the cost of ending relationships


One of the central themes of economic sociology is that relationships matter to market functioning. The theme is captured most succinctly with the “markets-as-networks” tag. Markets are just a special kind of network – one in which relationships facilitate trading of assets.

But sociologists are not alone in our efforts to analyze the relational aspects of markets. Economists and finance scholars are joining the fun, as can be seen in a set of papers assessing relationship effects on market exchange (e.g. Wilner, 2000). Sociological and economistic approaches to understanding the relational aspects of markets rarely speak to one another though, partly because of semantic differences or because we focus on different aspects of the relationships. Where sociologists see reciprocity, economists see monitoring. Where sociologists find embeddedness, economists find reputation. In fact, in many economics papers, the words “relationships” or “networks” never even make it into the text; instead they are supplanted with “reputation.” As a card-carrying sociologist, I see these as conceptually different (and operating according to different mechanisms), but I don’t want to quibble. My point here is that economists and finance scholars are actually doing quite a bit of network research that isn’t getting picked up by economic sociologists. Perhaps it’s time we pay attention.

For example, an article in the latest issue of The Journal of Finance (June, 2007), “Reputation Effects in Trading on the New York Stock Exchange ,” argues that relationships between traders and brokers in the New York Stock Exchange lower trading costs. Securities are traded in specific areas on the stock exchange floor. Specialists and floor brokers working in those areas get to know each other pretty well over time, perhaps not so well that Uzzi would cast their relationship as an embedded tie, but nevertheless the traders get to know each other well enough that accountability develops. In the words of the finance scholars, repeated face-to-face interaction “disciplines” traders from using private information to make trades at unfavorable prices. Over time brokers and specialists develop reputations for being fair and honest and this reputation causes them to self-monitor. They trade fairly in order to protect the reputations that they’ve earned through repeated interactions. As a result of discipline and reputation, relationships between specialists and brokers lower the costs of trading. Lacking relationships (i.e. the ideal-typical anonymous market), trading costs would be much higher.

The empirical evidence in the paper centers on observed location changes on the floor of the NYSE. When a specialist moves location (and brokers do not move also), the specialist has to develop a whole new set of relationships with brokers. The move, then, should lead to a temporary increase in trading costs. Costs will increase directly before the move, as brokers take advantage of the knowledge that the specialist will soon be gone and their relationship will no longer have any value to them. And costs will increase directly after the move while the specialist develops new relationships with brokers in the new location. This is exactly what the researchers found too. The evidence suggests that the bid-ask spread (a measure of trading costs) increases gradually beginning sixty days prior to the move and slowly returns to the same low levels fifty to sixty days after the move.

The study is interesting because it demonstrates the tangible costs of terminating relationships. In this case, the relationship is purely a market relationship (i.e. an arms-length tie), but even a tie of this type has real value. Just because you don’t spend weekday afternoons playing golf or don’t exchange Christmas cards doesn’t mean that the relationship (if repeated over time) will not develop some of the same predictable qualities of embedded ties. Repeated interaction creates reliability and forms a source of stability for markets that are needed to make the market work efficiently. The finding seems to support Wayne Baker’s (1984) study that showed that as options markets increased in size, they actually became more volatile. Given that there is a positive correlation between volatility and trading costs (demonstrated empirically in this paper), it’s possible that this may be one microexplanation for the observed phenomenon. Markets characterized by repeated exchange among a small group of actors may actually be more cost-efficient than the large, anonymous masses of the efficient markets hypothesis.

The study also demonstrates, rather nicely, that social networks help people lower the costs of doing business. Most network studies focus on the revenue side (e.g. getting a job), but as Joel Podolny maintains, the gains associated with certain structural properties like status or reputation may be mostly due to cost-cutting. High status investment banks hire the best employees at a lower cost than their low status competitors. Securities specialists with good reputations have lower trading costs.

The study also points to an important reason why social networks are so sticky. Actors simply can’t afford to find new acquaintances whenever they grow unhappy with their old ones. Cutting old ties and forming new relationships is costly. Of course, you probably have that figured out already.

Written by brayden king

June 5, 2007 at 6:58 am

Posted in brayden, economics, networks

One Response

Subscribe to comments with RSS.

  1. Brayden, thanks for referring to this paper. It is, as you say, very relevant to some of economic sociology’s central theoretical notions.



    June 10, 2007 at 11:33 am

Comments are closed.

%d bloggers like this: