On Labor’s Right to Strike
Why do unions strike? And who benefits?
You are reading Economic Forces, a free weekly newsletter on economics, especially price theory, without the politics. Economic Forces arrives weekly in the inboxes of over 14,500 subscribers. You can support our newsletter by sharing this free post or becoming a paid subscriber:
When it is my week to write a post, I generally try to stick to general price theoretical principles and economic arguments that I find interesting, insightful, and/or provocative. People often encourage me not to do this, but rather write about current events. Although I occasionally write about current events, it is difficult to do so. Readers tend to have opinions about current events, but our approach to interpreting current events is through price theory. Sometimes the insights of price theory conflict with people’s opinions. Maybe that is because they don’t understand price theory. Maybe that is because I am missing an important cost or constraint in my application of price theory. Or maybe people mistake price theoretic analysis for an ethical judgment. Regardless, I try to only write about current events when I think that there is some clear price theoretic lesson.
Thus, when news broke that the International Longshoremen’s Association was going to strike, I decided that it would be a great idea to write a post on labor strikes. If would give me an opportunity to do three things: (a) write about current events, (b) use some price theory, and (c) discuss a provocative price theoretic argument on labor strikes from one of my favorite economists, Earl Thompson. As luck would have it, the strike didn’t last long and the contract dispute was resolved during Brian’s week to write. As a result, I spent days thinking about what other topic I would write about. But then I had an epiphany. Writing about things either way before or long after the rest of the world cares about those things is actually my brand, so let’s discuss Thompson’s theory of strikes.
According to the conventional view, labor strikes are a tool used by unions to harm the firm (or industry) in order to extract higher wages and/or benefits. Thompson is skeptical of this claim. For example, in Thompson’s own work on bilateral monopolies, he argued that the side that could credibly communicate its price commitment and corresponding reaction function could extract most of the surplus. If one thinks about the negotiation between the union and a firm, for example, the firm seems to have a distinct advantage in being the first-mover to commit. The reason is that the manager of the firm (say, the CEO) is free to make such a commitment without taking a vote of the firm’s shareholders. By contrast, the labor union is typically democratic. Credible commitment is hard when one has to rely on coordinating membership behind a particular outcome. He also suggests that this position is evident in empirical analysis, which typically shows that management tends to win the negotiations. Thus, Thompson looks for an alternative explanation.
Thompson starts his discussion by considering the implications of a strike using price theory. Let’s first operate under the assumption that the union negotiates with the industry rather than one particular firm (I will then discuss when this makes little difference to the analysis). Thompson then considers the following example. Suppose the good that is being produced has a higher intertemporal rate of substitution (people are very willing to substitute their consumption across time) and suppose that the industry operates near capacity such that the short-run supply curve is very inelastic (quantity won’t respond much to changes in demand, but price will respond significantly). Under these circumstances, Thompson argues that the industry actually benefits from the strike. The reason is that the quantity of production is restricted over time, but that the total demand does not change because people are willing to substitute consumption now for consumption later (or vice versa).
In short, a strike allows the industry to act as though it is a cartel to restrict quantity to raise price above marginal cost (increasing its markup). Furthermore, since the industry (or the firm) is the credible first-mover, it can effectively induce a strike by beginning with an insulting, strike-inducing lowball offer. Finally, he also notes that even in the case when unions are not negotiating with the industry and rather the firm, the industry still effectively acts as a cartel as strikes tend to happen firm-by-firm with the net effect of the strikes on the industry being the same.
One notable observation that Thompson makes is that the ability of a labor union to negotiate pari passu with an entire industry arose in the U.S. in the 1930s. He notes that the timing seems significant. In the decades prior, the creation and enforcement of antitrust law eliminated traditional cartel behavior. The fact that industries acquiesced to such legislative changes in the 1930s might be indicative of their understanding of the cartel-like outcomes that striking behavior produces.
As I said, this is a provocative theory of labor strikes. It is therefore natural to ask what testable hypotheses are produced by the theory.
Perhaps the most obvious test of his theory is that it implies that strikes should be more likely in industries that typically operate at capacity and have high intertemporal rates of substitution for their product. Indeed, Thompson cites data on the frequency of strikes from the early- to mid-20th century in the U.S. and argues that strikes are most frequent in industries like mining, shipbuilding, long shoring, lumber, and textiles. At the same time, the industries with the lowest prevalence of strikes are those like agriculture and hotel and restaurant services. This certainly seems to fit with the theory. He also notes (writing in 1980) that examination of strikes from 1950 onward show that they tend to be concentrated in durable goods industries, which are characterized by high “pent-up” demand when supply is limited. He also notes that there are almost never strikes by wholesale and retail trades, which tend to operate with substantial excess capacity. He also notes the general pattern that strikes tend to be procyclical, implying strikes are more likely when firms are operating at capacity.
He also notes that strikes in agriculture start to become more frequent after significant innovations in refrigerated transportation. This is significant because “the cumulative demand for certain farm products over an entire production season or longer, has come to replace the weekly demand as the relevant determinant of farm prices.” In other words, his theory suggests strikes are more likely in durable goods industries and refrigerated transportation made agricultural goods more durable (and substitutable across time), at least at the margin.
He also considers alternative explanations for strikes. Some have argued that strikes come about as a result of informational differences between the industry (firm) and the union. However, if this was true, then past negotiations would provide valuable information to both sides. This should mitigate the informational differences over time. If that theory is true, one should expect the frequency of strikes in particular industries today to be negatively related to the frequency of strikes in the past. By contrast, his theory suggests that strikes should be concentrated in industries with particular characteristics. Thus, if Thompson’s theory is correct, the frequency of strikes in particular industries today should be the same or similar to the frequency of strikes in years past. Indeed, he examines the frequency of strikes in a “base” period and compares them to strikes most recent to his writing. He finds that in both the base period and the recent period, the industries with the highest frequencies of strikes are the same.
Another typical explanation of a strike is that it is simply a demonstration of the commitment and resolve of the union and its willingness to impose costs on the firm in order to extract higher wages. Thompson doesn’t deny this general idea, but rather argues that it does specifically explain strikes as the chosen means. Given the greater ability to commit to a wage offer and to communicate a reaction function, firms have the upper hand in the negotiations. Strikes evidently create better outcomes for workers by getting them a larger share of the surplus than they would otherwise get, especially in comparison to the firm’s initial, strike-inducing low-ball offer. But a strike is not the only way to demonstrate resolve or impose costs on the firm. In fact, by striking, the union is imposing a temporary cost on itself by forgoing wages during the strike period. If the union wanted to inflict costs on the firm and demonstrate its resolve, it could do so without creating costs for itself. For example, the union could simply shirk their responsibilities, work slowly, and/or make deliberate, costly mistakes. Instead, they strike. Thompson’s argument is that his theory can explain why unions resort to strikes: striking benefits both parties in comparison to the relevant counterfactual.
Thompson is also quick to note that his theory is a theory of strikes; it is not a theory of unions or union behavior generally. In fact, he argues that unions might provide some of the roles that people incorrectly attribute to the action of a strike. For example, the increased prevalence of unions might reduce information asymmetries and in the process raise the share of the surplus that goes to union workers, independent of a strike. Unions also provide other valuable services to its members independent of wage negotiations, like legal and information services.
Thompson’s theory is now 40 years old and likely deserves a new empirical examination. Regardless, it is provocative, topical, and grounded in price theory — something Thompson did well.