Most Americans employed in the private sector do not have any job protection. They are what are known as "employees at will." They can quit or be fired at any time for any reason other than a reason forbidden by law, such as race. Unionized workers (now a very small percentage of the private-sector work force) have some job protection; they can be laid off if their employer experiences a fall in demand and therefore doesn‚Äôt need as many workers, but they can be fired only "for cause," normally some form of deficient job performance. In the public sector, most employees below the top political level have extensive job protection (including teachers), except in the military and other national-security employment, such as the CIA. Generally, civil employees of the government can be discharged only for cause, which often is very difficult to prove. The Supreme Court has largely abolished, in the name of free speech, the "spoils" system whereby state and local government jobs were given to the political supporters of the party in power. Federal judges can be removed (barring physical or mental disability) only by the cumbersome process of impeachment by the House of Representatives and conviction by the Senate. An important category of job-protected workers that bridges the public and private divide is tenured professors, who cannot be fired without cause. Finally, in Europe most workers have far more extensive job protections than American workers do.
The question I wish to address is whether this pattern makes any economic sense. One way to pose the question is to ask why--since employment at will is the cheapest form of employment contract--aren't all employees employees at will? In the otherwise dissimilar cases of unionized workers and public employees protected by the Supreme Court‚Äôs interpretation of the First Amendment against political firing, tenure (employment protection) is imposed from the outside. Employers would like greater flexibility, but outsiders--unions or judges--impose tenure for their own reasons. Unions worry that without tenure protection, employers will pick off the union's supporters; the Supreme Court worries that without tenure protection public employees will be afraid to express political views opposed to those of their superiors, and so freedom of expression will be curtailed. But surely the curtailment would be slight, since few public employees will engage in public disagreement with their superiors even if they can't be disciplined for doing so. Moreover, there is a tradeoff between professional competence and personal loyalty. A slightly less able employee who is loyal to his superiors because of political compatibility or even nepotism will work more harmoniously with them, and the reduction in friction may offset a (modest) competence deficit.
Tenure is an efficient system in what organizational economists call a "high commitment" workplace. Contrast two types of enterprise. In one, the contribution of the individual employee to the enterprise‚Äôs output is readily measured. Ordinarily this will be a business firm. Revenues, costs, and ultimately profits provide objective measures of performance. The individual employee's contribution to those measures may be more difficult to measure, especially when employees work in teams. But reasonable estimates are usually possible--employees and their superiors negotiate reasonable goals for the coming year relating to sales, markups, and cost reductions and progress toward those goals is measured throughout the year. Employees can therefore be paid a salary or wage that approximates their marginal product. With their productivity continuously measurable, there is no need for job protection.
Or so it seems; for even in a firm, there may be some benefits to providing a degree of job protection. Suppose employees are in a position where by sharing their know-how they could increase the productivity of other employees. They may be reluctant to do this if they fear losing their jobs because they have helped the other employees become more productive than they. Some firms deal with this problem by making an employee's annual bonus depend not only on his own contribution but also on the overall performance of the firm that year. This is a more flexible method than giving workers tenure.
The sharing problem is sometimes offered as an argument for how unionization might actually increase productivity. But it is a weak argument. If tenure is an efficient employment contract, employers will institute it without union prodding. The steep decline of unionization in the private sector is a convincing "Darwinian" refutation of the argument one used to hear that unions actually promote efficiency.
Although performance measures are generally most feasible for business firms, some governmental or other noncommercial activities lend themselves to such measures. Criminal-investigation agencies such as the FBI provide good examples. An FBI agent can be evaluated by the number of arrests he makes weighted by convictions (arrests that do not lead to convictions are not productive), with the convictions in turn weighted by the length of the sentence and the value of any property recovered as a result of the prosecution. Note that the measure here, as in a firm, is not a simple quantitative measure of contribution to output, but rather is a value measure.
In activities (some of which may be team production within business firms) in which performance measures are infeasible, usually because either the value of output or the employee's contribution to that output cannot be quantified, other methods of employee motivation than performance-based compensation must be sought. The "high commitment" workplace is a recognition that, fortunately, employees have other motivations for working productively besides the hope of salary increments, such as identification with the goals of the employer, as when judges and (other) civil servants internalize a "public service" ethic that induces them to work productively for a modest wage with limited hope of advancement. Tenure in such a setting both encourages sharing and discourages "influence activities," a term organizational economists use to refer to the kind of jockeying for position that occurs in the workplace when the absence of objective performance measures opens the door to worker competition based on personality, connections, and intrigue.
Even in a high-commitment environment, additional motivation may be provided by a tournament-style promotion system. Even if an employee's output cannot be measured with any precision, it may be possible to identify the best employee because the gap between his contribution and that of the next best may be large enough to be perceived without being quantifiable. Promoting the best employee to the next rank is therefore a method of incentivizing employees to do their best.
Both judicial and academic tenure are defended as needed to encourage independent thought and prevent political retaliation for unpopular views. This rationale is more persuasive in these contexts than in that of ordinary public employees, but it is not very satisfactory. In most nations, including nations that we consider our peers, the judiciary is insulated from political pressures but the judicial career is much like that of other employees. Judges start at the bottom rung of the judiciary when they are appointed and work their way up by impressing their superiors. The U.S. federal judicial system (also the British judiciary, and that of the other former British possessions) is unusual in being a system of lateral appointments (from practice or the academy, generally) with very limited promotion. The difference may be due to the fact that the Anglo-American and especially the U.S. legal system gives much more discretionary authority to judges than other foreign systems do, so that identifying the "best" for promotion is difficult and even arbitrary.
I do not think tenure makes a great deal of sense any longer in the academic setting, and I expect to see it gradually abandoned. (It has already been abandoned in England, for example.) If a university wishes to offer its faculty protection against political retaliation for unpopular views, it can do that by writing into the employment contract that politics is an impermissible ground for termination. Tenure is no longer needed because of an absence of performance measures. These measures exist in abundance. Quality of teaching is readily measurable by student evaluations, provided care is taken to prevent teachers from courting popularity by easy grading and light assignments and student evaluations are supplemented by faculty observation of the classroom. Quality of research is readily measurable by grants, prizes, and above all by citations to the professor‚Äôs scholarly publications, weighted by the quality of the journal in which the citations appear.
In some fields, such as mathematics, there is generally a significant falling off in academic output at a young age, and there is fear that without tenure these faculty would be turned out to pasture long before retirement age. But this is no different from the situation in professional sports, modeling, and other youthful occupations, where it is handled by an alteration in the wage profile. If a career in mathematics entails a sharp fall-off in market wages after, say, age 40, the academic market will compensate by offering disproportionately high wages to young mathematicians; otherwise, talented mathematicians will choose professions, such as economics, in which math skills are valued but productivity does not decline steeply with age.
One reason for the superior productivity of U.S. compared to European workers is that tenure encourages lazinesss by reducing the cost of laziness to the worker. But that is not the principal problem. Tenure removes the stick but not necessarily the carrot. More productive professors can be paid more and, even if their university has a lock-step compensation system, can obtain prestige and outside income by outstanding performance. The greater cost of tenure is simply in forcing retention of inferior employees. The 80-year-old mathematician may be working hard, but he may be incapable of achieving the output of the 25-year-old mathematician who would take his place were it not for tenure. Note how governmental prohibition of compulsory retirement at a fixed age aggravates the inefficiency of tenure--and is no doubt contributing to its eventual abandonment.
Perhaps the strongest argument for academic tenure is that without it academics would be reluctant to undertake promising projects with a high risk of failure. But the situation is no different in "knowledge" firms such as software and pharmaceutical-drug producers, which encourage their scientists to undertake high-risk projects--and do not think it necessary to offer tenure. If most good new ideas are produced by young academics, then an institution that raises the average age of faculty, namely tenure, seems likely to reduce academic productivity. An interesting empirical project, therefore, would be to study the effect of England's abolition of tenure on the average age and productivity of English university faculties.