When distributed work became normal at scale, a confident prediction followed: if work can be done anywhere, employers will hire wherever talent is cheapest, and wages will converge globally.
Several years on, that hasn't happened to the degree predicted, and the reasons say something useful about what actually constrains labour markets.
What did happen
Real changes, worth acknowledging before the qualifications.
Hiring ranges expanded within countries. Companies in expensive cities hired from cheaper regions of the same country at meaningful scale, which had genuine effects on regional labour markets and property prices.
Cross-border contracting grew substantially. Not employment — contracting — with individuals engaged as independent contractors rather than employees.
And employer-of-record services expanded, allowing companies to employ people in countries where they have no legal entity, by using an intermediary that formally employs them.
So the direction was real. The magnitude was smaller than predicted and the form was different.
Why full convergence didn't occur
Several constraints, each individually significant.
Legal and tax complexity. Employing somebody in another country creates obligations — payroll, social contributions, employment law compliance, and potentially a taxable presence for the company. That last point is the serious one: an employee working in a country can create a permanent establishment for tax purposes, with substantial consequences.
Employer-of-record services exist precisely to manage this, at a cost that erodes the wage saving.
Time zones. The constraint that receives least attention and binds hardest. Work requiring interaction with colleagues has a limited overlap window, and a gap beyond a few hours makes collaboration genuinely difficult.
This is why cross-border remote hiring has concentrated in compatible time zones rather than following the lowest cost globally.
Coordination costs. Distributed teams require more explicit communication and more documentation. Those costs are real and they scale with distribution, offsetting some of the wage differential.
Regulatory attention. Classification of remote workers as contractors rather than employees has attracted enforcement in multiple jurisdictions. Companies structuring international hiring as contracting face genuine risk if the working relationship resembles employment.
What actually determines the wage
The more interesting finding: wages for remote roles have generally tracked the employer's location more than the worker's, at least in the higher end of the market.
The reasons are competitive rather than charitable. If a company pays local rates in a low-cost location, its best workers are recruited by companies paying more. Since those competitors are also remote, geography provides no protection.
So the market for genuinely scarce skills has tended towards a global rate, set by the highest-paying employers, rather than converging downward.
The effect is concentrated in roles where skills are genuinely scarce and quality differences are large. In roles where work is more substitutable, downward pressure has been considerably more real.
The location-based pay debate
A live argument with reasonable positions on both sides.
The case for adjusting pay by location: cost of living differs enormously, paying uniformly means employees in cheaper locations receive far greater real compensation for the same work, and it allows a company to afford more people.
The case against: pay should reflect the value of the work, not the employee's housing costs. Location-based pay creates an incentive not to move and produces awkward outcomes when people relocate. And it's frequently applied asymmetrically — reduced for those moving to cheaper places, rarely increased for those moving to expensive ones.
Practice varies and there's been movement in both directions, with some companies abandoning location adjustment and others introducing it.
The distribution of gains
Worth stating clearly. The people who benefited most were those in expensive countries able to move to cheaper regions while retaining their salary.
Workers in lower-wage countries benefited where they gained access to higher-paying international employers, which is a genuine and significant gain for those who accessed it.
But that access has been concentrated among those with strong language skills, compatible time zones, established credentials and existing networks — which is to say the already advantaged within those markets.
The broad convergence that would have benefited the many did not occur. What occurred was a widening of opportunity for a specific segment, which is a real improvement and a much narrower one than was forecast.
The office question, settled unsatisfyingly
Several years on, most large organisations have landed somewhere in the middle, and the pattern is fairly consistent internationally: some required days, some flexibility, and a great deal of local variation by team.
What is notable is how little the arrangements correlate with any evidence. Organisations doing the same work in the same sector have arrived at quite different policies, which suggests the decisions were driven by leadership preference, property commitments and competitive positioning rather than by analysis.
Property commitments deserve particular mention. An organisation holding a long lease on expensive space has a financial reason to want it occupied, and that reason is rarely stated in the communication explaining why attendance is required.
None of which makes the arrangements wrong. It does suggest that the debate about productivity was largely a proxy for a set of decisions being made on other grounds entirely.