Failure by design? Why public works programmes don't always deliver what they promise.

We weren’t looking for trouble. 

In fact, we went to great lengths to avoid finding it.  Our slightly geeky work was to develop ways of assessing the impacts over time of development interventions such as public works programmes. Public works programmes (also called things like ‘cash for work’, ‘food for assets’ and ‘employment generation schemes’) are when people in need of food or money are given a small wage in exchange for working. 

To test our tools, we had to have some impacts to study, so we were only looking for successful projects. We studied public works projects where a few years earlier people had been made to work combatting soil erosion in Ethiopia and building dams for water ponds in Kenya. The projects, of a kind commonly used, had been successful, we were assured. That didn’t just mean that people had worked and been paid – that brings immediate benefits. We were studying how to assess the longer-term benefits: so one where the work done on soil conservation or creating water ponds had created something useful. 

What we found should be shocking. The work that people had done – or, in the jargon, the ‘assets’ that they had constructed – brought virtually no benefits for anyone’s livelihoods at all. 

Why this matters

The wages that people earn are often regarded as the most important part of public works programmes, which may be targeted at people in need of money to meet immediate needs. But there are two consequences of using public works programmes instead of simply giving people a welfare payment: people had to sacrifice their time and effort to work; and the programme typically costs twice as much. It is expensive to find work for people to do, buy tools and materials, supervise construction, etc. This adds up to billions of dollars a year being spent not on welfare payments for people who need the money, but simply on making people work to get their benefits. 

This cost is justified by the claim that the assets that people build bring longer-term benefits, such as for their livelihoods or resilience. The assumptions are that soil conservation will bring people higher agricultural yields and water ponds will allow them to water vegetable gardens. So, as a result of the public works, they won’t need cash hand-outs in the future.

Why do we say the failure was ‘by design’?

These assumptions proved completely false in the projects that we studied. That can happen: it isn’t shocking or even particularly surprising to find projects that haven’t been very successful, because it is much harder than we like to admit to design interventions that work well. But the closer we looked, the clearer it became that the assets hadn’t been useful because they were never taken seriously. Only one thing had been considered in every step of the project cycle, from deciding what to build and where, to evaluating the project’s success: making the wage payments (whether in food or cash) to alleviate food insecurity. No priority was given to ensuring that the work they did, i.e. the assets they built, would bring any longer-term benefits. 

Any number of examples, some detailed in our report, would illustrate this. Trees were planted on hillsides to reduce erosion. But because people needed to be kept busy to justify making wage payments, thousands of trees had to be planted. To achieve this target, they had to be planted so close together that there never chance they would survive or grow. In other words, the trees were never really the point: making people plant them was all that had mattered.

Dams for water ponds were constructed in sites where the geology and topography made it inevitable that they could not hold water into the dry season. They were sited there because that was where the people lived who needed to be paid. 

Most damning of all is that we were directed to these projects as successes because no one involved in the projects realised that the assets had failed to deliver. Nobody ever checked to see whether people were benefiting from a new water source, or whether farming had been improved by reducing soil erosion. 

That didn’t stop plenty of success stories being told. For example, we were frequently regaled with entirely false stories of honey being exported from Ethiopian hillsides as a result of the public works. If anyone had visited the hillsides to check, it wouldn’t have taken long to find out the truth.

What lessons can we draw?

We obviously cannot use a few case studies to claim that assets created by public works programming are never useful. But we do know that with almost no exceptions, no one has ever bothered to find out. This means that billions of dollars are being spent based entirely on assumptions that no one has been interested in checking. And we did find something inherent in the way in which many such programmes are designed and managed that makes it less likely that the assets will be useful.

We don’t need to find out how many billions of dollars have been wasted making people construct assets that were never likely to be useful. It is more important to worry instead about the future. We can make sure that in designing this kind of programming, as much attention is given in the design, implementation, monitoring and evaluation to what people will create as is give to how wages are paid.  We can still go back to learn from the previous investments to find out in which circumstances the assumptions justifying spending the money have a chance of being true.

We know that plenty of projects will not be as successful as we would like.  But there is no excuse for not making every effort to maximising the chances that the next billion dollars that we spend on PWP assets are worth it.

Read the report here: 'Do public works programmes create valuable assets for livelihoods and resilience? A retrospective study of the impacts of assets for natural resource management in Ethiopia and Kenya'.

Join our webinar at socialprotection.org on Thursday 22 May at 14:00 UK and West Africa time (GMT+1) / 15:00 Central Africa time (GMT+2) / 16:00 East Africa time (GMT+3) to find out more and to help shape a better way forward.