Wednesday, April 19, 2017

The problem with targeting outcomes in development

Outcomes-based payments has been a trend in international development for more than a decade-and-half, though with limited results to show. In fact, except for the deforestation programs, there may be no example of a pure outcomes project. Even the totemic examples skirt around outcomes and end up focusing, at best, on poorly specified outputs. In short outcomes-based financing itself has shown little evidence of outcomes. 

The fundamental premise with outcomes based approach to addressing development problems is the assumption that a financial incentive, adequate and appropriately structured, can force systems to get their acts together and achieve outcomes. The financial structure, it is believed, can help vault past inputs, processes, and outputs, to realize outcomes. 

Here is a listing of three important failings with this approach (assuming the focus being on outcomes and not outputs)

1. The assumption that a financial incentive is sufficient to overcome antecedent deficiencies and achieve outcomes betrays a very high degree of ignorance. In simple terms, the assumption is that financial incentive can ensure that processes are put in place, processes get adhered to, complementary inputs can be and are brought to bear, and execution is reasonably assured. And all this happens in scale. Take a break!

While it may be logical to target outcomes, in practice, the financial incentives alignment alone may not be sufficient to help leapfrog the process compliances that may be necessary to help achieve the objective. In most environments (places and sectors), just process compliance requirements (simple things like teacher attendance or individual student learning tracking, which are critical to achieving the objective) can be daunting enough, even if they are targeted directly. Now achieving them as an incidental benefit from an outcomes-based contract, may be expecting too much. 

2. The other assumption is that governments can afford to pay for outcomes. This, in turn, assumes that the same set of resources can be redeployed more efficiently to achieve the desired outcomes. 

But what if the real cost of achieving outcomes is much greater than the current expenditures? And what if the principles of government budgeting conflict with the requirements of outcomes-based budgeting? 

Answers to these can be found in the way government budgets are made. Logic would have it that allocations be made based on an objective assessment of the cost required to achieve the goals. But in the real world, allocations are made with another consideration as the foremost priority – balance and equity (in the coverage of geographies, sectors, and populations). 

The result is extensive skimping on the allocations, especially on maintenance, consumables, procedural compliances, newer types of activities, discretionary spending, and new investments. Nothing is sacrosanct except the salaries of employees and long-term contractual obligations. Spreading the limited fiscal butter thin, often to the point of being imperceptible (or seriously detracting from the objectives), becomes inevitable. 

Now imagine an outcomes-based project where procedural compliances, maintenance schedules, quality of engagement, rigour of monitoring, and so on are critical to its success. All this inflates the price tag for the actual achievement of outcomes. Governments will find the increment fiscally unsustainable. The option of a phased expansion, apart from raising several practical problems, would also be politically unacceptable. 

3. Finally, in case of private production, the assumption that the supply side is large enough to support outcomes based contracting in scale is most likely to be unrealistic. Forget education or health, try calling tenders for outcomes based water treatment facilities and, in most developing countries, within only a handful of tenders the no-response bids start to bind.

So, the argument would go that the markets take time to mature. We only need to catalyse it and will develop. But, as the far simpler and easier to contract market for infrastructure services has shown, such markets are more likely to remain elusive for a very long time. And, even plain simple outcomes contracts like long-term road concessions are rife with renegotiations and attendant moral hazard which distort contractual obligations. 

So, is it futile to target outcomes in development? It is realistic to target outcomes in at least two areas. One, logistics based activities, especially infrastructure services, are amenable to outcomes targeting. This is largely because, there is little difference between outputs and outcomes. Two, where there is private production and public provisioning of services, and where outcomes can be quantified without any of the standard challenges, it should be possible to hold private providers accountable and demand outcomes. Skill trainings linked to placement is an example.

Yet with even these limited areas, the aforementioned three problems will bind to varying degrees, hampering the achievement of outcomes. 

Similarly, it is unrealistic to target outcomes in cases of public production involving engagement intensive activities. I would define engagement intensive activities as those where outcomes depend intimately on the quality of engagement. Education and learning outcomes is the best example. In all such cases, especially when done in scale in developing countries, state capacity becomes the binding constraint.

No comments: