Akholi Research · 2026

The wage floor stopped falling. The models did not notice.

Eleven delivery geographies. 4.2 million employment records. Every cost model in circulation assumes a convergence rate the data does not support.

+6.1%Actual annual real wage growth, tier-2 delivery cities

+2.3%Rate assumed by standard cost models

4 / 11Geographies where arbitrage is booked but not realized

14 / 20Largest economies facing a binding supply constraint by 2040

Modelled vs. realized real compensation, tier-2 delivery citiesExhibit 1

Index, 2016 = 100. Constructed from national employment microdata and firm-level payroll returns. Method: Appendix A.

Evidence

Annual real wage growth by delivery geographyExhibit 2

All eleven exceed the 2.3% modelled rate. The error has one direction, which points at the model rather than the market.

Wage trajectory against capability depthExhibit 3

Bubble area denotes headcount. Capability moved before cost did, and the receiving markets repriced accordingly.

#	Finding	Basis	Confidence
01	Real delivery wages converged faster than modelled	4.2m payroll records, 11 geographies	High
02	Attrition is the true cost line, not salary	Firm-level replacement cost returns	High
03	Capability moved before cost did	Role-level task decomposition	Moderate
04	The next decade is a supply problem	UN WPP medium variant, 20 economies	High
05	National skills forecasts share a directional bias	11 forecasts vs realized employment	Moderate

Diagnostic, target operating model, and sequencing for ministries of labor, finance, and education.

Location strategy and the honest total cost of distributed delivery for multinational employers.

Country assessments built on employment microdata rather than survey sentiment.

Standing counsel to multilateral and donor institutions on design, measurement, and evaluation.

Closed-door briefings for boards and cabinets on the forces reshaping their labor supply.

Legislative and regulatory testimony grounded in published, reproducible method.