Last time we talked about the basic structure of the New York Stock Exchange and American Stock Exchange, along with the basic workings of the specialist system used on those exchanges. Today I would like to cover the basics and give a brief history of the NASDAQ trading system, traditionally known as the over-the-counter market. This competing system has grown rapidly along with the growth in trade at the traditional exchanges, and a lively debate can always be assured when the conversation turns to which system is better.
    
Instead of a building with huge history and storied addresses, the NASDAQ system relies on a computer system linking a vast network of dealers making markets in one or more stocks listed on the system. The NASDAQ system essentially has three levels:

• Level 1 allows access to the composite best bid and offer for a particular stock. For example, when your broker calls up a quote on Oracle stock he or she may see a composite quote of $17.10 bid- $17.12 ask. What that tells you is that out of all the dealers making a market in Oracle stock the best bid is $17.10 and the best offer is $17.12.
      
• Level 2 allows the broker, or individual willing to pay for the service, access to the list of individual dealers bidding and offering, as well as the prices and sizes of those bids and offers.
      
• Level 3 allows those dealers actually making markets in the stock to access the system and list bids and offers. The traditional methodology for trading, then, was to access the composite quote for the best bid or offer, then proceed to the next level to find the individual dealer, and then execute the trade with that dealer.

The design of this system was thought to have several advantages over the exchange specialist systems, especially to those academics that look to promote increased competition in an industry. Instead of appointing one specialist with essentially monopoly power to set spread widths and control trading in a stock, literally anyone with the desire and the capital could become a market maker in any NASDAQ stock tomorrow. In the early 1970's much was written concerning the inherent monopoly power and unfairness in the single specialist system, really corresponding with similar attitudes and writings at the time directed towards trucking and airline cartels. Also, the early returns on the system reliability seemed to indicate that it was at least as reliable as the exchanges in regards to outages and occasions where stocks were forced to stop trading.

The disadvantages to the system were really threefold. The lack of a centralized market made it impossible for a stock to open at a single price, something routine and required at a traditional exchange. On the opening at the NYSE it can be said that IBM opened at one price, and that all market orders and marketable limit orders are filled at that price on the opening. The NASDAQ makes no such claim, and in fact multiple price openings are the norm. Broker A can sell Oracle for a client at the opening at $17.10, while Broker B, putting in an opening buy order can and probably will be filled at $17.12. The second issue is one of trade-throughs. A person willing to pay $120 for IBM on the NYSE will be filled before the stock trades at $119 7/8. In fact, if a large block were to trade at $119 7/8, he may be included in the block at the lower price. On the NASDAQ system there is no such guarantee. Even if we discount for a moment nefarious motives, even a simple trade could have involved a trade-through. Broker A is trying to buy 500 shares of Oracle at $17.10, when Broker B receives an order to sell 1700 shares at the market. Broker B looks at his level 2 system and sees 500 shares bid for at $17.10 and 3000 shares bid at a different dealer at $17.08. Under the original design of the system broker B was free to sell all 1700 shares at the $17.08 price, and was under no obligation to split the trade to take out the smaller $17.10 bid.

The third disadvantage of the system is the difficulty in executing a traditional market order. It is fairly simple and routine to send relatively large market or limit orders to the NYSE, where there really was no central place on the original NASDAQ system. The fulfillment of such an order really involved the communication with one or more dealers in an attempt to fill the order, with no guarantees that the remaining dealers would maintain their bids and offers once the process began. Other dealers were not above pulling their offers when they noticed the offers of others on the system being taken. Many times a large order from an investor or hedger from the options floor would leave the execution firm with the question "Where exactly do I go to fill this order?"
              
The NASDAQ system also did not solve the issue of various firms now acting with the same inherent conflicts of being both principal and agent as the specialist. In fact, in many cases it was much worse, with firms bidding lower for a stock than their customer was bid, and then selling it immediately at the higher price to the customer. That seems incredible, but it was a practice for decades. In a simple case, the stock would be initially $17 bid $18 ask, with a customer entering an order to buy at $17 ¾. The firm would then bid for the firms account at $17 ¼, and if filled sell to the customer at $17 ¾. Great oversight by the SEC, wouldn't you say?