Equity in Financing and distribution of
health benefits in Zambia
Collins Chansa
Netsanet Walelign Workie
Bona Chitah
Oliver Kaonga
August 2018
© 2018 International Bank for Reconstruction and Development / The World Bank
1818 H Street NW, Washington, DC 20433
Telephone: 202-473-1000; Internet: www.worldbank.org
Some rights reserved
1 2 3 4 19 18 17 16
This work is a product of the staff of The World Bank with external contributions. The findings,
interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World
Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not
guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and
other information shown on any map in this work do not imply any judgment on the part of The World
Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries.
Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and
immunities of The World Bank, all of which are specifically reserved.
Rights and Permissions
This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO)
http://creativecommons.org/licenses/by/3.0/igo. Under the Creative Commons Attribution license, you
are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the
following conditions:
Attribution—Please cite the work as follows: World Bank. 2019. Zambia Health Sector Public
Expenditure Review: 2006-2016. Washington, DC: World Bank.
Translations—If you create a translation of this work, please add the following disclaimer along with
the attribution: This translation was not created by The World Bank and should not be considered an
official World Bank translation. The World Bank shall not be liable for any content or error in this
translation.
Adaptations—If you create an adaptation of this work, please add the following disclaimer along with
the attribution: This is an adaptation of an original work by The World Bank. Views and opinions
expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and
are not endorsed by The World Bank.
Third-party content—The World Bank does not necessarily own each component of the content
contained within the work. The World Bank therefore does not warrant that the use of any third-party-
owned individual component or part contained in the work will not infringe on the rights of those third
parties. The risk of claims resulting from such infringement rests solely with you. If you wish to re-use
a component of the work, it is your responsibility to determine whether permission is needed for that
re-use and to obtain permission from the copyright owner. Examples of components can include, but
are not limited to, tables, figures, or images.
All queries on rights and licenses should be addressed to the Publishing and Knowledge Division, The
World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail:
pubrights@worldbank.org.
1
Table of Contents
LIST OF TABLES ................................................................................................................................................ 2
LIST OF FIGURES .............................................................................................................................................. 3
ACKNOWLEDGMENT ...................................................................................................................................... 4
ABBREVIATIONS ............................................................................................................................................... 5
1. INTRODUCTION ....................................................................................................................................... 6
1.2 OBJECTIVES OF THE STUDY.................................................................................................................... 7
2. KEY ISSUES IN HEALTH FINANCING IN ZAMBIA .......................................................................... 7
3. CONCEPTUAL FRAMEWORK ............................................................................................................... 8
3.1 FINANCING INCIDENCE ANALYSIS (FIA) ....................................................................................................... 8
3.2 CATASTROPHIC HEALTH EXPENDITURE (CHE) ............................................................................................. 8
3.2.1 Incidence and Intensity of Catastrophic Expenditures .......................................................................... 9
3.3 BENEFIT INCIDENCE ANALYSIS (BIA) ........................................................................................................... 9
4. METHODS ................................................................................................................................................... 9
4.1 DATA SOURCES ............................................................................................................................................. 9
4.2 ASSESSING HEALTH CARE FINANCING INCIDENCE ....................................................................................... 10
4.2.1 Tax Computations ............................................................................................................................... 10
4.2.2 Ability to pay ....................................................................................................................................... 10
4.2.3 Health care payments ......................................................................................................................... 10
4.3 ESTIMATING CATASTROPHIC HEALTH EXPENDITURE ................................................................................... 11
4.4 ESTIMATING BENEFICIARY INCIDENCE ........................................................................................................ 11
5. RESULTS ................................................................................................................................................... 12
5.1 FINANCING INCIDENCE ANALYSIS ............................................................................................................... 12
5.1.1 Structural Progressivity of Health Financing: Taxes and OOP ......................................................... 12
5.1.2 Effective Progressivity of Health Care Payments: Taxes and OOP ................................................... 13
5.2 CATASTROPHIC HEALTH EXPENDITURE ...................................................................................................... 16
5.2.1 Incidence and Intensity of Catastrophic Health Expenditures ................................................... 16
5.3 BENEFIT INCIDENCE ANALYSIS (BIA) ......................................................................................................... 18
5.3.1 Distribution of health subsidies and outpatient visits at public health facilities by province ............. 18
5.3.2 Distribution of total subsidies in comparison to reported illnesses at provincial level ...................... 19
5.3.3 Distribution of outpatient and inpatient health care benefits by income groups ................................ 19
5.3.4 Distribution of total benefits in comparison to need by income groups .............................................. 21
6. CONCLUSION .......................................................................................................................................... 22
REFERENCES ................................................................................................................................................... 24
APPENDIX 1: KEY HEALTH REFORM AREAS AND ELEMENTS, ZAMBIA: 1992-2018 .................. 26
APPENDIX 2: INCIDENCE AND INTENSITY OF CATASTROPHIC HEALTH EXPENDITURE ...... 27
APPENDIX 3: TECHNICAL NOTES .............................................................................................................. 28
1.0 FINANCING INCIDENCE ANALYSIS ............................................................................................................... 28
Progressivity of funding sources .................................................................................................................. 28
2.0 BENEFIT INCIDENCE ANALYSIS - CONSTANT UNIT COST ASSUMPTION ....................................................... 29
3.0 HEALTH EXPENDITURES AND IMPOVERISHMENT EFFECTS .......................................................................... 30
1
List of Tables
TABLE 1: TAX SOURCES, DEFINITIONS AND COMPUTATION TECHNIQUES ............................................................... 11
TABLE 2: PERCENTAGE CONTRIBUTIONS TO TOTAL TAX REVENUE BY MAJOR TAX SOURCES (2010–2014)............ 12
TABLE 3: DISTRIBUTION OF HEALTH FINANCING BURDEN IN ZAMBIA: TAXES AND OOP - 2010............................ 15
TABLE 4: BENEFIT INCIDENCE TEST RESULTS – OUTPATIENT HEALTH SERVICES.................................................... 20
TABLE 5: BENEFIT INCIDENCE TEST RESULTS – OUTPATIENT AND INPATIENT SERVICES - 2014 ............................. 20
TABLE A.1: OOP EXPENDITURE ON HEALTH – 2010 .............................................................................................. 27
TABLE A.2: OOP EXPENDITURE ON HEALTH – 2015 .............................................................................................. 27
2
List of Figures
FIGURE 1: STRUCTURAL PROGRESSIVITY OF HEALTH CARE PAYMENTS - 2010 ...................................................... 13
FIGURE 2: STRUCTURAL PROGRESSIVITY OF HEALTH CARE PAYMENTS - 2014 ...................................................... 13
FIGURE 3: EFFECTIVE PROGRESSIVITY OF INDIVIDUAL TAXES AND OOP, ZAMBIA 2010 VS 2014.......................... 14
FIGURE 4: PREVALENCE OF SMOKING CIGARETTES AMONG MALE ADULTS IN ZAMBIA .......................................... 15
FIGURE 5: INCIDENCE OF CATASTROPHIC HEALTH SPENDING AT HOUSEHOLD LEVEL, 2010-2015 ......................... 16
FIGURE 6: INTENSITY OF CATASTROPHIC HEALTH SPENDING AT HOUSEHOLD LEVEL, 2010-2015 .......................... 17
FIGURE 7: CATASTROPHIC HEALTH SPENDING BY RESIDENCE AND FACILITY TYPE (PERCENT), 2014 ................... 17
FIGURE 8: SHARES OF TOTAL HOUSEHOLD HEALTH EXPENDITURE (PERCENT), 2014 ............................................ 17
FIGURE 9: DISTRIBUTION OF HEALTH SUBSIDIES AND OUTPATIENT VISITS AT PUBLIC HEALTH FACILITIES BY
PROVINCE ..................................................................................................................................................... 18
FIGURE 10: DISTRIBUTION OF TOTAL SUBSIDIES IN COMPARISON TO REPORTED ILLNESSES BY PROVINCE ............. 19
FIGURE 11: DISTRIBUTION OF TOTAL BENEFITS IN COMPARISON TO NEED FOR HEALTH CARE: 2010 VS 2015 ........ 21
3
Acknowledgment
This report was prepared by the Health, Nutrition, and Population (HNP) Global Practice of
the World Bank Group in collaboration with the Department of Economics at the University
of Zambia (UNZA). Financial support was provided by the U.K. Department for International
Development. From the HNP, the main authors are Collins Chansa (Senior Health Economist
and Task Team Leader) and Netsanet Walelign Workie (Senior Health Economist); and from
UNZA, Bona Chitah and Oliver Kaonga. Technical guidance was provided by Pia Schneider
(Lead Economist), Patrick Hoang-Vu Eozenou (Senior Health Economist), Magnus Lindelow
(Practice Manager), and Prof. Felix Masiye (UNZA).
The draft report and initial findings were discussed and validated through the national Health
Care Financing Technical Working Group (HCFTWG) meetings. The authors would like to
thank all the members of the HCFTWG for their invaluable contributions. Sincere appreciation
goes to Henry Kansembe, Mpuma Kamanga, Patrick Banda, Anita Kaluba, and Wesley
Mwambazi (Ministry of Health, Zambia) for supporting the team in interpreting the findings.
The final draft report was peer reviewed by: Pia Schneider, Laura Di Giorgio, Ellen Van De
Poel, Reem Hafez, and Patrick Hoang-Vu Eozenou (World Bank, HNP); and Prof. Felix
Masiye (UNZA). Administrative and editorial support were provided by Gertrude Mulenga
Banda, Yvette Atkins, Charity Inonge Mbangweta, and Claire Jones..
4
Abbreviations
AE Adult Equivalent
ATP Ability to Pay
BIA Benefit Incidence Analysis
CBoH Central Board of Health
CC Concentration curves
CHE Catastrophic Health Expenditure
CI Concentration Index
CIT Corporate Income Tax
FIA Financing Incidence Analysis
KI Kakwani Index
LC Lorenz Curve
LCMS Living Conditions Monitoring Survey
MPO Mean Positive Overshoot
MSL Medical Stores Limited
NHA National Health Accounts
OOP Out-of-Pocket
PAYE Pay as You Earn
PBC Performance-Based Contracting
PIT Personal Income Tax
RBF Results-Based Financing
SWAp Sector-Wide Approach
VAT Value Added Tax
ZHHEUS Zambia Household Health Expenditure and Utilization Survey
ZRA Zambia Revenue Authority
5
1. Introduction
The Zambian health care system continues to undergo various reforms. The system has
experienced health financing and organisational reforms since 2006. Among the notable and
common themes underpinning all the health reforms across the different timelines have been
the following:
a. Equity. The policy commitment continues to be associated with the desire to ensure that
resources and health care services are distributed and provided as close to the household
or family as possible.
b. Access and affordability. Access to quality and affordable health care services
continues to be one of the central themes of the reform process. In this regard, health
services are supposed to be generally available, adequate, and of reasonable quality and
cost.
Given the above, government expenditures in Zambia have been focused on provision of
primary health care services. It is assumed that this strategy can help to guarantee equitable
availability of health care services to the population irrespective of socio-economic status.
Some of the other policies and instruments which have been used to enhance equity in Zambia
over the years are:
(a) Resource allocation formula. Allocation of resources in the health sector has evolved from
a historical budgeting approach in the 1990s to a needs-based approach. In 2004, a needs-
based resource allocation formula was developed (and later revised in 2009 and 2017) to
facilitate an evidence-based distribution of resources from the Ministry of Health to the
districts. This formula accounts for material deprivation using variables such as prevalence
of poverty; ownership of assets; disease burden; and access to banks, markets, and fuel
stations. The deprivation index is then used as a weight on district population to derive
relative shares of the resource envelope.
(b) Linked to the resource allocation formula is the basic or essential health care package
which has been defined at all levels of the health system in Zambia. The main aim of this
package is to rationalize planning subject to resource constraints and epidemiological
considerations. Thus, priority setting and resource allocation have to be in line with
requisites of the basic health care package.
(c) To increase access and utilization of health services, user fees were abolished at primary
healthcare level (health posts, health centers and district hospitals) in 2006 in rural areas,
2007 in peri-urban areas, and 2012 at the entire at primary healthcare level nationwide.
(d) Human resource distribution and placement. Distribution of human resources is
predominantly urban biased. The policy goal has been to achieve parity by re-distributing
the available health workers and posting new graduates to rural areas to achieve a balance
among the different geographical areas.
(e) Infrastructure development. Over the years, there have been huge investments in
infrastructure (health posts, health centers and hospitals), medical equipment, staff
housing in rural areas, and vehicles (including ambulances and motorbikes).
6
Determining the extent to which the above goals are progressing has raised the need for
evidence. It is anticipated that if the policy measures are being attained, then equity in financing
and utilisation of quality healthcare services ought to be attained. This is important because the
overarching objective of the Zambia health care system, as with many other health systems
worldwide, is to ensure that its population is provided with a minimum level and quality of
health care. High expenditure on healthcare has the potential to expose poor households and
individuals into further deprivation or poverty. The impoverishing or catastrophic effects of
health care expenditures through out-of-pocket (OOP) payments are major causes of concern
for policy makers worldwide. In Zambia, the health policy is focused on achieving fairness in
financing and consumption of health care, and reducing inequities in health outcomes.
Given that a number of health financing reforms have been implemented in Zambia over the
years, this study provides an updated analysis on the extent to which government expenditures
on health provide an effective intervention in redistributing health care resources in an
equitable manner. The study looks at the health system holistically and does not look at each
of the individual reforms.
1.2 Objectives of the Study
The objective of the study is to analyse equity in financing, and beneficiary incidence. The
study uses the 2014 Household Health Expenditure and Utilization Survey; and the 2010 and
2015 Living Conditions Monitoring Survey. The specific objectives are:
a. Financing incidence analysis – To estimate the distribution of health care financing
burden between socio-economic groups, distinguishing between public and private
financing mechanisms, and the factors influencing this distribution;
b. Catastrophic health expenditure analysis – to assess the extent of catastrophic payments
for health care at household level; and
c. Benefit incidence analysis – To explore whether poor households benefit from public
spending by facility level (primary, secondary and tertiary), by ownership (public and
private), and by service (inpatient and outpatient).
2. Key Issues in Health Financing in Zambia
Zambia has over the past two decades embodied equity as a key element in the financing and
distribution of health benefits. The policy debate on health equity in Zambia is linked to the
political-economic history of the country and is one of the key elements of the 1991 health
reforms (Kalumba 1997), and subsequent reforms. Motivation to incorporate equity in health
financing and provision was triggered by the deterioration of the economy in the 1980s, which
contributed to inequities in health outcomes and access to health care (Ministry of Health
1991). Consequently, the Zambian government has been unwavering in its pursuit of inclusive
growth centered on eliminating the risk and prevalence of poverty and deprivation through the
social sectors such as health and education. Thus, over the years, the Zambian government
through the Ministry of Health has implemented a number of financing and organisational
reforms aimed at achieving equity and other key health system goals (Gilson et al. 2003; Lake
and Musumali 1999). Thus, the 1991 health reform vision of ‘equity of access to cost-effective
quality health care as close to the family as possible’ has remained unchanged since 1991
7
(Ministry of Health 1991). A summary of the main elements of the health reforms in Zambia
is provided in Appendix 1.
Some of the key health financing reforms which have been implemented over the years include
the abolition of user fees in rural areas, peri-urban areas, and all primary health care facilities
in 2006, 2007, and 2012, respectively (Carasso et al. 2012; Masiye, Kaonga, and Kirigia 2016).
To enhance value-for-money and results-focus, Zambia also implemented a nation-wide
performance-based contracting (PBC) system through a sector-wide approach programming
(SWAp) framework between 1996 and 2006 (Chansa et al. forthcoming). PBC was abandoned
in 2006 but later reappeared in form of results-based financing (RBF) in 11 districts between
2011 and 2014, and in 58 districts between 2016 and 2018. Further, a needs-based resource
allocation formula for allocating operational grants flowing through the public health system
at district level has been in implementation since 2004 (Chitah and Masiye 2007).
But though Zambia has implemented several health reforms, and has a fiscal redistributive
system comprising social expenditures and taxes, the impact of these reforms and policies on
poverty reduction and shared prosperity have not been adequately evaluated, especially in the
health sector. Assessing fairness in financing of healthcare, resource allocation, and impact of
public policies on the poor is critical to monitoring and evaluating the attainment of health
systems goals of: (i) improved health status, (ii) financial risk protection, (iii) responsiveness
to needs, and (iv) client satisfaction. This study applies three methods: (i) financing incidence
analysis, (ii) benefit incidence analysis, and (iii) catastrophic health expenditure analysis to
determine whether equity in health care financing and utilisation have been enhanced or not.
3. Conceptual Framework
3.1 Financing Incidence Analysis (FIA)
Equity in the sources of financing for health assumes an important dimension as it represents
how financing is distributed. To determine how financing is distributed, analysis showing who
or where the burden or tax incidence lies among the various socio-economic groups in society
is conducted. This is done by looking at the progressivity of the sources of revenue for health
care financing. A measure of progressivity advocated for is based on an approach which
measures the extent of the departure of proportionality in the relationship between payments
(resources generated) and the ability to pay (O’Donnell et al. 2008). The main sources of health
financing in Zambia are: public financing through taxation, and donors; and to a small extent
households and private health insurance.
3.2 Catastrophic Health Expenditure (CHE)
There are two alternative approaches that are used to consider the distribution of health care
spending by households. The underlying principle for each of these is that health care payments
should be at a certain level which should not exceed a given threshold such that the household
suffers undue financial ruin or suffering as a consequence of experiencing expenditures above
the given threshold (Wagstaff and Doorslaer 2003). Households that may experience
expenditures on health care such that such expenses are above the threshold are said to
experience catastrophic expenses. Alternatively, the minimum expenses on health care by
households are set such that household income does not go below a set limit and experience
poverty as a consequence of health payments.
8
3.2.1 Incidence and Intensity of Catastrophic Expenditures
In order to extend the interpretation of catastrophic expenditures reference is made to the
intensity and incidence. The incidence or headcount is the percentage of individuals whose
health care costs expressed as a proportion of income, exceeds a given discretionary threshold.
The intensity or mean gap is the average amount by which payments as a proportion of income
exceeds the threshold (Bredenkamp, Mendola, and Gragnolati 2010; O’Donnell et al. 2008).
3.3 Benefit Incidence Analysis (BIA)
Conceptually, BIA considers the distribution of public expenditures for services among
different groups in the population particularly among different income groups. The main
objective of BIA is to assess whether public spending is progressive, that is, whether public
services serve the intended beneficiaries and whether it can redistribute resources to the poor.
This is important because the effectiveness of public expenditure is usually determined by its
ability to have a positive impact on the poor. By undertaking a BIA, we develop a relationship
between public expenditure and the health outcomes that are produced (Demery 2000).
Ultimately, BIA addresses the question of how public expenditures and benefits are distributed
among the different socio-economic groups, and if any of the groups are advantaged or
disadvantaged by the other (McIntyre and Ataguba 2011; Castro-Leal et al. 2000).
4. Methods
4.1 Data Sources
This study applies the standard methodologies for FIA, BIA, and CHE analysis. Raw data was
obtained from the Living Conditions Monitoring Survey (LCMS) for 2010, and 2015; the 2014
Zambia Household Health Expenditure and Utilisation Survey (ZHHEUS); the Zambia
Revenue Authority (ZRA), and National Health Accounts. The LCMS is a repeated nationally
representative cross‐sectional household survey which use a two-stage stratified cluster
sampling method to generate household and individual-level information. The LCMS is
designed to provide data on living conditions and welfare (including poverty estimates)
overtime; and each survey includes modules on health, education, agriculture, household
consumption and expenditure, economic and labour market activity and so forth. The 2004 and
2010 LCMS were administered to approximately 20,000 households while the 2015 LCMS
was administered to 12,260 households. Considering that the health modules in the LCMS’ are
too general and do not contain adequate data on health choices and spending, Zambia
conducted a nationally representative health-sector specific household survey in 2014 which
generated comprehensive data on health expenditure and utilisation. This study (the ZHHEUS)
used a two-stage stratified sampling approach (similar to the LCMS), and gathered household-
and individual-level information from 11,927 households. Tax revenue data was obtained from
the ZRA (specifically to serve the FIA), and health expenditure data from previous National
Health Accounts (NHA) surveys and the ZHHEUS.
9
4.2 Assessing health care financing incidence
This study applies both structural and effective approaches to assessing health care financing
incidence.
4.2.1 Tax Computations
The analysis of health care financing incidence was limited to general taxation, and OOP
payments. The key taxes covered in the analysis include personal income tax (PIT), corporate
income tax (CIT), value added tax (VAT), fuel levy, and excise tax. Individual household tax
contribution for various taxes was extracted from tax revenue data that was obtained from ZRA.
This is summarised in Table 1 below which shows an overview of each type of tax, and the
estimation method. The PIT component was estimated based on reported income while excise
tax, VAT and fuel levy were estimated based on reported consumption expenditure on items
where tax was applicable. For CIT, the study assumed that 50 percent of the tax burden falls
on shareholder (those that reported dividends) and 50 percent on consumers. This is because it
is difficult to know with certainty whether the burden of CIT is borne by shareholders or if it
is passed on to consumers. In the estimations we take a graduated assessment of the different
percentiles for assessing the incidence of CIT. However, as has been argued in the literature
there is really no consensus or gold standard that has been established (Jenkins, Kuo, and
Shukla 2000).
4.2.2 Ability to pay
FIA studies require an assessment of household’s capacity or ability to pay. This is because
health financing equity is analysed with respect to ability to pay. But while reported income is
used as a measure of socio‐economic status in higher income settings, reported income
generally suffers from under‐reporting in low income countries, and is unreliable. In low
income countries, consumption expenditure and composite indices of socio‐economic status
have been proposed as more reliable measures of socio‐economic status. However, the use of
a composite index is limited to the analysis of the concentration of either payments or income
distribution and one cannot use it to calculate distribution indices. Therefore, adult equivalent
consumption expenditure was used as the measure of ability to pay in this study. Specifically,
per capita consumption and household expenditure were used to measure ability to pay.
4.2.3 Health care payments
At the household level, all payments relating to consumption of health care by the household
were aggregated to obtain the total sum of health care expenditures which are considered as
OOP. At the macro level, we obtained health care payments from general tax sources as the
Zambian health care system is primarily funded through general taxes.
Technical details on the FIA approach are presented in Appendix 3.
10
Table 1: Tax sources, definitions and computation techniques
Tax Rates Computation Techniques
CIT 35 percent on dividends Apportioning CIT receipts based on assumptions on
tax shifting. Shifting assumptions include certain
percentage borne by shareholders (the LCMS includes
information on those who receive dividends) and the
rest of the households through consumption. We
assumed that 50 percent of the tax burden is borne by
shareholders and the remainder 50 percent by
households through consumption.
VAT 16 percent on standard rated goods This was computed on all commodities where VAT
was applicable. We excluded all commodities that are
either zero rated or exempt.
PIT • Equal to or below K800,000 – no tax We first removed a non-taxable pension allowance of
• 25 percent on income above K550,000 before computing the amounts of taxes.
K800,000 but equal to or below Then applied rates to the appropriate tax bands.
K1,335,000
• 30 percent on income above
K1,335,000 but equal to or below
K3,300,000
• 35 percent on income above
K3,300,000
Excise tax • 60 percent on clear bear All these taxes were charged on respective product
• 10 percent on opaque beer values. No tax was charged on unprocessed tobacco
• 125 percent on spirits and wines because this is assumed to be sold informally.
• 145 percent on cigarettes
Health 1 percent on interest earned from The 1 percent was charged on interest earned from all
Levy savings formal bank savings accounts.
Fuel Levy 15 percent on petrol and diesel Fuel levy was estimated for those who use fuel directly
by applying a 15 percent rate while for those using
commercial buses and other modes of transport it was
estimated based on the total spent on public transport.
4.3 Estimating catastrophic health expenditure
Technical details on the CHE approach are presented in Appendix 3.
4.4 Estimating beneficiary incidence
This study used the standard methodology for BIA which involves three steps (McIntyre and
Ataguba 2011; O’Donnell et al. 2008). These are:
i. Estimation of unit expenditures for health services.
ii. Imputation of the unit expenditure to households or individuals. This is in effect
considered an in-kind transfer and BIA is used in measuring the distribution of the
in-kind transfer across the population for health care
iii. Grouping of households (or individuals) by sub-groups of the population (based on
some measure of socio-economic outcome). A common criterion for grouping is
income or wealth.
Technical details on the BIA approach are presented in Appendix 3.
11
5. Results
5.1 Financing Incidence Analysis
5.1.1 Structural Progressivity of Health Financing: Taxes and OOP
This section looks at the distribution of the health financing burden through taxes and OOP
between 2010 and 2014. First, results on the percentage contribution to total tax revenue by
major tax sources is provided in Table 2. The results show that income taxes generate more
than half of the total government revenue between 2010 and 2014. On average, income from
individuals through Pay as You Earn (PAYE) or personal income contributed 25 percent of the
total annual tax revenue between 2010 and 2014 while revenue from domestic taxes on goods
and services was 14 percent of the total tax revenue over the same period. Income from Excise
taxes was the largest share of domestic taxes on goods and services.
Table 2: Percentage contributions to total tax revenue by major tax sources (2010–2014)
2010 2011 2012 2013 2014 Average
Income Tax 56 52 57 50 48 53
CIT 18 19 21 12 13 17
PAYE 29 24 24 25 23 25
Withholding Tax 5 4 5 6 6 5
Extraction Royalty 3 5 7 8 6 6
Domestic Goods & Services 14 9 9 15 22 14
Excise Duties 10 9 11 10 10 10
Domestic VAT 4 0 -1 5 11 4
Trade Taxes 30 30 34 35 30 32
Import VAT 20 21 24 27 23 23
Import Tariffs 10 9 10 8 7 9
Source: Authors’ compilation from ZRA data.
Figures 1 and 2 presents results from an assessment of the distribution of health care payments
as a share of ability to pay (ATP) across the five quintiles for 2010 and 2014. The results show
that for both 2010 and 2014, the share of ATP that each quintile spent on health care through
corporate income tax (CIT), Value Added Tax (VAT), and Excise taxes are increasing with
each quantile. This suggests that CIT, VAT, and Excise taxes are progressive i.e. wealthier
households pay a higher share of their income on these taxes. On the other hand, results for
both 2010 and 2014 show that PIT is not progressive for the second poorest and middle
quintiles. This implies that households in the second poorest quantile pay a higher share of their
personal income than the middle quantile. For OOP spending, results for 2010 shows a weak
progression or proportional distribution of ATP as compared to 2014 which shows a
progressive distribution.
12
Figure 1: Structural progressivity of health care payments - 2010
25
Health payments as % of ATP
20
15
10
5
0
CIT VAT PIT Excise taxes OOP
Poorest 20% Second Poorest 20% Middle 20%
Second Richest 20% Richest 20%
Source: Authors’ compilation from LCMS 2010 data.
Figure 2: Structural progressivity of health care payments - 2014
70
60
Health payments as % of ATP
50
40
30
20
10
0
CIT VAT PIT Excise taxes OOP
Poorest 20% Second Poorest 20% Middle 20% Second Richest 20%
Source: Authors’ compilation from ZHHEUS 2014 data.
5.1.2 Effective Progressivity of Health Care Payments: Taxes and OOP
Assessing progressivity in health financing by using ratios as highlighted in 5.1.1 has some
limitations (Ataguba et al. 2018). In particular, structural progressivity does not provide a
comprehensive picture of how health care payments as a share of ATP are distributed across
the entire spectra over time. To avert this problem, the analysis was repeated by using
concentration curves with a view of measuring effective progressivity by using the Kakwani
index (Figure 3). In addition, dominance tests were also calculated (Table 3) in order to
establish whether progressivity or regressivity is the same for the entire distribution of ATP.
This approach is consistent with guidelines on conducting FIA studies (Ataguba et al. 2018),
and the multiple comparison approach was used to conduct the dominance tests (O’Donnell et
al. 2008).
13
Results from the effective progressivity analysis for the year 2010 (Figure 3A and Table 3)
show that taxes as a whole are progressive (Kakwani index = 0.18, p˂0.01). However, a review
of individual taxes shows that Excise tax is regressive (Kakwani index = -0.015, p˂0.01).
Specifically, the concentration curve for Excise tax lies inside the Lorenz curve after the 60th
cumulative population share. Below this threshold, the concentration curve oscillates up and
down the Lorenz curve and this explains why the dominance test for Excise tax is not dominant.
Non-dominance for Excise tax could be explained by cigarette tax that has a negative Kakwani
index (-0.16, p˂0.01) (which suggests regressivity). However, the 45 degree line and the
Lorenz curve for cigarettes are non-dominant. On the other hand, taxes on alcohol are
progressive (Kakwani index = 0.21, p˂0.01). The reason why cigarette tax in Zambia is
regressive is because the percentage of adult men smoking cigarettes is prevalent among the
poor (Figure 4). Review of the progressivity of CIT, VAT, and PIT for the year 2010 shows
varying levels of progressivity with PIT being the least progressive (Kakwani index = 0.01,
p<0.01) while CIT (Kakwani index = 0.20, p<0.01) and VAT (Kakwani index = 0.24, p<0.01)
are more progressive. For 2014, results show that VAT and Excise taxes are progressive while
PIT is regressive (Figure 3B).
For OOP spending, results for 2010 (Figure 3C and Table 3) shows that OOP is progressive
(Kakwani index = 0.06, p˂0.01). However, the level of progressivity is very weak and this is
confirmed from the dominance test for OOP which shows that the 45 degree line dominants
while the Lorenz curve is non-dominant. In 2014, the concentration curve for OOP dominates
the Lorenz curve (Figure 3D) and this indicates that OOP health spending is regressive in 2014.
Figure 3: Effective progressivity of individual taxes and OOP, Zambia 2010 vs 2014
A. Concentration curve for taxes, 2010 B. Concentration curve for taxes, 2014
Source: Authors’ compilation from LCMS 2010 and ZHHEUS 2014 data.
14
C. Concentration curve for OOP, 2010 D. Concentration curve for OOP, 2014
Source: Authors’ compilation from LCMS 2010 and ZHHEUS 2014 data.
Table 3: Distribution of health financing burden in Zambia: Taxes and OOP - 2010
Concentration Index Kakwani Index (standard Dominance Test
(standard error) error) 45 degree line Lorenz
CIT 0.78*** (0.06) 0.20*** (0.32) - -
Domestic VAT 0.81*** (0.02) 0.24*** (0.02) - -
PIT 0.58*** (0.16) 0.01*** (0.07) - -
Excise taxes 0.55*** (0.044) -0.015*** (0.13) - -+
Alcohol 0.78*** (0.045) 0.21*** (0.12) - -
Cigarettes 0.4*** (0.06) -0.16*** (0.19) -+ -+
Fuel 0.82*** (0.02) 0.25*** (0.12) - -
Health levy 0.75*** (0.05) 0.17*** (0.1) - -
All taxes 0.76*** (0.06) 0.18*** (0.05) - -
OOP 0.64*** (0.02) 0.06*** (0.04) - +
***p˂0.01, **p˂0.05, *p˂0.1
Dominance tests: – indicates that the 45 degree line or Lorenz curve dominates the concentration curve
+ indicates that the concentration curve dominates the 45 degree line or Lorenz curve
-+ indicates non-dominance i.e. the concentration curve crosses the Lorenz curve at a certain point
Figure 4: Prevalence of smoking cigarettes among male adults in Zambia
35
30 33.8
25
20
Percent
22.1
15 17.2 17.4
10
10.7
5
0
Poorest 2nd 3rd 4th Wealthiest
Source: Central Statistical Office, Ministry of Health, ICF International (2014).
15
5.2 Catastrophic Health Expenditure
5.2.1 Incidence and Intensity of Catastrophic Health Expenditures
This section examines changes in the incidence and intensity of OOP health spending across
households from different income status between 2010 and 2015. Having established that OOP
health spending is regressive from the previous section, the aim is to determine whether OOP
health spending leads to catastrophic health expenditure i.e. whether OOP health spending
exceeds 20 percent or 40 percent of a household’s ‘capacity to pay.’ In this case, capacit y-to-
pay is defined as total household expenditure minus expenditure on subsistence, essentially
food. The 20 percent and 40 percent thresholds were selected based on empirical precedence
from the literature. Using data from the 2010 and 2015 LCMS, the incidence (headcount) and
intensity (mean overshoot) of catastrophic health expenses due to OOP payments were
calculated and the results are presented in Figures 5 and 6 for the 20 percent and 40 percent
thresholds. The results show that between 2010 and 2015, fewer households experienced
catastrophic health spending at both the 20 percent and 40 percent thresholds (Figures 5A and
5B). However, for the households experiencing catastrophic health spending in 2010 and 2015,
the level of intensity was more profound for the poor at both the 20 percent and 40 percent
thresholds (Figures 6A and 6B). In Appendix 2, results for the 10 percent and 30 percent
thresholds are provided.
To further understand the socio-economic status of the location, facility type, and main
components of OOP spending at household level, results from the 2014 ZHHEUS were used.
The results show that households in rural areas are more likely to experience catastrophic health
spending than those in urban areas (Figure 7A). Further, visiting a private facility and hospitals
(2nd, 3rd, and district level) is more likely to lead to catastrophic health spending as compared
to the other health facilities (Figure 7B). And Figure 8 shows that 42 percent of health-related
spending at household level is on drugs (42 percent) followed by transportation and food (26
percent).
Figure 5: Incidence of catastrophic health spending at household level, 2010-2015
A. Head count at 20% threshold B. Head count at 40% threshold
12.2
9.8
Percentage of households
Percentage of households
5.8
4.8 4.7 4.3
3.6 3.4
3.1 2.7 3.2
1.8 2.2 2.3
1.9 1.8 2.1
1.5
1.0 1.1
Poorest 2nd 3rd 4th Richest
Poorest 2nd 3rd 4th Richest
2010 2015 2010 2015
Source: Authors’ compilation from LCMS 2010 and 2015 data.
16
Figure 6: Intensity of catastrophic health spending at household level, 2010-2015
A. Mean Overshoot at 20% threshold B. Mean Overshoot at 40% threshold
21.6 20.8
Percentage of households
Percentage of households
5.6 6.0 5.2 5.6
1.5 1.4 1.1 1.0 1.2 1.1 1.0 0.8 0.7 0.5 0.4 0.9 0.6 0.7
Poorest 2nd 3rd 4th Richest Poorest 2nd 3rd 4th Richest
2010 2015 2010 2015
Source: Authors’ compilation from LCMS 2010 and 2015 data.
Figure 7: Catastrophic health spending by Residence and Facility Type (percent), 2014
A. Residence B. Facility Type
12
10 10.7
8
6.3 8
6
6.2 5.9 5.8
4
2
2.6 1.90
2.6 0
Private 2nd and Public Public Public Public Other
facility 3rd district Health rural urban
level hospital post health health
Urban Rural Hospital centre centre
Source: MoH (2014).
Figure 8: Shares of total household health expenditure (percent), 2014
Consultation Surgery
7% 1%
Medical Exam
6%
Other Drugs
18% 42%
Transport & Food
26%
Source: MoH (2014).
17
5.3 Benefit Incidence Analysis (BIA)
As earlier stated, Zambia has implemented a number of health reforms over the years aimed at
improving access to health care for all Zambians, particularly the poor. This section assesses
the distributional impact of the health reforms in Zambia on public spending and equity across
regions and income groups by using the 2010 and 2015 LCMS, and the 2014 ZHHEUS. By
using multiple surveys, the study evaluates changes in the distributional impact of the health
reforms over a period of time. The study does not look at each individual element of the health
reforms but examines changes in benefit incidence across different income groups over time.
5.3.1 Distribution of health subsidies and outpatient visits at public health facilities by province
Figure 9A shows that four provinces (Luapula, Southern, Copperbelt, and Eastern) recorded a
reduction in their share of total health subsidies in 2015 in comparison to 2010. The largest
reduction in the share of health subsidies was in Copperbelt and Southern provinces at 7 percent
and 5 percent, respectively. Eastern province received the highest share of total health subsidies
from the government in 2010 and 2015 despite a two-percentage point reduction between 2010
and 2015. This is followed by Lusaka and Copperbelt provinces which ranked second and third
overall, respectively.1 Outpatient visits at public health facilities also shows a reduction in four
provinces (Southern, Luapula, Copperbelt, and Eastern) in 2015 as compared to 2010 (Figure
9B). These four provinces had ranked highest in outpatient visits in 2010. The largest reduction
in outpatient visits was observed in Eastern and Copperbelt provinces while the highest gain
of 3 percent was recorded in Lusaka and Central provinces. Eastern province ranked first in the
overall share of outpatient visits for 2010 and 2015 while Lusaka and Southern provinces
ranked second and third, respectively.
Figure 9: Distribution of health subsidies and outpatient visits at public health facilities by
province
A. Distribution of health subsidies B. Outpatient visits
35 35
30 30
25 25
20 20
15 15
10 10
5 5
0 0
2010 2015 2010 2015
Source: Authors’ compilation from LCMS 2010 and 2015 data.
1Lusaka province recorded a 3-percentage point increase between 2010 and 2015
18
5.3.2 Distribution of total subsidies in comparison to reported illnesses at provincial level
To assess if health subsidies are distributed in line with reported illnesses for health care at
provincial level in Zambia, the share of health subsidies for each province was compared with
the share of the population reporting illnesses for each province. Data for this exercise were
drawn from the 2014 ZHHEUS. The results show that distribution of health subsidies at
provincial level is not in line with reported illnesses in each province in Zambia (Figure 10).
Specifically, Eastern, Lusaka, and Copperbelt provinces received a greater share of the
subsidies even though the percentage shares of the population reporting illnesses were
significantly lower. All the other seven (7) provinces, which are predominantly rural, received
a lower share of health subsidies despite having a larger share of the population reporting
illnesses.
Figure 10: Distribution of total subsidies in comparison to reported illnesses by province
100%
6
90% 20 7
8
80%
6
10
70%
12
10
60% 3
5
11
50% 8
40% 11
30% 27 12
20% 12
9
10%
3 13
6
0%
% share of public subsidies % share of population reporting illness
Luapula Northern Western Eastern Central
North-Western Muchinga Copperbelt Southern Lusaka
Source: Authors’ compilation from ZHHEUS data.
5.3.3 Distribution of outpatient and inpatient health care benefits by income groups
Table 4 shows dominance test results for utilization of outpatient health services (or distribution
of health benefits) across the various health providers and facilities. In 2010, mission health
facilities are pro-poor with a concentration index of -0.114 (p-value < 0.05) while public
hospitals and private health facilities are pro-rich with concentration indices of 0.058 (p-value
< 0.01) and 0.324 (p-value < 0.05), respectively. In 2015, results from both the LCMS and
ZHHEUS show that the distribution of benefits is pro-rich at public hospitals and private health
facilities. For mission health facilities, the 2015 LCMS shows a pro-rich distribution of benefits
with a concentration index of 0.093 (p-value < 0.1) while results from the ZHHEUS are
statistically insignificant. However, at 10 percent level of significance, results from the LCMS
are barely statistically significant. On the other hand, overall distribution of benefits at all
public health facilities (hospitals and health centers) was evenly distributed in 2010 with a
concentration index of 0.014 (p-value < 0.01) but became pro-rich in 2014 with a concentration
index of 0.046 (p-value < 0.05).
19
Table 4: Benefit incidence test results – outpatient health services
LCMS ZHHEUS
2010 2015 2014
Provider/facility type CI SE DT CI SE DT CI SE DT
Public
All hospitals 0.058*** 0.028 – 0.048* 0.030 – 0.214*** 0.024 –
Health centers –0.0486 0.011 + –0.023 0.017 + 0.013 0.018 n-Dom
All health facilities 0.014*** 0.009 n- 0.002 0.007 n- 0.046** 0.018 –
(hospitals & health centers) Dom Dom
Mission health facilities –0.114** 0.031 + 0.093* 0.054 – –0.106 0.068 +
Private health facilities 0.324** 0.050 – 0.597*** 0.063 – 0.686*** 0.027 –
***p<0.01; **p<0.05; *p<0.1
Note: CI = Concentration Index; SE = Standard Error; DT = Dominance Test; – means that the 45 degree line dominates
(pro-rich); + means that the concentration curve dominates (pro-poor); n-Dom means non-dominance
LCMS – Living Conditions Monitoring Survey; ZHHEUS – Zambia Household Expenditure and Utilisation Survey
Using results from the 2014 ZHHEUS, health facilities are further broken down by provider
and facility type, and by outpatient and inpatient care. The benefit incidence test results are
shown in Table 5. The results show that the distribution of benefits at all public health facilities
(all types of hospitals and health centers) are generally pro-rich for both inpatient and outpatient
services except for district hospitals and health centers which are pro-poor for inpatient services
with concentration indices of -0.09 (p-value < 0.1) and -0.179 (p-value < 0.01), respectively.
Furthermore, while the results for beneficiary incidence for outpatient services at mission
health facilities are statistically insignificant, results for inpatient services are pro-poor with
concentration index of -0.158 (p-value < 0.1). Meanwhile, the distribution of benefits for both
outpatient and inpatient services at private health facilities is pro-rich.
Table 5: Benefit incidence test results – outpatient and inpatient services - 2014
Outpatient Inpatient
Provider/facility type CI SE DT CI SE DT
Public
Tertiary (3rd level) hospitals 0.523*** 0.065 – 0.528*** 0.044 –
General (2nd level) hospitals 0.385*** 0.032 – 0.222*** 0.033 –
District (1st level) hospitals 0.091** 0.037 – -0.090* 0.052 +
Health centers 0.013 0.018 n-Dom -0.179*** 0.022 +
All hospitals (3rd+2nd+1st) 0.214*** 0.024 – 0.243*** 0.015 –
All health facilities (hospitals & health centers) 0.046** 0.018 – 0.160*** 0.017 –
All health facilities (inpatient & outpatient) 0.059*** 0.018 –
Mission health facilities -0.106 0.068 + -0.158* 0.091 +
Private health facilities 0.686*** 0.027 – 0.804*** 0.071 –
***p<0.01; **p<0.05; *p<0.1
Note: CI = Concentration Index; SE = Standard Error; DT = Dominance Test; – means that the 45 degree line dominates
(pro-rich); + means that the concentration curve dominates (pro-poor); n-Dom means non-dominance
20
5.3.4 Distribution of total benefits in comparison to need by income groups
To further assess the distribution of health care benefits, we compared the need for health care
with the benefits received by wealth quintile (Figure 11). Overall, there has been an
improvement in the cumulative proportion of the population receiving benefits relative to their
need. The lowest or poorest 60 percent of the population received a lower share of benefits
relative to their share of need in 2010 (Figure 11A), but the situation improved in 2015 with
only the poorest 40 percent of the population receiving a lower share of benefits relative to
their share of need (Figure 11B). Furthermore, the poorest 20 percent of the population received
a much higher percentage share of benefits in 2015 (22.7 percent) as compared to 2010 when
they received 17 percent of the benefits.
On the other hand, the richest 20 percent of the population received a much lower percentage
share of benefits in 2015 (17.5 percent) as compared to 2010 when they received 18 percent of
the benefits. This suggests that inequities have reduced between 2010 and 2015. Even though
there has a pro-poor redistribution of benefits in 2015 whereby the bottom 20 percent and 40
percent of the population received more than a 20 percent share of benefits in each quintile, the
distribution of benefits is still inappropriate because the lowest two income groups have higher
health needs. For instance, the poorest 20 percent of the population only received 17 percent
of the benefits in 2010 despite having a 18.7 percent share of health need. In 2015, the
percentage share of benefits for the poorest 20 percent of the population increased but the
benefits (22.7 percent) were still less than the health need (23.6 percent). Meanwhile, for the
richest 20 percent of the population, the share of benefits received were relatively higher than
their health needs in both 2010 and 2015.
Figure 11: Distribution of total benefits in comparison to need for health care: 2010 vs 2015
A. 2010 B. 2015
100% 100%
90% 18.0 15.5 90% 17.5 17.1
80% 80%
% share of benefits/need
% share of benefits/need
19.6 17.1 16.6
70% 20.7 70%
60% 60%
23.4 21.0 20.5
50% 22.0 50%
40% 40%
21.8 22.1
30% 22.3 23.1 30%
20% 20%
10% 18.7 10% 22.7 23.6
17.0
0% 0%
% share of benefits % share of need % share of benefits % share of need
Poorest Poor Middle Rich Richest Poorest Poor Middle Rich Richest
Source: Authors’ compilation from LCMS 2010 and 2015 data.
21
6. Conclusion
6.1 Zambia’s health financing incidence is relatively progressive
While several taxes which were assessed (CIT, VAT, PIT) were found to be progressive in
2010, this was not the case for Excise tax which was found to be regressive. Some components
of Excise tax such as cigarette tax is regressive while taxes on alcohol are progressive. The
reason why cigarette tax in Zambia is regressive is because the percentage of adult men
smoking cigarettes is prevalent among the poor. And though PIT was progressive in 2010, its
level of progressivity was weaker than the other taxes i.e. CIT and VAT which were more
progressive. In 2014, PIT was regressive. Considering that income from individuals through
PIT contributed 25 percent of the total annual tax revenue between 2010 and 2014, poor
households bear a larger burden of financing health services through PIT. For OOP health
spending, results shows that it was regressive in 2014 compared to 2010 when OOP was
marginally progressive (almost proportional).
6.2 Catastrophic health expenditures are prevalent among poor households and expose
them to greater financial risk
The results show that between 2010 and 2015, fewer households experienced catastrophic
health spending at both the 20 percent and 40 percent thresholds. However, the poor still
experienced higher incidence of catastrophic health spending than the rich. Further, between
2010 and 2015, the intensity of catastrophic health spending increased for almost all the
quintiles but was significantly higher for the poor. This means that for the households
experiencing catastrophic health spending in 2010 and 2015, the level of intensity was more
profound for the poor. Secondary data from the ZHHEUS (Ministry of Health 2014) shows
that households in rural areas are more likely to experience catastrophic health spending than
those in urban areas. In addition, a visit to a private facility and hospital is more likely to lead
to catastrophic health spending.
Decomposition of health-related household OOP spending further indicates that a huge share
is spent on medicines (42 percent) followed by transportation and food (26 percent) (Ministry
of Health 2014). This suggests that there could be a problem of physical access to health
facilities in Zambia (particularly in rural areas) while the quality of service provided at health
facilities is poor. Lack of medicines at health facilities forces patients to buy drugs from private
pharmacies and drug stores. These findings further suggest that Government policy on free
access to health care through the removal of user fees is not fully effective.
6.3 Distribution of subsidies and utilization of outpatient services at public health
facilities favours urban provinces
The results show that the distribution of subsidies and utilization of outpatient services at public
health facilities in Zambia has consistently been in favour of urban provinces, with the
exception of Eastern province. This suggests that allocation of resources across the 10
provinces in Zambia is not related to disease burden, poverty levels, size of the population, and
quantity of health facilities in each province. Apparently, disease burden, poverty levels, and
population size are key elements of the district-level resource allocation formula that has been
in use since 2004 in the public health sector in Zambia. Empirical studies that have evaluated
the application of the resource allocation formula in Zambia have concluded that the formula
has not been fully applied and this could be one of the reasons for the variations in the
22
distribution of public health subsidies by province in Zambia. And though Eastern province
had the highest percentage share of the subsidies and outpatient utilization of health services,
health outcomes in this province are among the poorest in Zambia as highlighted in previous
demographic and health surveys. This suggests poor quality of health services in Eastern
provinces and other rural areas, particularly for maternal health services. For example, urban
women were more likely than rural women to be provided information about pregnancy
complications, to be weighed, blood pressure measured, and urine and blood samples taken
during antenatal care.
Gaps in service coverage at facility-level raises questions on the effectiveness of the user fee
removal policy that was designed to increase access, and utilisation of quality health care. Some
studies find no evidence that removal of user fees has increased utilisation of health care in
Zambia, particularly for the poor. Further, service quality—a key factor in boosting utilisation
of health services— is low in Zambia and varied across provinces. This partially explains why
the user fees removal policy has had minimal impact on increased utilisation of health services
in Zambia. Another study shows that the richest 50 percent of the population benefit more from
income transfers that have been triggered by the user fees removal policy than the poor (Lépine,
Lagarde, and Le Nestour 2017). Thus, despite the existence of free health care, Zambians still
incur indirect costs when accessing health care such as transport, food, accommodation for
family members taking care of patients, and purchase of medicines not available at the health
facility (Chama-Chiliba and and Koch 2016).
6.4 Outpatient benefits by provider type and level of health care favours are pro-rich
This study shows that the distribution of outpatient benefits at private health facilities and
public hospitals have continually been in favour of the rich over the period 2010–2015. The
most glaring finding is that the distribution of outpatient benefits at faith-based health facilities
has moved from being pro-poor in 2010 to pro-rich in 2015. This suggests a deterioration in
access to health services by the poor over the years despite the fact that mission health facilities
are funded by government,2 are located in rural areas where most of the poor people reside,
and the services are free. This trend is similar to the distribution of overall benefits at all public
health facilities (hospitals and health centers) which favoured the rich in 2014 despite being
evenly distributed in 2010. Distribution of benefits for both inpatient and outpatient services
also shows that the rich benefit more than the poor at both public and private health facilities.
However, inpatient services for public district hospitals, public health centers, and mission
health facilities are pro-poor.
6.5 Distribution of total health benefits is not commensurate to need for health care
Distribution of total health care benefits received in comparison to need for health care shows
an improvement in the cumulative proportion of the poor population receiving benefits relative
to their need between 2010 and 2015. Nonetheless, the poorest 20 percent of the population
still received lesser health benefits in comparison to their needs in both 2010 and 2015 as
compared to the richer households who received a greater share of health benefits despite
having a lower share of health need.
2 The Zambian government funds all mission health facilities under the Churches Health Association of Zambia (CHAZ)
through a monthly operational grant and salaries for the health workers. CHAZ health facilities enjoy the same privileges as
government health facilities.
23
References
Ataguba, J.E., A.D. Asante, S. Limwattananon, and V. Wiseman. 2018. “How to do (or not to
do)… a health financing incidence analysis.” Health Policy and Planning 33(3): 436-444.
Berki, S.E. 1986. “A look at catastrophic medical expenses and the poor.” Health affairs 5(4):
138-145.
Bredenkamp, C., M. Mendola and M. Gragnolati. 2010. “Catastrophic and impoverishing
effects of health expenditure: new evidence from the Western Balkans.” Health Policy and
Planning 26 (4): 349-356.
Central Statistical Office [Zambia], Ministry of Health [Zambia], ICF International. 2014.
Zambia Demographic and Health Survey 2013-14. Rockville, Maryland, USA: Central
Statistical Office, Ministry of Health, ICF International.
Carasso, B.S., M. Lagarde, C. Cheelo, C. Chansa, and N. Palmer. 2012. “Health worker
perspectives on user fee removal in Zambia.” Human Resources Health 10:40.
Castrol–Leal, F., J. Dayton , L. Demery and K. Mehra. 2000. “Public spending on health in
Africa: Do the poor benefit?” Bulletin of the World Health Organization 78: 66–74.
Chama-Chiliba, C.M., and S.F. Koch. 2016. “An assessment of the effect of user fee policy
reform on facility-based deliveries in rural Zambia.” BMC Research Notes 9(1): 504.
Chansa, C. (forthcoming). “Evolution of the Zambia Health Sector – Key Reforms, Legislature,
& Major Events: 1992-2016”.
Chansa, C., M. Mukanu, C. M. Chama-Chiliba, M. Kamanga, N. Chikwenya, B. Bellows, and
N. Kuunibe. (forthcoming). “Looking at the Bigger Picture: Effect of Performance-Based
Contracting of District Health Services on Equity of Access to Maternal Health Services in
Zambia.”
Chitah, B. and F. Masiye. 2007. Deprivation-based resource allocation criteria in the Zambian
health service: A review of the implementation process. Harare: EQUINET.
Demery L. 2000. Benefit incidence: A Practitioner’s Guide. Washington DC: World Bank
Gilson, L., J. Doherty, S. Lake, D. McIntyre, C. Mwikisa, and S. Thomas. 2003. “The SAZA
study: implementing health financing reform in South Africa and Zambia.” Health Policy and
Planning 18: 31-46.
Jenkins, G.P., C.Y. Kuo, and G. Shukla. 2000. Tax Analysis and Revenue Forecasting.
Cambridge, Massachusetts: Harvard Institute for International Development, Harvard
University.
Lake, S. and C. Musumali. 1999. “Zambia: the role of aid management in sustaining visionary
reform.” Health Policy and Planning 14: 254-63.
24
Lépine, A., M. Lagarde, and A. Le Nestour. 2017. “How effective and fair is user fee removal?
Evidence from Zambia using a pooled synthetic control.” Health economics 2018: (27) 493–
508.
Kalumba K. 1997. Towards an equity-oriented policy of decentralization in health systems
under conditions of turbulence: the case of Zambia. Geneva: World Health Organization.
Masiye, F., O. Kaonga, and J.M. Kirigia. 2016. “Does User Fee Removal Policy Provide
Financial Protection from Catastrophic Health Care Payments? Evidence from Zambia.” PLoS
One. 2016. 11(1):e0146508.
McIntyre, D., and J.E. Ataguba. 2011. “How to do (or not to do) … a benefit incidence
analysis.” Health Policy and Planning 26 (2):174-182.
Ministry of Health. 1991. National Health Policies and Strategies. Lusaka: Ministry of Health.
———. 2014. Zambia Household Health Expenditure and Utilization Survey. Lusaka:
Ministry of Health.
O'Donnell, O., E.V. Doorslaer, A. Wagstaff, and M. Lindelow. 2008. Analyzing health equity
using household survey data: A guide to techniques and their implementation. Washington DC:
World Bank.
Ravallion, M., S. Chen and P. Sangraula. 2008. “Dollar a Day Revisited.” Policy Research
Working Paper 4620. Washington DC: World Bank.
Wagstaff, A. and E.V. Doorslaer. 2003. “Catastrophe and impoverishment in paying for health
care: with applications to Vietnam 1993–1998.” Health economics 12(11): 921-933.
25
Appendix 1: Key health reform areas and elements, Zambia: 1992-2018
Period Organization Finance Provider Payment
1992- - Devolution of health services - Tax-based finance
1993 - SWAp is introduced - Pooling of government &
donor funds for districts
- Medical user fees
introduced with
exemptions for the poor
1995- - Provider-purchaser split - Basic Health Care - Country-wide PBC
1996 - Central Board of Health (CBoH) Package developed
created - Population-based resource
- Functions of Medical Stores allocation formula
Limited (MSL) restricted to developed
storage and distribution
1998- - Functions of CBoH and Ministry
1999 of Health streamlined
- MSL contracted-out under a
lease agreement
2003- - MSL contracted-out under a - Medium Term
2004 management contract Expenditure Framework
- Re-organisation of SWAp - Pooled funding extended
coordination mechanisms to all levels
- Needs-based resource
allocation formula
developed
- Introduction of medical
levy
2006- - Dissolution of CBoH - Some donors transition - PBC discontinued
2007 - The Ministry of Health assumes from pooled funding at the
role of provider, purchaser & Ministry of Health to
regulator General Budget Support at
the Ministry of Finance
- Medical user fees
removed in all rural areas
(2006); and peri-urban
areas (2007
2011- - Transfer of the primary health - Medical user fees - RBF introduced in
2013 care function from the Ministry removed at the entire 11 districts
of Health to the Ministry of primary health care level
Community Development, (2012)
Mother and Child Health - Medical levy abolished
2015- - Re-merger of the primary health - RBF introduced in
2018 care function to the Ministry of 58 districts in five
Health (2015) of the 10 provinces
- Structural re-organization of the
Ministry of Health (2016-2018)
Source: Chansa (forthcoming).
26
Appendix 2: Incidence and Intensity of Catastrophic Health Expenditure
Table A.1: OOP expenditure on health – 2010
OOP health spending as a share of non-food expenditure (monthly)
Threshold 10% 20% 30% 40%
Quintile (1=Poor, 5=Rich) 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5
Head Count (%) 15.0 8.7 8.0 6.5 5.7 12.2 5.8 4.7 4.3 3.6 10.4 4.1 3.1 2.6 2.7 9.8 3.2 2.3 1.8 2.1
Mean Overshoot (%) 2.1 2.1 1.7 1.6 1.5 1.5 1.4 1.1 1.0 1.1 1.1 1.0 0.7 0.6 0.8 0.8 0.7 0.5 0.4 0.6
Mean Positive Overshoot (%) 27.3 25.4 21.9 23.2 27.8 29.9 27.0 24.3 22.7 31.4 33.5 27.6 24.9 24.8 30.8 31.0 25.7 23.2 22.9 28.7
OOP health spending as a share of total household expenditure (monthly)
Head Count (%) 4.6 3.5 3.3 3.1 3.7 2.8 1.8 1.6 1.4 2.1 2.0 1.2 1.1 0.8 1.5 1.7 0.7 0.5 0.5 1.1
Mean Overshoot (%) 0.7 0.6 0.6 0.5 0.9 0.4 0.4 0.3 0.3 0.6 0.3 0.2 0.2 0.2 0.4 0.2 0.1 0.1 0.1 0.3
Mean Positive Overshoot (%) 19.0 17.8 16.7 16.0 23.7 27.0 18.5 17.3 24.8 28.3 26.7 18.9 19.9 27.6 26.6
Source: Authors’ compilation from LCMS 2010 and 2015 data.
Table A.2: OOP expenditure on health – 2015
OOP health spending as a share of non-food expenditure (monthly)
Threshold 10% 20% 30% 40%
Quintile (1=Poor, 5=Rich) 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5
Head Count (%) 6.9 4.8 4.4 3.1 3.3 4.8 3.1 2.7 1.8 2.2 4.1 2.2 2.0 1.2 1.4 3.4 1.9 1.5 1.0 1.1
Mean Overshoot (%) 22.2 6.0 6.3 1.4 1.3 21.6 5.6 6.0 1.2 1.0 21.2 5.4 5.8 1.0 0.9 20.8 5.2 5.6 0.9 0.7
Mean Positive Overshoot (%) 327.6 125.5 143.3 43.5 38.9 456.7 181.1 222.2 64.0 46.1 524.4 242.9 293.8 84.6 62.1 622.0 268.4 380.2 91.5 64.6
OOP health spending as a share of total household expenditure (monthly)
Head Count (%) 3.9 2.3 2.3 1.7 2.5 2.5 1.4 1.3 1.0 1.4 1.7 1.1 0.8 0.7 1.0 1.5 1.1 0.6 0.5 0.8
Mean Overshoot (%) 10.7 1.6 1.5 0.4 0.7 10.4 1.5 1.3 0.3 0.5 10.3 1.3 1.2 0.2 0.4 12.1 1.2 1.1 0.2 0.3
Mean Positive Overshoot (%) 301.7 71.5 64.1 29.6 26.3 482.4 105.1 101.6 40.0 33.4 739.7 116.6 158.3 34.5 46.7 856.6 115.1 203.9 46.7 50.7
Authors’ compilation from LCMS 2010 and 2015 data.
27
Appendix 3: Technical Notes
1.0 Financing Incidence Analysis
1. Adult equivalent in household (AE):
In adjusting for the household size and composition in order to obtain individual estimates
where appropriate the LCMS survey uses the following deflator:
������������ = (������ + ∝ ������)Θ
Where,
A = number of adults in the household
K = the number of children
∝ = ������������������������ ������������ ������ℎ������ ������ℎ������������������������������������
Θ = degree of economies of scale
2. Concentration Index: the concentration index is by definition
1
C = 1 - 2∫0 ������ℎ (p)dp and for a discrete living standards variable is
2 1
C= ∑������
������=1 ℎ������ ������������ 1 −
������������ ������
Where,
Hi = the health sector variable
������ = ������ℎ������ ������������������������
������
ri = = the fractional rank of individual I in the living standards
������
distribution, with I = 1 for the poorest and i=N for the richest
3. Kakwani Index (KI). In order to overcome some limitations imposed by graphs, such as
identifying the variations or relativities in progressivity by various criteria such as type
of source, time or region, it is useful to compute and complement the use of graphs with
some other measure. The KI is such a measure. The KI is defined as twice the area
between a payments’ concentration curve and the Lorenz curve. It is estimated as:
πk = C – G, where C = health payments’ concentration index and G = the Gini
Coefficient of the ATP variable. The value of πk ranges from -2 to +1. A negative
number indicates a regressive relationship; LH (p) lies inside L(p). A positive number
indicates progressivity; LH (p) lies outside L(p). If the relationship is proportional, the
concentration lies on top of the Lorenz curve and the index is 0
The KI is computed from convenient regression of the model:
ℎ������ ������������
2������������ 2 [ − ]
������ ������
Progressivity of funding sources
According to O’Donnell et al. (2008, p.193), the overall or total progressivity of health
financing depends on both the progressivity of the different sources of finance and on the
proportion of revenue collected from each of the financing sources. This means that there is
need to understand the overall progressivity of the tax system as a whole especially in cases
where the public health system relies extensively on public financing. Determining who pays
for the final cost of health care provision is critical in understanding progressivity.
28
Concentration Curve and Concentration Index: Concentration curves (CC) can be used to
determine the extent of socio-economic inequality within a chosen health outcome variable and
how it varies over time (O’Donnell et al. 2008). However, concentration curves do not measure
the extent of inequality which can be compared across region, time or other comparative basis.
The concentration index (CI) is defined relative to the CC. This is twice the area between the
CC and the line of equality (45o line). The index is zero with no socio-economic inequality,
negative when the curve is above the line of equality indicative or there is inequality, indicative
of pro-poor dimensions of the variable under consideration. The index has bounds of -1 and
=+1.
KI: The Lorenz curve (LC) shows the proportion of health expenditures or income attributable
to cumulative shares of the population. It has a range (0,1). Dominance of the LC over another
distribution which holds for the CC as well occurs, when for any given distribution of the
population, p, the LC of a given income distribution is above that of the other distribution. This
implies that the dominating LC has a distribution with less inequality.
OOP expenditures: These are the household expenditures for all household related
consumption.
ATP: This is given as the difference between expenditures on non-discretionary items less
expenditures on food and food related items.
2.0 Benefit Incidence Analysis - Constant Unit Cost Assumption
This study uses a repeated cross-sectional survey design and applies the traditional BIA
methodology (McIntyre and Ataguba 2011; O’Donnell et al. 2008) to assess the distribution of
public subsidies and service benefits (utilization of health services). Listed below are the key
steps and activities which were undertaken.
i. Using household expenditure as a measure of socio-economic status, quintiles were
constructed and used to rank the population by wealth;
ii. Data on the utilization of health services was disaggregated by provider, level of health
care, outpatient/inpatient, and socio-economic status;
iii. Unit costs for outpatient and inpatient services was calculated by using expenditure
data, population, and utilization rates;
iv. “Benefits” were calculated by expressing utilization of health services in monetary
terms by multiplying utilization rates by unit costs for each socio-economic group. The
“benefits” were then aggregated across different types of health services for each socio-
economic group; and
v. Comparing the distribution of health expenditures (subsidies) and benefits by province,
providers, type of health services; and for the different socio-economic groups in order
to determine differences in benefit incidence and with respect to need.
This study uses constant unit subsidies and adapts the generic formula below from O’Donnell
et al. (2008).
������ ������
������������ ������������������
������������ ≡ ∑ ������������������ ≡ ∑ ������ (1)
������������ ������������ ������
������=1 ������=1
29
Where:
Xj = the value of total health subsidy imputed to the socio-economic group, j
Hij = number of health visits of group j to health facilities at level i (with i = health
facility type)
Hi =is the total number of visits by different levels of health care by different income
groups
Si = government recurrent net spending (less all private payments)
������������
= unit subsidy of funding health subsidy at level ������.
������������
The share of total health subsidy (S), accruing to the groups is given by the formula below
������������������ ������������
������������ = ∑������
������=1 ( ) ≡ ∑������
������=1 ℎ������������ ������������ (2)
������������ ������
From this equation, the share of total health subsidy to each group is determined by two factors:
(i) share of the group within the context of the total health visits at each level of care (hij) and
(ii) the share for each level of care in total health subsidy (si).
From the first equation, the provincial or regional analysis is derived as follows:
������ ������ ������ ������
������������������������ ������������������
������������ = ∑ ∑ ( ) ≡ ∑ ∑ ������������������������ ������������������
������������ ������
������=1 ������=1 ������=1 ������=1
In which k refers to the region specified in the unit subsidy, and n depicts the number of
provinces (regions) under consideration which in this case is ten (10). An assumption made in
the literature is that the unit subsidy ������ij is constant across all units of type i.
3.0 Health Expenditures and Impoverishment Effects
The principal interest or objective of the analysis is underpinned in the values and principles
centered around providing health care to all persons in a fair and equitable manner so as to
ensure dignity for all individuals. This extends the argument made by Berki (1986) that
financial catastrophic expenditures may affect a relatively small share of the population, and
still account for a substantial share of health expenditures. Yet there is limited evidence of the
extent and significance of health expenditures and their impoverishing effects (O’Donnell et
al. 2008). Although O’Donnell focuses on Asia, the circumstances hold equally for Sub-
Saharan Africa.
30
OOP expenditures
These are payments made by households at the point of service. OOPs will usually include
expenditures related to clinicians’ consultation fees, payment for medicines, payment for
clinical services incurred such as laboratory diagnosis, x-rays and hospitalisation. Insurance re-
imbursement are normally netted out of OOPs.
Household consumption expenditure (HCExp)
These are the household monetary and in-kind consumption and payment for all goods and
services.
Food expenditures
This is the amount of money spent in all foodstuff including own production monetary
equivalent by households. It excludes expenditures on cigarettes, alcohol and restaurant served
foodstuffs.
Poverty line and household subsistence expenses
Household subsistence spending is defined as the minimum requirement to maintain basic life.
This is defined at the internationally accepted conversion of US$1.25 (Ravallion, Chen,
Sangraula 2008). This minimum level poverty line has been adjusted with a purchasing power
parity rate of US$1 = ZMK10.00. The approach used in defining poverty is the extent of the
food and non-food expenditure in the household. The food share expenditures as a proportion
of total expenditures are taken as median expenditures within the range 45–55 percent. In
deriving the estimates, the following variables are used:
• Household size – the total number of household members at the time of the survey = N
• Stratum – the lowest sampling unit = the standard enumeration area
������������������������ ������������������������������������������������������������������������ ������������������ ������ℎ������ ℎ������������������������ℎ������������������ ������������������������ℎ
• Food expenditure = =
������������������������������ ������������������������������������������������������������������������ ������������ ������ℎ������ ℎ������������������������ℎ������������������ ������������������ℎ
• The household food expenditure equivalent = household food expenditure over
household size:
������������������������ℎ
Household food expenditure equivalent = ������������������������������������������������������������ ������������������������ ℎ
• A household is defined as poor when the total household expenditure is less than its
subsistence spending
poorh = 1, if exph < seh,, where seh = pl*eqsizeh. seh =subsistence expenditure for
each household
poorh = 0, if exph > seh,
Pl = weighted average of food expenditure
• The household’s ATP = the household’s non-subsistence spending and is the non-
subsistence effective income.
• OOP payments share of household ATP (oopatp)
31
• OOP payment share of household ATP: the burden of health payments is defined as the
OOP payments share of the household’s ATP
������������������
oopatp = ������������������ ℎ
ℎ
• Catastrophic health expenditures: These occur when a household’s total OOP
payments equal or exceed a given threshold, z, (30 – 40 percent) of the household’s
ATP or non-subsistence expenditure. The threshold, z, can adapted and be over a range
of values.
Adopting the methodology by O’Donnell et al. (2008), we define incidence and intensity of
catastrophic health expenditures as a share of health care costs in total expenditures and/or non-
food expenditures relative to a defined threshold (see above). Conceptually this is shown in
Figure A.1 below. In the figure, the horizontal axis shows the cumulative share of households
������
ordered in a ratio ������ from the largest to the smallest (equivalent to the cumulative density
function for the reciprocal of the health payments budget share with the axes reversed). The
interpretation of the graph is that at the threshold, z, exists the fraction H of households with
health care budget shares that exceed the threshold, z, which is equal to the catastrophic
payment head count. If an indicator E, is defined such that
������
E = 1 if, ������������ > ������ ������������������ 0, ������������ℎ������������������������������������, ������ℎ������������ ������ℎ������ ������������������������������������������������ ������������ ������ℎ������ ℎ������������������������������������������������ ������������
������
������
1
������ = ∑ Ei, where N is the sample size
������
������=1
Where, T = OOP
������ = ������������������������������ ℎ������������������������ℎ������������������ ������������������������������������������������������������������
������(������) = ������������������������ ������������������������������������������������������������������������ = ������������������ − ������������������������������������������������������������������������������ ������������������������������������������������������������������,
it follows that the household experiences catastrophic expenditures where
������⁄ or ������⁄
������ [������ − ������(������)]
������������������������������������������ ������ ������������������������������ ������ℎ������������������ℎ������������������, ������,
where z = represents expenditure levels where the absorption of household resources by
spending on health care imposes a severe disruption of living standards or the ATP is severely
limited.
The catastrophic payment overshoot captures the average by which payments (as a proportion
of total expenditures) exceed the threshold z.
������
������ℎ������ ℎ������������������������ℎ������������������ ������������������������������ℎ������������������ = ������������ = ������������ ( ������⁄������������ ) − ������, ������ℎ������ ������������������������������ℎ������������������ ������������ ������ℎ������ ������������������������������������������:
������
1
Oi = ∑ Oi, where N is the sample size
������ ������=1
According to O’Donnell et al. (2008), O which is shown in Figure A.1, is given as the area
under the payment share curve and above the threshold. Further, since H captures the incidence
of catastrophe occurrence, O captures the intensity of occurrence and are related through the
mean positive overshoot (MPO), which is defined as:
������������������ ������
������
32
Figure A.1: Health Payments Budget Share against Cumulative Percent of Households
Ranked by Decreasing Budget Share
Payment as a share of household expenditure
total catastrophic overshoot, O
Proportion H exceeding threshold
Cumulative percent of population, ranked by decreasing payment fraction
Source: O’Donnell et al. (2008).
33