Main
According to calculations by BTU researchers, employee earnings accounted for 98.1% of the nominal increase in Georgia’s average annual labour cost per employee between 2021 and 2025. Understanding the increase therefore requires examining pay alongside the work produced, while a declining share of other costs does not establish that their amount fell.
Key Takeaways
- Georgia’s average annual labour cost per employee rose by GEL 10,528.45 in nominal terms between 2021 and 2025; BTU researchers calculate that earnings accounted for GEL 10,328.64, or 98.1%, of this increase.
- Geostat reports that other labour costs declined from approximately 3.0% of Georgia’s total labour cost in 2021 to 2.6% in 2025, but BTU researchers calculate that their annual amount per employee increased from GEL 488.52 to GEL 688.33.
- Georgia’s 2021–2025 comparison reveals the composition of employer costs; it does not measure an individual employee’s pay rise or labour cost per unit of output.
A higher employer cost per worker may mean better pay, but the total cannot tell us exactly what changed. Did earnings rise, did additional services expand, or did the composition of employment change? Geostat’s release of 30 October brings this question into focus with data for 2025. Its detailed spreadsheet, including the earlier comparison year of 2021, allows the increase to be separated into components. [1, 2]
Average annual labour cost rose from GEL 16,142.76 to GEL 26,671.21. This represents a nominal increase of 65.2% over four years, not a rise during 2026 or a single year. The composition is more informative: almost the entire additional amount appears in the earnings category. Describing the finding only as “workers became more expensive” obscures the distinction between pay, other employment costs and the cost of producing actual work.
What employer labour cost includes
Geostat’s spreadsheet separates labour costs into earnings excluding time not worked, remuneration for time not worked, and other costs. The first two form the broader earnings component. Omitting paid leave or other paid non-working time would artificially reduce the earnings share. The calculation in this article therefore combines both categories. [2]
The official release describes other labour costs as social security expenditure, vocational training, cultural and educational activities, recreation and other activities. This category is neither a training budget alone nor simply a tax. Its aggregate amount can be analysed, but changes in individual components require additional evidence. A national average also cannot describe one company’s employee benefits package. [1]
Earnings are not take-home pay. Geostat’s earnings methodology uses accrued gross remuneration before personal income tax and includes supplementary and in-kind payments. Translating a rise in employer costs directly into a household’s disposable income would therefore be incorrect. Employment cost measures one side of the relationship; the amount a person retains and what it can buy are separate questions. [3]
Earnings account for almost all the increase
BTU researchers used the full-precision values in Geostat’s electronic table. For each year, average total cost was multiplied by the relevant component share. The 2021 amount was then subtracted from the 2025 amount. Dividing the earnings increase by the total increase produces a contribution of 98.1%. This is an accounting decomposition: it identifies where the increase was recorded, not which economic forces caused earnings to rise.
Change in labour cost in Georgia
Nominal GEL per employee per year; 2021 and 2025
| Component | 2021 | 2025 | Increase | Share of increase |
|---|---|---|---|---|
| Total cost | 16,142.76 | 26,671.21 | 10,528.45 | 100.0% |
| Earnings | 15,654.25 | 25,982.89 | 10,328.64 | 98.1% |
| Other costs | 488.52 | 688.33 | 199.81 | 1.9% |
Source: Geostat [2]; component amounts and contributions calculated by BTU. Earnings include remuneration for time worked and time not worked.
The calculation separates two different percentages. In 2025, earnings represented approximately 97.4% of the level of labour cost. The 98.1% figure is their share of the increase between the two years. The first describes the composition of existing expenditure; the second describes the distribution of its change. Confusing these denominators would misstate the result. Monetary amounts in the table are rounded to two decimal places, while percentages were calculated from unrounded inputs.
There are also two movements within earnings. Annual remuneration associated with time worked increased by approximately GEL 10,040.10 per employee; remuneration for time not worked rose by GEL 288.54. Together they produce the total earnings increase. These amounts do not measure changes in working hours or leave duration: remuneration rates and the composition of employees can also alter the totals.
A smaller share can conceal a larger amount
The share of other costs fell from approximately 3.0% to 2.6%. That can initially look like a reduction in additional employer expenditure. Reconstructing the amounts shows otherwise: GEL 688.33 rather than GEL 488.52 per employee per year, a nominal increase of 40.9%. The share fell because total labour cost grew faster. A smaller share and a smaller amount are different claims.
This matters when discussing workforce development. A falling share of the aggregate category cannot establish that Georgian organisations reduced training. Equally, its larger amount cannot prove that training expanded, because another component may have changed. The statistics show the overall size of additional labour costs; the volume, quality and results of investment in professional development need separate assessment.
There is a practical budgeting risk here. A manager who attributes rising labour costs mainly to small additional programmes may misread the cost structure. The national average cannot decide a company’s response, but it shows why separating the company’s own costs into comparable components is useful. Savings from cancelling a programme should be assessed alongside any resulting change in service quality or workforce capabilities.
A common pattern across different activities
The dominance of earnings is not confined to one broad ownership group. Earnings account for 99.4% of the increase in average annual labour cost in the public sector and 97.5% in the non-public sector. Total cost increased by 67.4% and 63.2%, respectively. Earnings dominate both, but the groups differ in their occupational and industrial composition. This is not a public-versus-private comparison of pay for identical work. [2]
Industry figures also show considerable differences in total cost growth. Annual labour cost in information and communication increased by 95.1% between 2021 and 2025, compared with 85.4% in accommodation and food services and 38.6% in human health and social work. These are comparisons of individual activities, not employment-weighted contributions to national growth. The national increase of 65.2% cannot describe every Georgian employer equally well.
Annual labour cost in selected activities
Nominal GEL per employee per year
| Activity | 2021 | 2025 | Growth 2021–2025 |
|---|---|---|---|
| Information and communication | 25,561.69 | 49,869.45 | 95.1% |
| Accommodation and food services | 10,864.64 | 20,139.21 | 85.4% |
| Human health and social work | 16,162.50 | 22,401.40 | 38.6% |
Source: Geostat [2]; growth rates calculated by BTU. Selected activities illustrate differences, not industry contributions to national growth.
A lower increase does not automatically indicate a better position, and a higher increase does not automatically indicate a problem. Higher average cost can reflect different jobs, a shift towards better-paid positions, or higher remuneration for existing employees. Two annual averages cannot separate those possibilities. The 2021 starting point also requires care: initial and final observations cannot establish when the four-year change accelerated.
Pay needs to be assessed alongside productivity
Higher labour costs have two sides. For an employer, the number records a higher nominal cost of labour. For an employee, remuneration is a source of income. One party’s expenditure forms part of another party’s income. Treating rising earnings as an automatic economic loss is therefore as incomplete as declaring an automatic improvement in welfare. Either conclusion requires additional evidence about output and prices.
For a company, a central question is labour cost per unit of product or service. If real output per employee increased, higher annual remuneration may be less burdensome than the monetary total alone suggests. If output did not rise correspondingly, covering the cost may require more sales, savings elsewhere or a change in margins. These are possible mechanisms, not outcomes measured for Georgian companies by this spreadsheet.
Measuring productivity requires comparable output or value added, the corresponding employment or hours, and adjustment for price changes. The labour-cost table does not supply these elements together. It also cannot establish that labour costs caused inflation. The eventual effect on prices depends on demand, competition and other costs. An accounting explanation is not a substitute for causal evidence.
Employees likewise need to distinguish the average from their own experience. A larger share of better-paid jobs can raise the national average even if an individual’s earnings remain unchanged. Other workers may receive increases above the average. The table does not follow the same people over four years or show the distribution of earnings. Its growth rate cannot therefore serve as an automatic benchmark for an individual pay agreement.
The economic choice for Georgia
In BTU researchers’ assessment, the new evidence sharpens the discussion about labour costs. Sustainable remuneration in Georgian business needs to be connected to work outcomes, rather than addressed only through cuts in small additional expenses. Staff retention, the use of skills and better work processes are relevant questions for investigation; the statistics promise neither a measured effect nor a universal solution for any of them.
For public organisations, a higher annual labour cost is not automatic evidence of better services either. Assessment needs to consider service volume, quality and the time citizens wait for a result. Private companies need an equivalent connection to production or service delivery. Across different organisations, the economic question is similar: what do society and customers receive in return for higher expenditure on labour?
The opportunity lies in separating an organisation’s own earnings and other costs, then relating them to corresponding outcomes. The main risk is applying the national average indiscriminately. The data establish neither a 65.2% increase in every salary nor an identical rise in every firm’s costs. Relevant industry and occupational evidence is more informative than one number used to characterise the entire labour market.
Conclusion
Most of Georgia’s increase in average annual labour cost between 2021 and 2025 was recorded in employee earnings. BTU’s calculated contribution of 98.1% identifies the composition of that increase, while the 40.9% nominal rise in other costs demonstrates why a declining share does not mean declining expenditure. The next analytical task is to connect cost with work outcomes, without presenting an average as an individual pay rise or a verified productivity change.
Questions and answers
Does this mean wages rose sharply in 2026?
No. The new release concerns 2025, and the main comparison is between average annual costs in 2021 and 2025. This calculation does not measure wage changes in 2026.
Why did other costs increase while their share fell?
Total labour cost grew faster. Other costs increased by approximately 40.9%, but their share of the larger total fell from approximately 3.0% to 2.6%.
Can the GEL 688 amount be treated as training expenditure?
No. This average annual amount per employee covers the broad other-labour-cost category, including social security and training. Training alone is not separately identified.
Does the analysis show worsening business competitiveness?
No. That assessment requires comparable output, productivity and other conditions alongside labour cost. This calculation describes the composition of the cost change.
Reference periods: 2021 and 2025. Release date: 30 October 2026. Verified on 2 November 2026. Amounts are nominal Georgian lari. Prepared for editorial review by the BTUAI Research Team.