Sub-national budget credibility: Kenya’s experience
Jason Lakin, Senior Research Fellow, IBP & John Kinuthia, Lead Research Analyst, IBP-Kenya— Jan 27, 2020
Kenya’s Big Bang Decentralization
In 2013, Kenya began an ambitious experiment in devolution, decentralizing considerable power to 47 newly created counties which would take over responsibility for key services like health care. As is often the case, this devolution was premised on the notion that it would bring services closer to the people, and lead to improved access and a fairer distribution of services. Accordingly, counties were given the authority to formulate their own budgets, independent of national government.
While it is too early to assess the success of these reforms, for the theory of devolution to hold, counties must clearly access and spend funds on services. This is not a sufficient condition for improvements in access or quality, but it is a necessary one.
As part of the International Budget Partnership’s (IBP) Addressing Budget Credibility (ABC) initiative, we undertook the first systematic assessment of sub-national budget credibility in Kenya, looking at the degree to which counties collect revenue and spend as budgeted. Our analysis is based on data from all 47 counties over a four-year period (2014-2018), drawn primarily from reports produced by Kenya’s Controller of Budget. We complemented this data with county budget implementation reports and audit information from the Office of the Auditor General.
Rates of recurrent versus development spending were quite different. While overall counties spent 80 percent of their approved budgets during the four-year period, they were able to spend, on average, 96 percent of their approved recurrent budgets, but only 61 percent of their approved development budgets. In 2017/18, development expenditure against budget fell to only 42 percent. On the other hand, recurrent budget execution has continuously improved, growing from 92 percent in 2014 to 99 percent in 2017.
Trends are very similar to those at the national level. In the same four-year period, the national government’s line ministries (MDAs) spent 82 percent of the approved budget. While recurrent expenditure absorption was reasonably high, just as at county level, development absorption was much lower (an average of 67 percent over the period). From this, we can conclude that there is nothing particularly remarkable about low credibility at county level, but also that decentralization has not (yet) led to any improvement in budget execution.
Challenges at the county level are especially notable in sectors with high development spending –roads, water, and agriculture – although health budgets are also consistently underspent, achieving only an average of 86 percent budget execution over the four-year period we investigate. This is driven mainly by low execution of the health sector development budget, which averaged less than 60 percent for the period.
Re-prioritization of spending occurs during budget execution and may undermine service delivery over time. The shares of expenditure on executive and legislative administration (i.e., operational costs) tends to rise relative to their share of the original budget, while the share for agriculture and water tends to fall. In other words, as the budget is implemented, those costs needed to run the executive but not linked to specific services take a larger share of spending than budgeted, while the share of the budget going to some key services like agriculture and water is less than what was originally budgeted.
Counties rely heavily on national transfers and receive them too late in the year to spend down before the books are closed. Less than half of the intergovernmental transfers are received by the middle of the year, and 65 percent or less by the end of the third quarter. This pattern was particularly severe in 2017. Receiving roughly 29 percent of total revenue in the last quarter is likely to undermine counties’ ability to spend.
Counties also exhibit poor budget formulation and management practices. Kenya’s counties over-budget for expenditure and are too optimistic about revenues. Furthermore, counties exhibit inappropriate usage of supplementary budgets. While these budgets should be used to make corrective adjustments that facilitate budget execution, we find that county-level supplementary budgets sometimes worsen budget credibility, rather than improve it.
Hope for improvement?
While the overall data do not show strong signs of improvement in budget credibility over time, roughly one in five counties did improve their budget performance over the period. These improving counties come from across Kenya and suggest that there is wide potential for addressing low credibility over time.
A starting point for measuring budget credibility is the Public Expenditure and Financial Accountability (PEFA) framework. In looking for signs of change in county performance, we started by using the PEFA scoring approach to measuring budget credibility. In this approach, a county that executes anything less than 85 percent of its budget earns the lowest grade of “D”.
Given generally low execution rates in Kenya, this scoring approach obscures the wide variation among poor performers and we therefore introduced a complementary approach. Our alternative ranks counties by quartiles across the full range of budget execution rates. Using this method, we identified nine counties that improved their performance by two quartiles over the period, more than double what we found looking at counties that improved by two PEFA grades (from say a D to a B). In other low credibility environments, such as Nigeria or Philippines, it is important to complement the PEFA approach with other measures.
Table 1: Kenyan counties that showed improved budget performance from 2014-2018 under differing scoring approaches: PEFA vs IBP
|PEFA approach||IBP approach|
Toward greater transparency and credibility
Counties are not transparent about budget implementation. In addition to our overall review of all 47 counties, we looked at eight in more detail. We tried to collect data on credibility from county budget documents and to seek out explanations for low credibility directly from county officers. This was challenging. Only one of the eight counties (Baringo) regularly produces a quarterly budget implementation report, and just over half publish the County Budget Review and Outlook Paper, which reviews annual performance. Both these documents are legally mandated and should provide insight into the reasons for poor credibility.
We found that counties had trouble explaining their performance, although we did learn more about the causes of poor credibility: challenges with contractor capacity, at least partly related to the requirement to fill demographic quotas (for women, youth, local contractors, etc.); poorly designed public participation processes that yield infeasible projects due to lack of technical guidance to inform public views; and lack of or delay in approving policies to guide the use of special public funds (such as bursary or business development funds) where such funds cannot be spent without local legislation.
The low rate of overall budget execution is almost certainly driven in part by delayed release of funds to counties from their share of intergovernmental transfers. Nevertheless, this cannot be the whole story. In general, there is also a fairly high degree of volatility within counties over time, meaning that fixed characteristics of counties, such as their size or level of development, cannot explain the variation.
Are there factors which might vary from year to year and which could have more explanatory value? One possibility is that there are variations in legislative amendments over time, and that sectors where legislators amend the budget to add more projects tend to have lower credibility. This would be consistent with research from other contexts, such as Nigeria or Philippines.
A related hypothesis is that poorly constructed participatory processes could be the culprit. Where the public is not properly guided with technical support, they may insert unrealistic projects into the budget that then cannot be implemented, and these proposals might also vary from year to year.
Our data also hint at the fact that executives and legislatures tend to claw back some budgetary funds for their administrative budgets, or at least budgets controlled by them directly, and understanding what exactly happens to these funds and how and why they are moved is also an area for further investigation