Portfolio Trading
Neil Chriss

This lecture focuses on equity portfolio trading (also known as program trading), a business whose primary objective is to provide customers with liquidity in trading baskets of stocks. A basket of stocks refers to a group of stocks, such as a stock index (e.g., the S&P 500 is a basket of 500 large cap US equities). A basket can be bought or sold, or sold short. Customers have numerous reasons for wanting to transact in basket trades. Among them are:

Basket trading is sometimes also called program trading or portfolio trading. Whatever its name, it generally refers to the service of offering large (e.g., large dollar value) trades in many stocks simultaneously. In what follows we will discuss what services a typical program trading desk offers, what the risks to the desk are and a little bit about how the trades are accomplished.

This lecture is divided into several parts. The first part discusses what program trading is, and the types of trades a desk will do for a customer. The second part discusses how business is conducted and the third part discusses market volatility and market impact. Before we begin with any of that we have to introduce a few useful pieces of terminology:


Types of Program Trades

Program trades can be broken into a few different categories, according to the costs and services offered to the customer. The different types of trades are:

The Business of the Business

Program trading is a customer driven business in which sales people have relationships with pension funds and money managers. These money managers are in constant need of new market exposure or changing old market exposures. Sales people often bring business to portfolio strategies groups who help money managers devise particular portfolios for particular purposes. These strategies can only be carried out by putting on the positions suggested, and this usually involves a basket trade.

At the point a basket trade is in the offing, there is a fundamental problem. The customer has a position he would like to buy or sell (or both, as in a long-short basket), but for many good reasons would not like to reveal the contents of the basket until such time as the prices have been agreed upon. This is the case for several reasons:

On the other hand, the dealer has to know as much as possible about the contents of the basket in advance of trading for several good reasons:

The obvious conclusion of these two points is that no two $100m (or $500m) portfolios are alike. Yet despite this, for a principal trade the dealer has to offer a price per share to take the basket and even for an agency baskset these are issues. As a consequence of this, most firms have resorted to the following compromise.

Program trading desks are willing to forego specific knowledge in exchange for generic knowledge about the basket they are bidding on. An example of the type of information that a dealer will require is;


Number of shares

    > 5 million


    1 million - 5 million


      500 - 1 million


    100,000 - 1 million


      < 100,000


Exercise: explain in detail why each of the above points is useful for a trader to know. What other facts about a portfolio might you want to know?

Principal baskets are usually usually sold to the lowest bidder, in the following sense. A customer wants to sell a principal basket and naturally wants to pay the lowest price. That is, if the customer is selling the basket, they want to sell it for as few cents below the market close as possible. If he is buying a basket he wants to pay as little a premium above the market as possible. To help esnure this, the customer usually gives the portfolio (in the sense that he gives the above information) to several program trading desks and asks for a bid. Naturally, each desk attempts to win the bid (if they want it) by bidding as low as possible. Ultimately, the desk would like to win business while still making money, so the bidding process requires the desk to deduce the cost of trading the basket. This leads us to the final section of this lecture.

Portfolio Trading Strategies

When a program trading desk engages in a risk bid, their profit is the price per share they bid less the cost of trading. In an agency trade, these costs are borne by the customer, but the dealers performance is measured in terms of this cost, and future business depends on good performance. There are three basic forces at work in portfolio trading:

As a consequence of the latter two bullet points, portfolio trading is a balancing act between the cost of immediate liquidation (resulting in maximum market impact) and longer term liquidation (resulting in market exposure).