Why hasn't the increased availability of information reduced the demand for labor market intermediaries?
The economics of labour market intermediation, by David Autor, Vox EU: The labour market depicted by undergraduate textbooks (e.g. Mankiw 2006) is a pure spot market with complete information and atomistic price-taking. Labour economists have long understood that this model is highly incomplete. Search is costly, information is typically imperfect and often asymmetric, firms are not always price takers, and atomistic actors are typically unable to resolve coordination and collective action failures.
In this ‘second-best of all worlds,’ numerous third parties inevitably arise to respond to, and profit from, market imperfections. I refer to these third parties as Labour Market Intermediaries – entities or institutions that interpose themselves between workers and firms to influence who is matched to whom, how work is accomplished, and how conflicts are resolved.
Labour market intermediaries have been around in various guises for centuries, as labour unions, craft guilds, and occupational licensing boards. But they have also gained prominence in contemporary incarnations as temporary help agencies, Internet job search boards, and centralised job-matching institutions like the ‘medical match’ that allocates physicians completing medical school to medical residencies. The venerable history and continued re-emergence of intermediaries in various forms raises the question: what precise economic function do these institutions serve? And are they likely to improve labour market operation or merely tax it?
Though heterogeneous, it is my contention that labour market intermediaries serve a common role. They address – and in some case exploit – a set of endemic departures of labour market operation from the first-best benchmark of full information, perfect competition, and decentralised price taking. Three labour market deficiencies, in particular, appear to ‘call forth’ the involvement of intermediaries: costly information, adverse selection, and (failures of) coordination or collective action. I give examples of each below. A unifying observation that ties these examples together is that participation in the activities of a given labour market intermediary is typically voluntary for one side of the market and compulsory for the other; workers cannot, for example, elect to suppress their criminal records and firms cannot opt out of collective bargaining. I argue below that the nature of participation in an intermediary’s activities – voluntary or compulsory, and for which parties – is largely dictated by the market imperfection that it addresses, and thus tells us much about its economic function.
In the textbook model of the labour market, there are no job search frictions. In reality, job search is costly. In addition to the costs of job advertisements, applications, and interviews, job seekers forego leisure while searching for work and firms forego production while holding vacancies. Information about job vacancies and job seekers is largely a public good and hence likely to be under-supplied by the market. This creates a business opportunity, which many labour market intermediaries step in to fill.
The leading contemporary example of an intermediary that provides job search information is the online job board (e.g., Monster.com, hotjobs.yahoo.com). Like their pre-Internet antecedents – such as help wanted advertisements, street corner day labour queues, and national job banks – online job boards are ‘information only’ intermediaries that aggregate, package and (in some case) resell information. Though I claimed above that compulsion is critical to the operation of most labour market intermediaries, it is apparent that information-only intermediaries exert only modest levels of compulsion: commercial job boards require users to post their CVs or job vacancies before they can view others’ listings; government-run job registries often stipulate that employers should post all vacancies. Why isn’t compulsion more forceful? The reason, I believe, is that the main challenge facing information-only intermediaries is free riding – all workers and firms would like to use the centralised resource, but none face full incentives to contribute to it. Since the cost to workers or firms of posting their own information is typically modest, the degree of compulsion needed to avert free riding is proportionately small.
Pure free riding, however, is unlikely to be the most significant market failure that results from costly information. Where information is incomplete, workers and firms face incentives to shade adverse information to raise their wages or profitability. This selective concealment generates negative externalities – adverse selection – that can depress the quality and quantity of trade in equilibrium (Akerlof, 1970). For example, if job candidates pad their resumes with bogus credentials or firms conceal hazardous working conditions, these actions reduce the equilibrium return to workers’ skill investments and firms’ provision of workplace safety. Mitigating adverse selection requires a mechanism that implicitly or explicitly compels market actors to disclose information that they would rather conceal. A set of labour market intermediaries appears purpose-built to perform this function.
A compelling example is the job search engine, AlmaLaurea, founded by a consortium of Italian universities. Distinct from commercial job boards like Monster.com, AlmaLaurea provides comprehensive administrative records – classes, grades, class rank – for essentially the full set of students graduating from the universities in the consortium, not just those applying for a particular job. This ‘full disclosure’ feature mitigates the risk of adverse selection that appears to pervade Internet search (Kuhn and Skuterud, 2004); employers screening candidates through AlmaLaurea face little uncertainty about applicants’ credentials or their standing relative to their peers. Do employers take this information seriously? Manuel Bagues and Mauro Sylos Labini find that graduates of universities joining AlmaLaurea during 1998 to 2001 experienced a several percentage point reduction in non-employment – an economically large effect – relative to graduates of universities that had not yet joined AlmaLaurea.
Employers as well as employees face incentives to exploit asymmetric information. At the turn of the twentieth century, U.S. job seekers—cross-state migrants and unskilled labourers in particular—made extensive use of private, for-profit employment agencies. These unsophisticated job-seekers were ripe for exploitation by agencies armed with vastly superior labour market information. Some of the abuses perpetrated by agencies included sending job-seekers to non-existent firms in distant locales; bribing firms to temporarily hire job-seekers so that the agency could collect a commission; and referring unwitting female job-seekers to brothels.
As exposited in an insightful paper by Woong Lee, U.S. state governments devised an ingenious response to these abuses. Rather than attempting to monitor for-profit employment agencies – a formidable task since many agencies were fly-by-night operations – states created their own public employment offices that offered high quality employment assistance at zero cost to job seekers. These free, reputable offices undercut the business model of the fly-by-night agencies – only for-profit agencies offering significant value added could effectively compete with the no-cost state agencies.
That state governments found it necessary to intercede to thwart abuses of asymmetric information underscores that purging the taint of adverse selection from the labour market requires a mechanism to force information disclosure or simply drive adversely selected agents from the market. In the case of AlmaLaurea, the mechanism was ‘full disclosure’ of the job qualifications of the entire graduating population of member universities. In the case of public employment offices, the mechanism was subsidised competition that placed a floor on the quality of services offered by private entrants.
Providing information – even compelling it – is not necessarily sufficient to resolve market failures. Notifying a bank’s deposit holders that the institution faces a small risk of insolvency does not reduce the risk of a bank run – in fact, it raises it. In such cases, rational agents acting with full information and mutually consistent expectations make decisions that are privately optimal yet collectively suboptimal – a coordination failure. There is potential in such settings for labour market intermediaries to improve upon competitive outcomes, but only if these intermediaries have the power to realign the maximising choices of agents with the common good.
One setting where coordination failures appear widespread is in entry-level labour markets for highly specialised professions, such as legal clerkships and medical specialties. Candidates in these fields are often compelled to sign binding employment contracts one or more years prior to the start of employment, well before the quality of job matches can be assessed. Such a matching process is likely allocatively inefficient.
What is the fundamental problem? Muriel Niederle and Alvin Roth argue that it is market congestion. In professions where an entire cohort of entry-level candidates enters the labour market simultaneously (e.g., medical school graduates), there is inadequate time for employers to screen all relevant candidates before competing offers have been made and accepted. To reduce the chance that they are forced to choose among the ‘leftover’ candidates, employers make early, exploding offers to job candidates. This naturally leads to market unravelling; anticipating that their competitors will make exploding offers, each employer accelerates its own offers. This harms allocative efficiency by making markets artificially thin – job candidates are forced to make major, irreversible decisions before knowing their full choice sets.
If congestion is the cause of unravelling, the problem could in theory be averted if employers were prevented from making offers until a sufficient search period had elapsed. The National Resident Matching Program (NRMP), studied by Niederle and Roth, performs this function. The NRMP is a centralised clearinghouse that allocates candidates to fellowships using deferred acceptance algorithms. Candidates participating in the match are not bound by job offers initiated prior to the resolution of the match – thus nullifying the power of exploding offers. Employer participation in the match is, however, voluntary.
After widespread unravelling, the gastroenterology profession (GI) adopted the NRMP in 1989. Analysis by Niederle and Roth suggests that the match ameliorated some major market maladies. The highly dispersed timing of job offers (reflecting early, exploding offers) was greatly compressed, and the mobility of GI residents out of the hospitals where they performed their residences (indicative of monopsony power) rose substantially. While more difficult to measure, the match probably improved allocative efficiency by better pairing fellows with fellowships according to intellectual fit and geographic preference.
Ironically, the GI match itself unravelled ten years after implementation. The cause of this failure is revealing. When the GI profession chose to curtail entry into the profession in 1996, the size of new cohorts fell much faster than expected – to the point where the there were fewer fellows than fellowships. This gave hospitals a strong incentive to jump the queue so their fellowship slots would not go wanting. The NRMP was helpless to respond to these employer choices since employer participation in the match was voluntary. Initial defections spurred further defections, and the NRMP became irrelevant by 2000. The unravelling of a labour market intermediary designed to inhibit unravelling again highlights that the power to compel participation by at least one side of the market – workers, firms or both – is necessary condition for an intermediary to address collective action failures. The GI fellowship match had this power when the labour market was slack but not when it was tight.
One might have speculated that in an era of rapid information flows and substantial job mobility, the importance of labour market intermediaries would wane. Indeed, the most prominent labour market intermediary, the traditional labour union, has been in secular declines for decades. Yet, the decline of labour unions is the exception rather than the rule. Two of the intermediaries discussed above – online search engines and centralised medical matches – have only recently gained prominence. And another labour market intermediary not even considered above, temporary help agencies, has risen from relative obscurity to international significance over the last two decades. (Figure 1 provides many additional examples.)
Why has the vastly increased availability of information on the Internet not reduced the demand for labour market intermediaries? The reason, I believe, is that cheap information alone is rarely sufficient to solve the failures endemic to labour markets. Redressing adverse selection and reversing coordination failures ultimately requires labour market institutions that can variously compel disclosure of hidden information, coordinate the actions of participants in a congested market, or solve collective action problems among parties with complementary interests. Despite widely heralded advances in the technology of job matching, I anticipate that labour market intermediaries will continue to address, ameliorate, and exploit the imperfect environment in which workers and employers interact.
Figure 1 Labour market intermediaries, by market function and type of participation
Market failure Nature of participation Information provision/search costs Worker-side adverse selection Firm-side adverse selection Coordination & collective action Voluntary for workers and firms Voluntary for firms not workers Voluntary for workers not firms Traditional board jobs X X Comprehensive board jobs (e.g. AlmaLaurea) X X Criminal record providers X X Public employment offices X X Labour standards regulations X X Centralised medical job match X X Labour unions X X Temporary help agencies X X
Akerlof, George A. 1970. “The Market for ‘Lemons:’ Quality Uncertainty and the Market Mechanism” Quarterly Journal of Economics, 84, 488-500.
Autor, David H. (ed.). Forthcoming. Studies of Labor Market Intermediation. Chicago: University of Chicago Press.
Bagues, Manuel and Mauro Sylos Labini. Forthcoming. “Do Online Labor Market Intermediaries Matter? The Impact of AlmaLaurea on the University-to-Work Transition.” in David H. Autor (ed), Studies of Labor Market Intermediation. Chicago: University of Chicago Press.
Kuhn, Peter, and Mikal Skuterud. 2004. “Internet Job Search and Unemployment Durations.” American Economic Review, 94(1), 218-232.
Lee, Wong. Forthcoming. “Private Deception and the Rise of Public Employment Offices, 1890 – 1930.” in David H. Autor (ed.), Studies of Labor Market Intermediation. Chicago: University of Chicago Press.
Mankiw, Gregory (2006). Principles of Economics, 4th Edition. South-Western College Pub.
Niederle, Muriel and Alvin E. Roth. Forthcoming. “The Effect of a Centralized Clearinghouse on Job Placement, Wages and Hiring Practices.” in David H. Autor (ed.), Studies of Labor Market Intermediation. Chicago: University of Chicago Press.
1 This column draws upon a forthcoming NBER-University of Chicago Press Volume, “Studies in Labor Market Intermediation.” The research cited below appears in this volume.