Current Revenue Sharing Practices in Kenya: How Fair? John Kinuthia Research Analyst 20/09/2016

Transcription

1 Current Revenue Sharing Practices in Kenya: How Fair? John Kinuthia Research Analyst 20/09/2016

2 Equity: From Principles to Practice We have discussed 6 key principles of equity Most revenue sharing approaches globally incorporate need Many incorporate at least some of the other 5 principles as well Let s look at how well Kenya s national and county programs incorporate principles in practice 2

3 National Level National Constituency Development Fund Level 5 Hospitals Equalization Fund County Level Equitable Development Act - Elgeyo-Marakwet Ward Development Funds in Baringo and Meru 3

4 National 4

5 National Government Constituency Development Fund (CDF) CDF is 2.5% of the national government share of revenue. 75% of the fund is distributed equally among the constituencies and 25% is shared based on the number of poor people as a share of the national total. CDF is mainly a capital fund but there are no direct measures of infrastructure needs in the distribution criteria. 5

6 Is the CDF distribution mechanism fair? CDF does not directly measure needs for all the issues it allocates money to. The high equal distribution does not take into account population differences. Lamu East has the lowest number of poor people but receives 20 times the per capita allocation in Mandera South which has the highest poor population. The poverty data used is adjusted by a factor of 0.23 for urban areas in favour of rural constituencies. Not clear why this adjustment was made 6

7 Level 5 Hospitals 7 former provincial hospitals and 4 other high volume facilities. They are managed by host counties but funded through a conditional grant. Before devolution allocations to level 5 facilities were based on the formula below Variable Weight Poverty 20% Beds Utilized 40% Outpatient Cases 20% Accident Prone Facilities 5% Fuel Costs 15% Total 100% 7

8 After devolution In 2013/14 and 2014/15 the criteria used has been unclear. In 2015/16 and 2016/17 the allocations were based on the bed occupancy rates. The higher the rate, the more money a facility receives. 8

9 Is the Level 5 grant distributed fairly? The bed occupancy rates ignore the actual number of beds whose occupancy drives costs. For example Nakuru and Meru have the same occupancy rates but Nakuru has almost twice the number of beds than Meru. However both facilities received Ksh 356 million each in 2016/17. Is this fair? In addition these facilities provide other services that drive their costs beside inpatient services 9

10 Share of allocations in 2016/17 based on bed occupancy Facility Allocation Based on CARB 2016 Share of County Number of Occupied Beds Bed Occupancy County Allocation Allocation based on Share of Occupied as a Share of the Total Rates Bed capacity Occupied Beds (Million) the CARB Criteria Beds to the Total (Million) Machakos 79% % 8% 339 Embu 62% % 11% 439 Garissa 71% % 5% 182 Kakamega 88% % 11% 453 Meru 77% % 7% 270 Mombasa 80% % 11% 457 Nakuru 77% % 13% 519 Nyeri 84% % 8% 311 Kisumu 76% % 10% 398 Thika 85% % 6% 258 Kisii 86% % 9% 373 Total 4,483 3,491 4, % 100% 4,

11 Equalization Fund EF established by constitution to target marginalized areas: access to health facilities, water, roads and electricity lags behind other areas CRA marginalization policy identifies 14 counties to receive funds based on assessment of development and historical marginalization. Half of the fund is shared equally and the other half is shared based on a formula based criteria. 11

12 Equalization Fund distribution parameters Category/Dimensions Weight Indicator Weight Poverty gap 16% i Poverty 16.0% Infrastructure 28% i Roads 9.3% ii Electricity 9.3% iii Water 9.3% Health 28% i Immunization 9.3% ii Sanitation 9.3% iii Deliveries in Health Facilities 9.3% Education 28% i Literacy 14.0% ii Secondary education 14.0% Total 100% 100% 12

13 Is the distribution of the Equalization Fund fair? The marginalization policy includes parameters that are not in the constitution: education and poverty. The CRA policy does not explain how these weights were arrived at. For example, health and education are given the same weight of 28 percent. Why? Why are certain variables used for the parameters and not others? For example, immunization is part of the health parameter but child nutrition is not. 13

14 EF continued The selection of counties as the measurement unit locks out poor areas in non-marginalized counties. When shared equitably the range of allocations is between 5%-11% When shared with the equal share the range reduces to 6%-9%. 14

15 Counties 15

16 County approaches to revenue distribution Two common approaches to distribution by counties: First, mimic the CRA formula in some form Second, adapt the CDF criteria in some form Many counties are not fully implementing these laws but they give an idea of thinking about equity 16

17 Quick review: CRA and CDF The first CRA formula had five parameters that determined the horizontal allocation of revenue among counties. The parameters are population, poverty, basic equal share, land area and fiscal responsibility. CDF approach has a big proportion of the fund shared equally (75%) and remainder shared based on poverty numbers in each constituency. 17

18 EDA in Elgeyo-Marakwet The Equitable Development Act guides distribution of the full development budget among EMC wards. 60% of development budget in each department is shared equally and 40% is shared based on a formula borrowed from CRA s 2012 criteria. However, the EDA appears to be more re-distributive compared to the CRA formula. More weight on poverty and less on population 18

19 EDA Formula EMC CRA (2012) Parameters Weights Weights Population 38% 45% County Flagship Projects 23% Poverty Index 22% 20% Land Area 8% 8% Emergencies 5% Fiscal Responsibility 2% 2% Arid and Semi-Arid Areas 2% 19

20 Is the EDA Fair? The formula does not directly address the needs and the high equal share assumes the development needs in wards are almost similar. Example: the share of people without access to water ranges from 2%-9%,while the allocations to wards in that department is between 2%-6%. 20

21 Is the EDA fair? (continued) The county distributes the whole development budget through the EDA: leaves very little funding for countywide projects. It does seem the county changed the formula (informally) to allocate more resources to county flagship projects. The adjustments made in the formula are not clear and the county has not published these changes (undermines fair process) 21

22 Baringo County WDF Seems to borrow from CDF in one part and CRA in another part of the law. 22

23 Baringo Approach to Ward Distribution The act requires 10% of the development budget be distributed to wards. Baringo s modified CDF approach requires 85% of the fund be distributed equally and 15% based on ward population. Does not measure poverty (unlike CDF). 23

24 Is the Baringo approach fair? The CDF approach has a very high equal share (75%) but Baringo s is higher at 85%. Population might give an idea of pressure on infrastructure but does not give a true picture of current deficits in the wards. A smaller portion is allocated to the wards compared to EMC but the distribution is highly equal. For example, share of people without access to improved water ranges between 1%-6% Fund allocations per ward range between 3%-4% 24

25 Meru County WDF The sets aside at least 22.5% of ordinary revenue to a Ward Development Fund. 85% of the fund is to be shared equally and the remaining 15% shared based on population size, poverty levels, and infrastructure differences among the wards. 25

26 How fair is Meru s approach? Meru is the only county among the three analyzed that has a direct measure of infrastructure need in its distribution criteria. Use of an infrastructure differences parameter means that Meru s fund take into account infrastructure needs directly. But, the very high equal share means the formula is unlikely to be equitable. 26

27 How does Kenya compare globally? Look at India In recent years, India's formulas have taken into account fiscal need, capacity and effort. There has been a relatively high weight attached to fiscal capacity compared to other parameters over time. Notable that India has deemphasized fiscal discipline in its current formula, which is an area where many Kenyans believe there should be more weight. 27

28 India s revenue sharing formula Criteria and Weights 13th Finance Commission (2007) 14th Finance Commission (2013) Population 25% 17.50% Demographic Change Not Included 10% Fiscal Capacity Distance 47.50% 50% Area 10% 15% Forest Cover Not Included 7.50% Fiscal Discipline 17.50% Not Included Total 100% 100% 28

29 South Africa South Africa s provincial revenue sharing approach mainly emphasizes fiscal needs. South Africa uses parameters that measure actual service needs, such as the number of school age children as a measure of education needs. The weight for the equal share in Kenya s formula (now 26%) is more than five times that of South Africa (5%) 29

30 South Africa s revenue sharing formula Component Weights Education 48% Health 27% Basic (Shared based on population) 16% Institutional (Equal share) 5% Poverty 3% Economic Output 1% Total 100% 30

31 Conclusions Fiscal need is the main driver of resource distribution in Kenya, but not always well-measured. Many programs use different approaches to measuring poverty and marginalization, not all of which are consistent or fair. Kenya tends to put inordinate weight on equality over equity. Equality stands out when compared to the low weight given to equal shares in South Africa and India. Little attention has been given to fiscal capacity or fiscal effort in Kenya, though these are given considerable weight in other countries like India 31

32 At County Level Counties are making attempts to share revenue among their wards and are borrowing from CDF and the CRA formula. This approach has meant some measure of need is applied by the formulas However, equality seems to outweigh equity in all the counties. Projects funded in each ward are determined exclusive of the main budget process and this parallel process might create challenges (as it has with CDF) 32