Multi-tier government — central, regional, local — exists almost everywhere. The economic question is how to assign taxing and spending powers across levels so that each level has the right incentives, capacity, and resources. Kenya's 2010 Constitution made devolution explicit; this module covers what that means in public-finance terms.
The assignment problem
Two questions:
- Expenditure assignment — which goods and services should be provided by which level? (Healthcare, education, roads, water, security, agriculture extension, etc.)
- Revenue assignment — which taxes should each level collect? (Income tax, VAT, property tax, motor-vehicle licensing, business permits, hotel levies, etc.)
These should align in the abstract, but they rarely do — the level that should provide a service often can't efficiently raise the tax revenue to fund it. The gap is closed by intergovernmental transfers.
Subsidiarity and the spending-assignment rules
Subsidiarity — provide services at the lowest level of government capable of providing them efficiently. The canonical Oates (1972) rules for expenditure assignment:
- Goods with national-scale economies or significant spillovers → central. National defence, foreign affairs, monetary policy, top-tier research universities, major infrastructure with cross-region externalities
- Goods with regional-scale economies → regional / state. Provincial roads, regional health-services administration, tertiary education in non-flagship institutions
- Goods with local-scale economies and clear local-preference variation → local. Primary education, primary health, water and sanitation, local roads, refuse collection, planning, business permits, markets
- Redistributive functions → central. Local governments competing for taxpayers race to the bottom on income redistribution. Cash transfers, unemployment insurance, social pensions belong centrally
Tax-assignment rules
Musgrave and Musgrave (1989) propose:
- Progressive redistributive taxes (PIT, CIT) → central. Mobile bases that would erode under local-level rate differentials. Income-tax competition between states/counties races rates to the bottom
- Stabilisation-relevant taxes (CIT, VAT) → central. Macroeconomic management needs a single hand on these instruments
- Taxes on immobile bases (property, land) → local. Land doesn't move; local property tax doesn't trigger flight
- User charges and benefit-based taxes → local. Bus fares, market fees, parking fees, water tariffs. Direct link between payment and service
- Natural-resource revenues → central in most cases (with revenue-sharing to producing regions), because resources are geographically concentrated and revenue-volatile
The vertical fiscal imbalance
Spending assignment and revenue assignment rarely produce an exact match at each level. The Oates rules push spending DOWN (local provision of local services) but the Musgrave rules push revenue UP (central collection of the major tax bases). The structural result is vertical fiscal imbalance: lower levels are assigned more spending than they can finance from their own-source revenue. The gap is filled by transfers from the centre — and the design of those transfers shapes the political economy of devolution.
Kenya's devolution architecture
Kenya's 2010 Constitution created 47 county governments alongside the national government. The Fourth Schedule assigns functions:
- National government functions (38 items) — foreign affairs, defence, monetary policy, national education policy, immigration, citizenship, judiciary, elections, statistics, national security, etc.
- County government functions (14 items) — agriculture, county health services, cultural activities, county transport, animal control, trade development and regulation, county planning, education (pre-primary, village polytechnics, day-care, home-craft centres), implementation of national policies on natural resources, public works at county level, fire-fighting and disaster, control of drugs and pornography, ensuring participation of communities in governance
The equitable share — Article 203
The Kenya Constitution at Article 203 requires that 'not less than fifteen per cent of all revenue collected by the national government' be allocated to counties as their 'equitable share'. The Commission on Revenue Allocation (CRA) recommends the actual share and the allocation formula; Parliament adopts.
The CRA formula — Third Generation (2020-2025)
The horizontal allocation across the 47 counties uses a weighted formula. Third Generation weights: • Population (45%) • Equal share (20%) — each county gets the same base • Poverty (16%) — share of national poverty headcount • Land area (8%) • Development factor (1%) — adjustment for development gap • Fiscal effort (2%) — rewards counties that raise their own-source revenue • Fiscal prudence (2%) — rewards counties with low pending bills and clean audits • Health (6%) — burden of disease Fourth Generation formula (under negotiation 2024-2025) is expected to increase the population weight, reduce equal-share, and introduce explicit revenue-effort incentives. The political economy is intense — formula changes shift hundreds of millions of shillings between counties.
Own-source revenue (OSR) of counties
Counties are assigned constitutional powers to levy:
- Property rates (land rates on urban property)
- Entertainment taxes
- Any other tax that the county is authorised to impose by an Act of Parliament — e.g., hotel-bed tax, advertisement fees, agricultural produce cess (rationalised post-2018 to remove cascading)
- Charges for services rendered (market fees, parking fees, business permits, public-transport fees)
OSR averages 10-25% of total county revenue across counties — the rest comes from the equitable share and conditional grants. Nairobi, Mombasa, and Kiambu are the highest OSR collectors; the predominantly rural counties (Tana River, Mandera, Wajir) have OSR in the 5-10% range.
Horizontal fiscal imbalance and equalisation
Beyond the vertical gap (central vs counties), there's a horizontal gap — some counties have richer tax bases or higher service-delivery costs than others. The constitutional response is the Equalisation Fund (Article 204): a fund equal to 0.5% of national revenue, used over 20 years from 2013 to bring marginalised areas up to the national average in basic services.
The 14 counties identified as eligible (the 'marginalised counties') receive Equalisation Fund grants on top of the equitable share. The Fund has been chronically slow to disburse and the original 20-year timeline (ending 2033) is widely seen as aspirational.
Conditional grants
Beyond the unconditional equitable share, the centre transfers conditional grants — money tied to specific spending purposes (health-sector grants for level-5 hospitals, road-maintenance grants from the Roads Fund). These improve target-specific spending but reduce county autonomy. The balance between unconditional and conditional transfers is a permanent policy question.
Exercise
Suppose the National Treasury proposes to amend the CRA Fourth Generation formula to weight as follows: Population 55%, Equal share 12%, Poverty 12%, Land area 7%, Fiscal effort 7%, Fiscal prudence 5%, Health 2%. The current Third Generation weights are population 45%, equal share 20%, poverty 16%, land area 8%, development 1%, fiscal effort 2%, fiscal prudence 2%, health 6%. (1) Identify which categories of counties (by size, geography, OSR capacity) would gain or lose. (2) What's the public-finance argument FOR the higher population weight? (3) What's the public-finance argument AGAINST it? (4) Recommend a position with reasoning.