What a Shorter Job Stay Actually Tells Us
This chapter examines what current job duration can reveal about the labour-market response to expanding services trade. Its focus is the difference between a stock—the time a person has already spent in the job they hold—and the flows that explain what happens next. By separating those measures, we can treat shorter tenure as evidence of aggregate dynamism while testing whether the movement also involves unemployment, sectoral displacement, or more temporary contracts.
Current tenure is a snapshot, not a complete job history
A survey that asks how long someone has been in their current job records a stock at a particular moment. It can tell us how much tenure has accumulated so far. If the average is lower in a group with greater services-trade exposure, the group appears to be moving through jobs at a faster rate. That is a meaningful change in the pace of employment relationships, but it is not a record of every job that began or ended, and it cannot show how long the current job will eventually last.
This distinction matters because the same short tenure can be produced by different routes. A worker may have moved from one employer to another without leaving employment. They may have changed position within a firm, entered a growing occupation recently, or been displaced and spent a period searching for work. Someone else may have a permanent contract but only a short tenure because they joined the employer last month. The tenure measure places all of these people near the same point on one scale even though their security, bargaining position, and expectations may be very different.
The measure also does not identify the direction of a move. A change can preserve or increase a worker’s skills, autonomy, compensation, and future options, or it can involve the loss of firm-specific knowledge and a weaker match. A move between employers may be invisible to a broad sector classification if both jobs sit inside the same sector. Conversely, a change in the composition of a sector—more new entrants and fewer long-tenured workers—can lower its average without every existing worker becoming more mobile. The observed average is therefore the result of both individual histories and who is present in the group being measured.
This is the stock-flow distinction in practical form. A stock tells us what has accumulated among the people visible today; a flow tells us who entered, who left, and what route they took between two observations. If a fast-growing activity hires many recent entrants, its average tenure can fall even if long-serving employees are not being pushed out. If an activity loses established workers and replaces them with short-lived jobs, the same average can fall for a very different reason. The tenure statistic cannot separate those stories without transition data and a clear denominator.
The denominator matters as much as the average. The OECD regressions connect trade exposure to groups defined by sectors and occupations, rather than following one named worker through an entire career. That design is useful for identifying broad labour-market patterns, but it means that a result for a sector-occupation group should not be translated into a statement about the hazard faced by every person in it. The group can change through hiring, exits, occupational sorting, and changes in the mix of employers. An aggregate measure can therefore register a real change in the labour market while remaining agnostic about the experience of a particular worker.
The measure also leaves the comparison implicit. A claim that tenure is “shorter” means shorter than in a relevant comparison—another level of exposure, another year, or another group after the model’s controls—not that every job in the economy has become short. Reading the result responsibly requires keeping that comparator in view. It also requires asking whether the exposure is attached to the worker’s sector, occupation, or region, because each assignment describes a different route by which services trade may matter.
The services-trade result should be read at exactly that level. The OECD analysis reports that greater services-trade exposure is generally associated with shorter average current job duration in sector- and occupation-level regressions. It describes this as greater labour-market dynamism, not as a direct measure of involuntary displacement. The result does not establish that services trade causes every worker to switch jobs more often, nor does it tell us whether the next job is better. It gives us a reason to inspect the transitions around the current job rather than a reason to skip them.
The states in the diagram are not competing definitions of tenure. They are different observations that have to be placed around it. Contract form is a separate dimension attached to the current job; it does not describe the route by which a worker arrived there. A sector change is another possible transition, but a move within the same broad sector can still be a substantial change in tasks or working conditions. And a worker can pass through a brief period outside employment without that spell appearing in a comparison of two annual observations.
The transitions that make tenure interpretable
The first diagnostic is movement into or out of employment. If greater services-trade exposure were shortening tenure mainly because more people were being pushed into unemployment or out of the labour force, the share of workers making those transitions should rise. The OECD quasi-panel analysis does not identify a significant systematic relationship of that kind. This makes a simple “shorter tenure equals more unemployment” explanation less consistent with the measured pattern.
The result is narrower than it may sound. In the European labour-force data, people report their employment status and sector in the prior period, generally one year earlier. A person who leaves a job, experiences a short spell of unemployment, and finds another job before the next observation may appear employed at both points. The comparison can record neither the interruption nor its duration. Many entries into and exits from employment also have personal causes, such as education, parenthood, or retirement, which makes a trade-linked signal difficult to isolate. A null estimate means that the study does not detect a systematic change in the measured share; it does not certify an uninterrupted path for every worker.
The second diagnostic is movement between sectors. Sector switching can indicate that workers are following new opportunities, or that they are being pushed away from a shrinking activity. It can also require them to give up knowledge that was valuable only in the old setting. The analysis finds no significant relationship between services-trade exposure and the shares of workers switching into or out of a sector. The report therefore suggests that the additional job moves associated with shorter duration are primarily occurring within the same broad sector.
That interpretation has an important boundary. A broad sector is an administrative category, not a complete description of a person’s work. Moving from one employer to another inside “transport,” “administrative services,” or “information” may change tasks, schedule, pay, supervision, or prospects while leaving the sector label unchanged. The absence of a detected sector transition should thus be read as evidence against a broad reallocation story, not as evidence that jobs and skills remain unchanged.
There is also a difference between changing employers and changing sectors. The first can preserve a worker’s occupation while changing the firm-specific knowledge, manager, schedule, or prospects that shape daily work. The second is visible in the broad classification but may not tell us whether the worker moved because of a better offer or because the old activity was shrinking. A shorter tenure result with stable sector-switching shares is therefore consistent with a labour market in which employers compete for workers inside the same broad activity. It is also consistent with hidden changes in tasks that the classification cannot see. The evidence narrows the route; it does not narrate the whole career.
The transition data have another asymmetry. It is easier to observe a person’s status at two dates than to observe the sequence of decisions, offers, constraints, and interruptions between them. A worker may have searched while still employed, accepted a new job before leaving the old one, or temporarily stepped away for a reason unrelated to trade. A one-year comparison compresses those possibilities into an origin and a destination. That is adequate for asking whether a systematic change appears in a measured share, but not for assigning motive or welfare consequences to an individual move.
The third diagnostic is contract form. Temporary employment is an observable source of uncertainty because the contract itself limits how long the employment relationship is guaranteed to last. If services trade were producing shorter tenure by forcing workers into a succession of short-term positions, a rise in temporary-contract prevalence would be an expected companion result. The OECD report finds no effect of services exports or imports on the use of temporary contracts. Where services are imported for use as inputs, it finds an association with lower temporary-contract use.
This is the strongest direct reason to resist calling the tenure result a broad rise in contractual precarity. But a contract category is still not a complete account of security. A permanent worker can face intense performance demands, weak prospects, strained social relations, or uncertainty about future restructuring. A temporary worker can sometimes be using a short contract to enter a desirable occupation, although the available evidence here does not classify workers’ reasons or outcomes that way. Contract status helps test one pathway; it does not settle the quality of the job.
It is useful to distinguish two questions that are often folded into the word “temporary.” The first is contractual duration: whether the employer has promised an open-ended relationship or a fixed-term one. The second is experienced stability: whether the worker expects hours, income, tasks, and continued demand to remain dependable. The available measure speaks mainly to the first. It cannot show, for example, whether a permanent employee is under pressure because an employer may outsource the role, or whether a temporary employee has a predictable sequence of projects and a strong route into a longer-term position. Those cases cannot be inferred from contract prevalence alone.
This is why the four diagnostics work as a set. Tenure tells us that the population’s relationship with current jobs is changing. Employment-status transitions ask whether the change is accompanied by visible entry into or exit from work. Sector transitions ask whether the adjustment crosses broad economic boundaries. Contract form asks whether one recognisable kind of uncertainty is becoming more common. None of the four directly measures the quality of the next job, but together they prevent a single proxy from dominating the interpretation.
Taken together, the measures produce a pattern rather than a single answer. Current job duration is shorter. The measured shares entering or leaving employment do not rise systematically. Broad sector switching does not rise systematically. Temporary-contract use does not rise for exports or imports, and falls for one input-exposure measure. The pattern is compatible with more movement between jobs within broad sectors, including movement toward better matches. It is not proof that all movement is voluntary, upward, or harmless.
A disciplined reading of any future tenure statistic follows the same order. First identify the unit: are we looking at workers, sector-occupation groups, firms, or regions? Next ask whether the outcome is current tenure or an observed transition. Then check the routes—employment status, sector, and contract—before making a claim about insecurity. Finally, inspect the dimensions that tenure leaves out: wages, work intensity, physical and social conditions, autonomy, skills, and prospects. A short stay can be a signal that the labour market is rearranging; only the surrounding evidence can show what that rearrangement means.
The next question is how services trade reaches those workers in the first place. Exports, imports, and imported services used as inputs do not expose the same sectors or people, and the composition of the workforce can change before any individual transition is observed.