Vertical Equity

Vertical Equity: Appropriately Unequal Treatment in Health Care Delivery and Financing

Vertical equity is one of the two classic equity principles in health economics, the one that deals with people who differ. Its partner, horizontal equity, asks whether equals are treated alike, while vertical equity concerns differences in the need for care and in the capacity to fund it. It matters because a health system can treat equals alike and still give too little extra care to those in greatest need, or take too large a share of income from poorer households. The principle says that treatment should differ without saying by how much. This page explains how health economists turn that question into a norm that can be tested, how vertical equity in financing is measured with the Kakwani index, and how the idea reaches health technology assessment. An illustrative worked example shows how a progressive and a regressive source of finance combine.

Two settings for one principle: delivery and financing

Wagstaff and van Doorslaer's survey of equity in health care treats financing and delivery separately, and the Equity page covers that split in general terms. In the World Bank guide by O'Donnell, van Doorslaer, Wagstaff and Lindelow, vertical equity is the appropriate unequal treatment of unequals, judged by need for care in delivery and by ability to pay in financing, where a progressive system takes a rising share of ability to pay.

Why measuring horizontal equity rests on a vertical norm

Most empirical work on equity in health care has studied horizontal equity, often stated as equal treatment for equal medical need regardless of income, race or place of residence. The World Bank guide notes that this cannot be tested without a vertical equity norm at the same time. The standard method predicts each person's use of care from need indicators such as age, sex and health status, using the average need and use relationship in the whole population. Income-related differences that remain after this standardisation are then read as horizontal inequity.

That method assumes the average need and use relationship is itself fair. If people in poor health receive too little extra care on average, the benchmark is too low for everyone, and a study can find no horizontal inequity in a system that falls short on vertical equity. Sutton made this point in a study of general practitioner contacts in Scotland, arguing that economic studies usually derive need from the current link between health and health care and so restrict their attention to horizontal inequity. His analysis produced pro-rich estimates of both horizontal and vertical inequity, neither of them statistically significant, and further analysis suggested that vertical inequity in health care may be larger than horizontal inequity.

Setting the vertical norm for health care delivery

Because the principle does not fix how much more care greater need should attract, an analyst has to choose a norm and defend it. Three approaches from the literature are outlined here.

1. Restrict how use responds to health. Sutton proposed replacing the observed health and health care relationship with a norm defined by a restriction on the health elasticity of care, the percentage change in contacts for a percentage change in health status. Observed use can then be compared with use under that norm.

2. Take the norm from a reference group. Vallejo-Torres and Morris, studying people with cardiovascular disease in England, estimated the appropriate relationship between need and use from target groups rather than from the whole population. Adding the vertical component changed their conclusions for several services, with some estimates of income-related inequity becoming more pro-rich.

3. Make the claims explicit. Mooney, Jan and Wiseman argued that vertical equity does not necessarily imply equalising health or any other quantity across groups, and that how groups are treated differently on a fair basis depends on the values adopted. In Australian Aboriginal health they suggested putting this into practice through a positive resource weighting for Aboriginality in resource allocation formulae.

Behind every norm lies a definition of need. Culyer and Wagstaff argued that the definitions of need in the literature were inadequate and proposed a new one, and they held that equality of health should be the dominant principle, so that care is distributed to come as close as feasible to an equal distribution of health. A vertical equity finding is therefore always relative to a stated definition of need and a stated norm.

Measuring vertical equity in financing with the Kakwani index

Progressivity, the usual empirical handle on vertical equity in financing, is measured as a departure from proportionality between payments and ability to pay. The comparison sets the concentration curve of payments, which plots the cumulative share of payments against the population ranked by ability to pay, beside the Lorenz curve of ability to pay. If payments take the same share of ability to pay at every level, the two curves coincide. The summary measure, introduced by Kakwani for tax progressivity, is the Kakwani index:

$$ \pi_K = C_P - G_X $$

where $\pi_K$ is the Kakwani index of progressivity, $C_P$ is the concentration index of health care payments with households ranked by ability to pay, and $G_X$ is the Gini coefficient of ability to pay. The index can take values from minus 2 to 1. A positive value indicates a progressive source, a negative value a regressive one, and zero is consistent with proportionality, although curves that cross can also produce zero, so the index is best read alongside the curves themselves.

Both indices equal twice the area between a curve and the line of equality, which is one minus twice the area under the curve. With grouped data, the curve joins the group points with straight lines, so the area under it is computed exactly as a sum of trapezoids:

$$ C = 1 - \sum_{t=1}^{T} (p_t - p_{t-1})(L_t + L_{t-1}) $$

where $C$ is the concentration index (or the Gini coefficient when the curve is a Lorenz curve), $T$ is the number of groups, $p_t$ is the cumulative population share up to group $t$ with the population ranked from poorest to richest, $L_t$ is the cumulative share of the variable up to group $t$, and $p_0 = L_0 = 0$.

The World Bank guide measures the progressivity of a whole financing system as a weighted average of the Kakwani indices of its sources, with weights equal to each source's share of total payments. In a ten-country comparison published in 1992, Wagstaff, van Doorslaer and colleagues concluded that tax-financed systems tended to be proportional or mildly progressive, social insurance systems regressive, and private systems more regressive still, with out-of-pocket payments an especially regressive way of raising revenue in most of the countries studied.

Worked example: a progressive and a regressive source of finance

The figures in this example are illustrative. A health system is financed from general taxation and from out-of-pocket payments, and households are grouped into income quintiles of 20 per cent of the population each.

1. Compute the Gini coefficient of income. The cumulative income shares are 0.06, 0.17, 0.33, 0.56 and 1.00. Adding each pair of neighbouring cumulative shares gives 0.06, 0.23, 0.50, 0.89 and 1.56, which sum to 3.24. Each quintile has width 0.2, so:

$$ G_X = 1 - 0.2 \times 3.24 = 1 - 0.648 = 0.352 $$

where $G_X$ is the Gini coefficient of pre-payment income.

2. Compute the concentration index of each payment. For tax, the cumulative shares are 0.03, 0.11, 0.25, 0.48 and 1.00, the pair sums are 0.03, 0.14, 0.36, 0.73 and 1.48, and their total is 2.74. For out-of-pocket payments, the cumulative shares are 0.12, 0.27, 0.46, 0.69 and 1.00, the pair sums are 0.12, 0.39, 0.73, 1.15 and 1.69, and their total is 4.08.

$$ C_{tax} = 1 - 0.2 \times 2.74 = 0.452 $$

where $C_{tax}$ is the concentration index of tax payments, with households ranked by income.

$$ C_{oop} = 1 - 0.2 \times 4.08 = 0.184 $$

where $C_{oop}$ is the concentration index of out-of-pocket payments, with the same ranking.

3. Compute the Kakwani indices. Each index subtracts the income Gini coefficient from the payment's concentration index.

$$ \pi_{tax} = 0.452 - 0.352 = 0.100 $$

where $\pi_{tax}$ is the Kakwani index of tax finance.

$$ \pi_{oop} = 0.184 - 0.352 = -0.168 $$

where $\pi_{oop}$ is the Kakwani index of out-of-pocket finance. Tax finance is progressive and out-of-pocket finance is regressive. The ratio of payment share to income share confirms the direction: for tax it rises from 0.5 in the poorest quintile (3 divided by 6) to about 1.18 in the richest (52 divided by 44), while for out-of-pocket payments it falls from 2.0 (12 divided by 6) to about 0.70 (31 divided by 44).

4. Combine the sources. Suppose 60 per cent of revenue comes from tax and 40 per cent from out-of-pocket payments. The overall index is the revenue-weighted average:

$$ \pi_{total} = 0.6 \times 0.100 + 0.4 \times (-0.168) = 0.060 - 0.0672 = -0.0072 $$

where $\pi_{total}$ is the Kakwani index for the whole financing system. The system as a whole is close to proportional and very slightly regressive, even though its larger source is progressive. The combined payment shares are 6.6, 10.8, 16.0, 23.0 and 43.6 per cent, so the cumulative payment curve (0.066, 0.174, 0.334 and 0.564 at the first four quintiles) lies just above the Lorenz curve at every quintile, which confirms the slight regressivity.

5. Translate progressivity into redistribution. The World Bank guide links progressivity to the redistributive effect of compulsory payments through the vertical redistribution term:

$$ V = \frac{g}{1 - g} K_E $$

where $V$ is the vertical redistributive effect (the fall in the income Gini coefficient that progressivity produces), $g$ is the average payment rate as a proportion of income, and $K_E$ is the Kakwani index that would apply if households with equal pre-payment income paid equally. If tax payments for health care take 5 per cent of income, with no horizontal inequity and no reranking, then $V = (0.05 / 0.95) \times 0.100 = 0.0053$, so the tax lowers the income Gini coefficient from 0.352 to about 0.347.

Vertical equity in health technology assessment

Economic evaluation usually builds in no vertical distinction: the reference case in NICE's manual for technology appraisal (PMG36) states that an additional QALY has the same weight regardless of the other characteristics of the people receiving the health benefit, except in specific circumstances. Mooney and Jan noted that the health policy literature on equity had mostly focused on horizontal equity and had tended to overlook vertical equity, and they examined whether it should enter decisions through weighting outcomes or through fair procedures. Equity-weighted analysis takes the first route by giving health gains for some groups more weight.

The severity modifier in PMG36 is one of those specific circumstances. The committee may apply a greater weight to QALYs when a technology is for a condition with a high degree of severity, judged by the absolute shortfall and proportional shortfall in future QALYs with current care, whichever implies the greater severity. QALYs are weighted by 1.2 when the proportional shortfall is 0.85 to 0.95 or the absolute shortfall is 12 to 18, and by 1.7 when the proportional shortfall is at least 0.95 or the absolute shortfall is at least 18. The manual states that the modifier will not initially be applied to technology appraisals of HealthTech. It does not describe the modifier as vertical equity, but giving more weight to health gains for those facing larger losses of future health has the same structure: unequal treatment justified by an unequal starting position.

Vertical equity, horizontal equity and vertical inequity

The nearby terms differ in what they compare and in whether they name a goal or a failure. Keeping them apart avoids reading a measure of one as evidence about another.

Vertical inequity names the observed failure to meet the vertical equity norm. Vertical equity also differs from equality of health. A distribution of care can meet a vertical norm based on need while health still differs between groups, and Mooney, Jan and Wiseman argued that vertical equity does not necessarily imply equalising health.

Where vertical equity findings mislead

A frequent misunderstanding is to treat vertical equity as a measurable fact. Any finding depends on two choices: how need or ability to pay is measured, and what pattern counts as appropriate.

The progressivity of the largest source does not settle the direction for the system, as the worked example shows. Progressivity results also rest on incidence assumptions about who really bears each payment, such as the common assumption that employer insurance contributions fall on employees through lower wages.

More progressive financing is not automatically fairer. The verdict depends on how much more those with greater ability to pay ought to contribute, which is a value judgement, and on whether high out-of-pocket payments are deterring people with low incomes from using care. A regressive payment that people avoid by going without treatment can look small in financing data while causing a failure in delivery.

Sources

• Culyer AJ, Wagstaff A. Equity and equality in health and health care. Journal of Health Economics. 1993;12(4):431-457.
• Kakwani NC. Measurement of tax progressivity: an international comparison. Economic Journal. 1977;87(345):71-80.
• Mooney G, Jan S. Vertical equity: weighting outcomes? or establishing procedures? Health Policy. 1997;39(1):79-87.
• Mooney G, Jan S, Wiseman V. Staking a claim for claims: a case study of resource allocation in Australian Aboriginal health care. Social Science and Medicine. 2002;54:1657-1667.
• National Institute for Health and Care Excellence. NICE technology appraisal and highly specialised technologies guidance: the manual (PMG36). London: NICE; 2022, updated March 2026. Table 4.1 and section 6.2.
• O'Donnell O, van Doorslaer E, Wagstaff A, Lindelow M. Analyzing Health Equity Using Household Survey Data: A Guide to Techniques and Their Implementation. Washington, DC: World Bank; 2008. Chapters 8, 15, 16 and 17.
• Sutton M. Vertical and horizontal aspects of socio-economic inequity in general practitioner contacts in Scotland. Health Economics. 2002;11(6):537-549.
• Vallejo-Torres L, Morris S. Income-related inequity in healthcare utilisation among individuals with cardiovascular disease in England: accounting for vertical inequity. Health Economics. 2013;22(5):533-553.
• Wagstaff A, van Doorslaer E. Equity in health care finance and delivery. In: Culyer AJ, Newhouse JP, editors. Handbook of Health Economics. Volume 1B. Amsterdam: Elsevier; 2000. p. 1803-1862.
• Wagstaff A, van Doorslaer E, Calonge S, Christiansen T, Gerfin M, Gottschalk P, et al. Equity in the finance of health care: some international comparisons. Journal of Health Economics. 1992;11(4):361-387.