trading, where they trade with their own money, using their own research. These strategies get pretty obscure, with investment banks typically looking to make a fast buck when different assets are mispriced, even for a few milliseconds, relative to each other.