Shadow Price of Labour

Shadow Price of Labour in Health Programme Appraisal

The shadow price of labour is the shadow price an appraisal attaches to work when the wage paid does not show what society gives up by employing that worker. In the cost-benefit analysis of a health programme it replaces the wage of nurses, community health workers or volunteers with the value of what they would otherwise have produced, or the least they would accept in order to work. The European Commission and Asian Development Bank appraisal guides call it the shadow wage and apply it to the wage bill as a conversion factor. The gap is widest where unemployment is high or much of the work is unpaid, as in many community health programmes in low-income countries, while UK appraisal and the NICE reference case cost staff at what they are paid.

Why the wage can misstate the cost of health workers

The HM Treasury Green Book (2026) values goods and services by their opportunity cost, the value of the next best alternative use (paragraph 8.1). Market prices are the starting point, with the market wages of engineers as its example (paragraph 8.2). Paragraph 8.5 calls cases where market prices fail to represent social costs market failures, and a distorted labour market can be read in the same way.

The European Commission guide explains that wages can be a distorted signal when labour markets are imperfect, as shown by high and persistent unemployment or a large informal economy. Its examples include a legal minimum wage that some unemployed people would work below and informal work with no formal wage but a positive opportunity cost. Annex IV states the appraisal rule. The financial analysis values labour at the market wage, the economic analysis uses a shadow wage reflecting the social opportunity cost of labour, and the market wage usually overstates that cost, less often understating it.

Shadow wage rules by where workers come from

Both guides tie the shadow wage to what each worker was doing before being hired. In the European Commission guide, skilled workers previously employed in similar work have a shadow wage equal or close to the market wage. Unskilled workers who were unemployed have one no lower than the value of unemployment benefits or a similar proxy, and unskilled workers who leave informal activities have one equal to the output forgone there.

The Asian Development Bank guidelines describe the economic price of labour as the cost to the economy of employing one more worker. Scarce skilled labour is costed at the actual wage including benefits, and the openly unemployed at the reservation wage, the payment needed before a person takes work, usually set by family income or state benefits. Underemployed workers are costed at the output forgone in informal work, plus any housing, health or education provision needed for migrant workers that the project does not pay for.

A nurse recruited from another hospital leaves a post whose output is lost, so the wage is close to the opportunity cost. A community health worker recruited from subsistence farming gives up crop output that may be worth far less than the programme pays. A programme drawing workers from several origins has a weighted average shadow wage.

$$SW = \sum_{k} s_k \, OC_k$$

where $SW$ is the shadow wage per worker, $s_k$ is the share of workers in origin group $k$ (the shares sum to one) and $OC_k$ is the opportunity cost per worker for that origin, such as the wage given up elsewhere, the reservation wage or the informal output forgone.

Example 2 in Appendix 8 of the Asian Development Bank guidelines works one case. A project pays L120 a month, a 20 per cent premium over the wage elsewhere. Seventy per cent of its workers come from employers paying L100, and 30 per cent would otherwise be out of work and would accept L30. The shadow wage is 0.7 × 100 + 0.3 × 30 = 79, and the ratio to the wage paid is 79 / 120, about 0.66. The premium is not part of the shadow wage, since those workers produced L100 of output before.

Conversion factors and shortcut formulas

The guides express the shadow wage as a ratio to the wage. The Asian Development Bank calls this ratio the shadow wage rate factor, and the European Commission guide calls it the shadow wage conversion factor.

$$CF = \frac{SW}{W}$$

where $CF$ is the conversion factor, $SW$ is the shadow wage and $W$ is the market wage paid. Multiplying the labour cost lines of a financial appraisal by $CF$ turns them into economic costs.

Without national or regional estimates, the European Commission guide offers a shortcut that cuts unit labour costs by the share of income tax. For unskilled workers on project sites in a region with persistently high unemployment it adds an unemployment term.

$$SW = W \, (1 - t)(1 - u)$$

where $SW$ is the shadow wage, $W$ is the market wage, $t$ is the income tax rate and $u$ is the regional unemployment rate. Dropping the $(1 - u)$ term gives the simpler shortcut.

Annex IV reports regional estimates by Del Bo and colleagues from 2007 data, with a conversion factor of 0.99 in fairly socially efficient labour markets, 0.80 in urban dual markets, 0.62 in rural dual markets and 0.54 in regions with quasi-Keynesian unemployment. Their baseline formula weights a worker's previous and new social value by a regional welfare weight, which links the shadow wage to the distributional weights of weighted cost-benefit analysis. The Asian Development Bank guidelines say the national factor for skilled labour can normally be taken as 1.0 and that alternative factors for surplus labour can be tested in sensitivity analysis when local data are poor.

UK appraisal rules and the NICE reference case

Paragraph 8.26 of the Green Book defines the social cost of labour as the full-time equivalent cost of employees' time, including pension costs, national insurance, allowances, benefits and basic salary. Tax and national insurance paid from gross earnings are part of the value the workforce produces, so they are not transfers, whereas the European Commission shortcut takes income tax out of the wage. Paragraph 8.41 sets the principle of full displacement of jobs, under which every worker a proposal hires leaves another job, not unemployment or inactivity. Paragraphs 8.42 and 8.43 allow local exceptions only with compelling evidence signed off by the originating organisation's chief economist. Read together, these paragraphs make the full employment cost the shadow price of labour in UK appraisal.

The NICE manual (PMG36) is narrower. Section 4.4.1 states that reference-case costs relate to resources under the control of the NHS and personal social services and are valued at the prices relevant to them. Section 4.2.9 excludes productivity costs, and section 4.4.23 allows them to be presented separately when they may be a critical part of a technology's value. A nurse hour in a NICE cost-effectiveness analysis is therefore costed at what the NHS pays, and the opportunity cost that matters is the health that money would buy elsewhere in the budget, which the shadow price of health measures.

The contrast is mainly one of perspective, and it maps only loosely onto the welfarist and extra-welfarist split. The shadow wage belongs to welfare economics, which counts what the whole economy gives up, while extra-welfarism, in the account of Brouwer and colleagues, permits outcomes other than utility. They note that forgone health is the principal opportunity cost within a health care sector without being the sole opportunity cost of public sector resources. Shadow-pricing labour therefore sits with the societal perspective, not a fixed-budget health system view.

Volunteers and lost working time

Unpaid labour is where health appraisal most often meets the shadow wage. Kasteng and colleagues describe the replacement cost method, which uses a shadow wage equal to the cost of buying the same work on the labour market, and the opportunity cost method, which values time at the person's second best use, with a reservation wage as proxy for the unemployed. In their study of 45 unpaid community health workers in rural Uganda, a median workload of 19.3 hours a month was worth USD 6.9 a month at the replacement wage and a median of USD 4.1 at the workers' own opportunity cost. They argue that leaving unpaid work out makes a programme look more cost-effective from the provider perspective than from a societal one.

Turner and colleagues found that volunteer time in mass drug administration for neglected tropical diseases averaged between USD 0.05 and USD 0.16 per treatment across studies, and they valued the time donated to the African Programme for Onchocerciasis Control at USD 60 million to 90 million.

Working time lost to illness raises the reverse question. The European Commission guide notes that wage-based values of time and life assume the gross wage equals the value of the marginal product, and it suggests correcting the wage by the shadow wage under severe unemployment. The human capital approach values lost production at gross earnings. Pike and Grosse report that the friction cost method assumes a worker who leaves can be readily replaced, including by someone unemployed, and gives smaller estimates. In effect it assigns a low shadow price to a long-term absent worker's labour once a replacement is found.

Worked example with a vaccination outreach programme

All numbers are illustrative. A health authority in an EU region with high unemployment appraises a one-year vaccination outreach programme by social cost-benefit analysis. It employs 20 nurses recruited from other clinics, at EUR 30,000 each, and 30 outreach assistants recruited from long-term unemployment, at EUR 18,000 each. Non-labour costs are EUR 460,000. The programme gains 75 QALYs, valued here at EUR 20,000 per QALY.

1. Wage bill. Labour costs 20 × 30,000 = 600,000 plus 30 × 18,000 = 540,000, a total of EUR 1,140,000.

2. Nurses. They leave posts of similar skill, so the conversion factor is 1.0 and their cost stays at EUR 600,000.

3. Assistants. Applying the European Commission shortcut by analogy, with a tax rate of 0.20 and unemployment of 0.25, the shadow wage is 18,000 × 0.80 × 0.75 = 10,800, a conversion factor of 0.60. Their cost is 30 × 10,800 = EUR 324,000.

4. Net benefit. The benefit is 75 × 20,000 = EUR 1,500,000. At market wages the cost is 1,140,000 + 460,000 = 1,600,000, a net loss of EUR 100,000. At shadow prices it is 600,000 + 324,000 + 460,000 = 1,384,000, a net gain of EUR 116,000.

5. Cost per QALY. At market wages it is 1,600,000 / 75, about EUR 21,333, above the EUR 20,000 value. At shadow prices it is 1,384,000 / 75, about EUR 18,453, below it.

6. Sensitivity. With unemployment of 0.10 the shadow wage is 18,000 × 0.80 × 0.90 = 12,960, and the net gain falls to EUR 51,200. Without the unemployment term it is 14,400 and the net gain is EUR 8,000. Under full displacement, as in the Green Book, the net loss of EUR 100,000 returns. Break-even comes at 440,000 / 30, about EUR 14,667 per assistant, a factor of about 0.81.

7. Interpretation. With the health gain and its value held fixed, the decision turns on where the assistants come from.

Choices, errors and limits

Worker origin is the largest choice, and it needs evidence on recruitment. Nurses or doctors in short supply fall in the Asian Development Bank's scarce skilled category, with a factor of about 1.0 even where general unemployment is high, so the unskilled shortcut should not be applied to a whole hospital wage bill. A factor estimated in a recession may overstate the saving in a long programme if unemployment falls.

Two of the common mistakes the European Commission guide lists bear on labour. One is counting job creation as a benefit while also applying the shadow wage, since wages are part of the cost, not the benefits, and the shadow wage already captures the social benefit of employment. In the example above, adding the 30 jobs as a benefit would count the same gain twice. The other is borrowing conversion factors from other countries without justification. Treating volunteer time as free understates what a community programme needs in order to scale up, as Kasteng and colleagues argue.

A reservation wage or informal output value is hard to observe, and regional averages may not describe the workers a health programme hires. A low shadow wage makes labour-intensive options look cheaper, which can favour one staffing model over another.

Sources

• Asian Development Bank. Guidelines for the Economic Analysis of Projects. Mandaluyong City, Philippines: Asian Development Bank; 2017. Paragraphs 87 to 94 and Appendix 8. https://doi.org/10.22617/TIM178607-2
• Brouwer WBF, Culyer AJ, van Exel NJA, Rutten FFH. Welfarism vs. extra-welfarism. Journal of Health Economics. 2008;27(2):325-338. https://doi.org/10.1016/j.jhealeco.2007.07.003
• European Commission, Directorate-General for Regional and Urban Policy. Guide to Cost-Benefit Analysis of Investment Projects: Economic appraisal tool for Cohesion Policy 2014-2020. Brussels: European Commission; December 2014. Sections 2.8.5, 2.8.10, 2.8.11, 3.8.1 and 3.8.4 and Annex IV. https://ec.europa.eu/regional_policy/sources/studies/cba_guide.pdf
• HM Treasury. The Green Book (2026). Updated 28 October 2026. Paragraphs 8.1 to 8.5, 8.26 and 8.41 to 8.43. https://www.gov.uk/government/publications/the-green-book-appraisal-and-evaluation-in-central-government/the-green-book-2026
• Kasteng F, Settumba S, Källander K, Vassall A; inSCALE Study Group. Valuing the work of unpaid community health workers and exploring the incentives to volunteering in rural Africa. Health Policy and Planning. 2016;31(2):205-216 (published online 2015). https://doi.org/10.1093/heapol/czv042
• National Institute for Health and Care Excellence. NICE technology appraisal and highly specialised technologies guidance: the manual (PMG36). Published 31 January 2022, last updated 13 October 2026. Sections 4.2.9, 4.4.1 and 4.4.23. https://www.nice.org.uk/process/pmg36
• Pike J, Grosse SD. Friction cost estimates of productivity costs in cost-of-illness studies in comparison with human capital estimates: a review. Applied Health Economics and Health Policy. 2018;16(6):765-778. https://doi.org/10.1007/s40258-018-0416-4
• Turner HC, Toor J, Bettis AA, Hopkins AD, Kyaw SS, Onwujekwe O, Thwaites GE, Lubell Y, Fitzpatrick C. Valuing the unpaid contribution of community health volunteers to mass drug administration programs. Clinical Infectious Diseases. 2019;68(9):1588-1595. https://doi.org/10.1093/cid/ciy741