The call for more money for Kenya’s counties has resonated with many citizens. This reflexive support for funding devolution is more a leap of faith than a rational calculation.

Kenyans are not particularly confident that governors are managing their funds well. Nor do they have any particular information about the basis for the current funding going to counties or why it might or might not be adequate. But just as one instinctively leaps to the defence of a blood relative, defending funds for counties is more of a litmus test of one’s dedication to devolution than an analytic position.

In a sense, this instinctive circling of the wagons is one of Kenya’s greatest assets to ensure that devolution prospers. State reform is an inherently political process, and sentiment and perception are crucial in determining outcomes.

At the same time, instincts should not blind us to facts. If we do not understand the figures behind our demands, we can actually end up undermining our own cause by gutting existing institutions and disrupting services, leading to a backlash against devolution that brings in its wake recentralisation of key functions.

So what are the facts about how revenue is shared in Kenya today?

Fact #1. The revenue that counties receive is not determined by the last audited and approved accounts of Parliament.

The best way to think about this issue is the following: There are two steps in the process of determining the equitable share that goes to counties (the so-called “15 per cent”). First, we estimate the costs of performing county functions. Second, we check to see if it is more than 15 per cent of the last audited and approved accounts.

The first stage is by far the most important, but has gotten almost no attention, while the second stage is almost pro forma, but has gotten all the attention. This is how politics blinds us to reality.

How does the first step work? Actually, it is based on the 2012/13 budget year. The 2012/13 year is used as it is the last year in which the national budget contained a consolidated assessment of how much Kenya spends on all functions of government.

Why? Because from 2013/14, various functions had been devolved to counties, and the spending on those functions now appears in county budgets, not the national budget. If we want to know how much the government devoted to all functions, both national and county, relative to each other (to calculate the percentage for each level of government), we need to carefully examine the 2012/13 budget and try to assign the costs in that budget to each level of government based on their functions.

National Treasury did exactly this in 2012/13, coding everything that it thought should be devolved with the number “98.” Now, the intention was that the Transition Authority would lead a process with line ministries of validating and improving on the 98 coding. But this never happened.

Therefore, the figures that Treasury proposed in 2013/14 for counties were based on this 98 coding, with some adjustments for administrative costs at county level and inflation.

Recommendations

This is what formed the recommendations to Parliament in that year. The Commission on Revenue Allocation’s (CRA) own recommendations were also based on the 98 coding, but their figures were higher (Ksh231 billion) because they added more for administration and various “contingencies.”

The figures from Treasury were imperfect. They included things that shouldn’t have been included, like donor funds, that later had to be removed. They also included Constituency Development Fund (CDF), which Treasury expected to be devolved but was maintained.

The inclusion of CDF alone means that the county share was overestimated by more than Ksh23 billion. On the other hand, because line ministries did not do their jobs, and wanted to hold on to certain functions, it is also possible that Treasury excluded funds that should have been coded 98.

Using its imperfect analysis, Treasury initially proposed Ksh199 billion for counties (a mix of unconditional and conditional grants). After political negotiations, the figure was raised to Ksh210 billion (with a different mix of conditional and unconditional grants).

Because of Treasury’s errors in including donor funds, the final amount that went to counties was actually only Ksh190 billion in equitable share (unconditional), and a Ksh3.4 billion conditional grant for regional hospitals.

It is important that we understand that the final revenue given to counties in 2013/14 was over 20 per cent of estimated revenues for 2013/14.

I point this out because we have not even gotten to step 2 in the process yet, and we already have over the 15 per cent minimum for current year revenues (as opposed to what the constitution requires, which is 15 per cent of last audited and approved). Indeed, the truth is that when we get a result like this in step 1, we don’t even need to go to step 2, because we already know the answer.

For clarity, however, let us go to step 2. Since total shareable revenue grows every year, we know that if the county share is above 15 per cent of current revenues, it must be above 15 per cent of every previous year’s revenues. In examining this, however, we are led into the next fact.

Fact #2. Every serious Kenyan institution (Parliament, Treasury and CRA) erred in 2013/14 in carrying out step 2 by using the 2010/11 audited accounts as the base for calculating the 15 per cent.

If you look at what Parliament approved in 2013, the Act states clearly the county allocations as a share of last audited and approved, and of “last audited but not approved”. The years used for these calculations were 2010/11 and 2011/12, respectively.

This year, however, we learned that Parliament had not actually approved 2010/11 and should have used 2009/10 accounts in 2013. Indeed, by the time of the 2014/15 division, it still had not approved the 2010/11 accounts. So step 2 in 2013 and 2014 should have been carried out using the 2009/10 accounts.

Now, because of this mistake, some politicians have been claiming that they increased the percentage share going to counties by more than they have. In 2013/14, the percentage share using 2010/11 accounts was 31.20 per cent.

In 2014/15, using 2009/10 accounts, the share was 43 per cent. But of course, we cannot compare these two numbers. The proper thing to do is to recalculate 2013/14 numbers using the 2009/10 base. When we do this, we find that the share in 2013/14 was actually almost 37 per cent.

This demonstrates that counties did increase their share of funding against a common base, but this percentage is not particularly meaningful in the context of the overall budget, since the base is so old.

A more meaningful comparison between the two years is to look at shares of current revenues. We have seen that counties received about 21 per cent of current revenues in 2013/14. In 2014/15, they received 22 per cent of current revenues.

This gives us a better picture of how county funds are changing relative to the total size of government, and indicates a small increment in the county share of total public spending.

While this may not satisfy those who think counties need more resources relative to national government, it is also not consistent with a narrative of national government trying to squeeze counties.

Fact #3. The total amount going to counties has always been projected to be about 20 per cent of current revenues and it is inappropriate to speak of the county share relative to the whole budget.

The discourse circulating in some corners that counties only get 12 per cent of the budget misunderstands the structure of the national budget.

If we look at the national budget in 2014/15, the total consists of roughly three parts: what we spend on ministries, what we transfer to counties, and what we spend on debt repayment and other obligatory expenditure. The budget also includes donor funds that cannot be devolved.

In 2014/15, the budget is about Ksh1.7 trillion. About Ksh325 billion is allocated for debt payments, plus another Ksh37 billion for pensions and other obligatory spending. Ksh1.13 trillion goes to ministries.

Of this, about Ksh185 billion is foreign-funded. So we should remove from the Ksh1.7 trillion the obligatory expenditure and foreign funding before we look at the county share. When we do this, the budget is closer to Ksh1.2 trillion. Of this, Ksh226.6 billion went to counties.

Using this measure, counties received about 19 per cent of the total budget. This is slightly different than the 22 per cent we calculated before. But recall that the national budget has a deficit, and county shares are based not on total expenditure, but total revenue, which is always smaller than the budget.

Fact #4. Comparing Kenya’s level of funding for counties to other nations is misleading because of the constitutional decision to leave primary and secondary education at national level.
Many Kenyans think we should spend even more on education because they believe there are too few teachers or inadequate classrooms.
If we look at how much goes to education, however, we will see that it is already enormous and (virtually alone) explains why county revenues will be constrained as long as education remains a national function. In 2012/13, Kenya allocated 21 per cent of the total budget for all education functions (national and county), summing to Ksh236 billion. That was more than even CRA recommended for total allocation to counties in 2013/14. In 2014/15, the national government has allocated just over Ksh300 billion to education (including primary, secondary, tertiary and Teachers Service Commission).
Most federal countries have devolved education, which explains why a country like Nigeria might have given much more to states and local government than Kenya. From our figures, it is clear that if Kenya devolved education, the county share of current revenues would increase from roughly 20 to roughly 50 per cent immediately. This calculation also reveals, however, that Kenya’s current allocations to counties are actually exactly where we would expect them to be given that education has not been devolved.

Fact #5. The Council of Governors has not identified Ksh100 billion in funds at national level for devolved functions that should have been given to counties.

Given the fact that we never completed the functional assignment process, and that Kenya’s 2012/13 budget was opaque, it is not easy to know for sure if there are resources at national level that should have been devolved. It is extremely likely that there are some, but the quantum is unknown.

It is also the case that a number of parastatals performing devolved functions, such as those in roads and water, have not been reformed and control resources that could be devolved.

The Council recently published an analysis in the daily newspapers claiming they had identified over Ksh100 billion in national funds that should be devolved. I have already challenged those numbers in this newspaper. They include a number of errors.

For example, donor funds, which cannot be devolved, were included in the county share. A “miscellaneous” 10 per cent was included on top, which cannot be justified. These errors constitute at least Ksh30 billion of what the governors found.

Of the remainder, the funding for roads and water may be justifiable, but requires state corporation reform. In the case of water, it is not clear that regional institutions, like Water Service Boards, should simply be dissolved and their budgets transferred.

Regional entities that promote multi-county approaches to dealing with shared resources are vital to the success of devolution and need to be supported somehow. Whether we should do parastatal reform by referendum is an open question.

More broadly, the push for increased county funding through a referendum suggests that we are not properly using the policy tools bequeathed to us by the new constitution. For the katiba envisions reasoned, evidence-based policy debate. And there is simply too little of that.

Jason Lakin is a senior programme officer and research fellow at the International Budget Partnership. E-mail: [email protected]