World Bank Investment Projects Aren’t Designed for Crises

In a recent note, Zack Gehan and I used a database of World Bank projects to examine the association between World Bank project types and environmental review procedures and project preparation times. We found that policy lending projects are significantly faster to design, while projects that undergo the stricter forms of environmental and social screening are slower (these are “Category A” projects where borrowers and the bank need to undertake assessments and agree mitigation measures and “Category B,” which still requires some assessment and mitigation measures). Policy lending takes 107 fewer days to reach the board than investment lending, while category A loans take 189 days longer than non-category A/B loans and Category B projects take 43 days longer.

This blog uses the same dataset to explore how long projects take to complete: to disburse funds and finish the project program. Unsurprisingly, the picture is similar with regard to what takes more time to complete as what takes more time to begin. 

I take all board projects from the database approved 2010-2017 that involve an IDA or IBRD commitment, report a lending instrument, and have a closing date where that date is reported as before the end of 2022. I divide the sample into development policy lending, program for results projects (a small subset of projects that pay on the achievement of agreed development outcomes and do not undergo environmental screening), and all other (including investment projects, adaptable program loans, technical assistance, financial intermediary and sectoral investment loans). This “other” group of loans I divide into category A, B and other or no environmental screening categories. The figure below reports the average length of time reported from project approval to closure for each of these groups.

World Bank projects: Days from approval to close

Reported Category A projects take an average of 2,689 days (7.4 years) to close. Category B projects take 287 days less to close, while other projects that aren’t policy or program for results loans take 195 fewer days again. Program for results projects close in about 5.8 years, while policy loans close out in less than 15 months, or about one-sixth the time of a category A project. In fact, because my sample of projects runs up to those approved in 2017 (only six years ago), this is probably an underestimate of the difference between policy lending project lengths and other projects, as it will only include the more rapidly closing investment and program for result projects in recent years.   

The difference in bureaucratic overhead suggested by those numbers is considerable. Seven or eight years of World Bank task team missions, often multiple times a year, with procurement reviews and financial and technical oversight, carries a considerable cost to both borrowers and the World Bank. That will be a major factor behind client complaints about the hassle of borrowing from the bank, and a deterrent to borrow for zero carbon investments. But it also speaks to the question of what’s a good tool to respond to climate crises. A recent TED Talk suggested we have seven years to save the planet. If that were true, it would be about equal to one World Bank investment project cycle. Even if it is an irresponsible exaggeration, maybe other tools are more appropriate?


CGD blog posts reflect the views of the authors, drawing on prior research and experience in their areas of expertise. CGD is a nonpartisan, independent organization and does not take institutional positions.

Image credit for social media/web: Adobe Stock