Over the past few years, Kenyans have engaged in an on-again, off-again debate about the “wage bill” and the “cost of the new Constitution.”

One of the big unknowns in these larger discussions about the price of government is the cost of running counties in particular. This matters for several reasons.

First, we want to know how much these new structures of government contribute to the cost of the Constitution. While we may carry on about constitutional commissions and our bloated parliament, the real money is actually in devolution.

A second reason why county running costs matter is that devolution presupposes an active citizenry at the county level that is engaged in budgeting. This requires transparency about running costs in each county (among other items) for purposes of public debate about annual priorities.

Moreover, the share of county costs that goes to operations versus the share that goes to services is a far more important issue than the others that claim our attention: The recurrent versus development split, or the “wage bill.”

A lot of the core services we care about, such as health or education or agriculture, require heavy wage expenditure for the doctors, teachers, and extension workers that make these services valuable.

Other recurrent spending on medicines and fertilisers is also essential. We should not mind this. What matters is not the recurrent to capital ratio, but the share of the budget that goes to administrative overheads and the share that goes to actual service delivery.

Finally, county running costs are important because they are the basis for the “basic equal share” in the revenue sharing formula that determines how much each county receives annually from the total share for counties.

The theory behind this parameter is that all counties have certain fixed costs that are basically the same regardless of their other characteristics, and all of them should receive an equal share for these costs.

For example, every county, no matter how rich or poor, or large or small, has to pay its governor the same salary. Nevertheless, the size of the running costs that are the same across all counties is unknown, and it is therefore unclear whether the current 25 per cent weight for the equal share is appropriate.

Recognising these imperatives, my organisation analysed available data to try to estimate county running costs. We released a brief last week that provides estimates for each county on an annual basis.

The main data sources used were circulars from the Salaries and Remuneration Commission, as well as some estimates by the Commission on Revenue Allocation and the National Treasury. Unfortunately, while all of this data is theoretically public, it has been difficult to find for ordinary citizens and has not been consolidated in a single place. That was the gap we sought to rectify.

So what did we find?

First, the topline figure: We estimate that it costs about Ksh48.5 billion per year to run the counties. These costs are only for the salaries and benefits of state officers and administrative officers on the executive and assembly side, plus operational costs.

They do not include any costs of service delivery, like health services. These figures are not perfect, due to lack of data and guidelines, but we believe they are fairly close to the true costs.

To put this figure in context, it is approximately 19 per cent of the total equitable share received by counties this year (Ksh260 billion) and a little over 2 per cent of the total national budget (including county transfers). The average county costs approximately Ksh1 billion a year to run, and running costs range from Ksh695 million a year in Lamu/Isiolo to Ksh1.8 billion a year in Nairobi.

So what are the implications of this? First, the overall cost of running counties is not prohibitive in the context of the national budget. At the same time, we would probably not want to see running costs that are more than 20 per cent of county budgets, so there is a call for vigilance around rising administrative expenditure at county level.

Finally, the basic equal share in the formula is too large. The data shows wide variations in individual county costs, which are mainly caused by differences in population and therefore assembly size. It is not appropriate to give counties an equal share for all running costs, but only those that are identical across all counties (otherwise we are punishing larger counties).

We would therefore anchor the basic equal share on the lowest county. If we multiply Lamu’s running costs by 47, we get roughly Ksh33 billion, which is about 13 per cent of the current equitable share. This means that the current 25 per cent is far above what is needed for this purpose and should be reduced as the formula is revised.

Jason Lakin is Kenya country manager of the International Budget Partnership. E-mail: [email protected]