I have just published a note about enforced local preferences in multilateral development bank (MDB)-financed procurement. The paper assumes the reader believes such preferences might be good for client countries. Of course, many of those calling for such preferences are not as concerned with development outcomes as they are with their distaste at the fact that China wins so many contracts. To the extent that is what is driving the move away from best practices in procurement, less damaging policy changes are available.

All of this is rather silly. The evidence suggests Chinese contractors deliver a quality product at a low price and in part that is because the Chinese government is subsidizing costs—saving MDB clients money. And the amounts involved are small: MDB-financed procurement accounts for about 0.1 percent of public procurement in developing countries. Given that, the very least that concerned shareholders can do is try to minimize the harm done by proposed fixes to an arguable problem.

The current proposals favor imposing a standard universal local preference (as it might be: 30 percent of labor costs on all MDB-financed civil works contracts must be local). The note explains why this is a terrible idea from a development perspective: it is a sludge-ridden one-size-fits-all approach for a policy that requires nuance and country specificity to have any shot at working (and despite all that, it might not even reduce Chinese contractor share). The note pushes for a more flexible approach instead: allowing client governments to use their own national labor and content rules in MDB-financed procurements if MDB country teams think they make sense from a development perspective.

Another approach would see MDB shareholders trying to shift lending to sectors where Chinese contractors are not as dominant in procurements: health, education, and social protection, for example. But absent client demand, any such effort could significantly reduce development impact and the attractiveness of MDB finance. Depending on your view of climate finance, a move away from the core infrastructure sectors involved in mitigation and adaptation might be a small silver lining or another reason to think this is a bad idea.

More attractive would be moving to lending modalities that don’t involve direct bank-financed procurement at all: payment for results or policy lending. That’s because there are good reasons to think that MDBs should be doing more of this lending regardless. World Bank Independent Evaluation Group (IEG) reviews of payment for results projects suggest they outperform the average quality of projects (see chart). More broadly, given the World Bank’s IBRD in particular appears to be primarily a lender of penultimate resort (before the IMF) and development banks are a fairly marginal source of investment finance, the project model as a whole looks increasingly antiquated, and policy lending simply makes more sense.

Percentage of World Bank projects by IEG evaluation, payment for results (all years) versus all projects (2024)

Image

Data: IEG

To repeat, Chinese firms winning competitive bids on projects financed by MDBs is a weak justification for changing MDB lending models, sectoral focus, or procurement rules. The best response would be for rich country donors to stop obsessing over an insignificant problem. But if they can’t, they should abide by the principle “first do no harm.” And simply reducing the proportion of MDB finance covered by burdensome (if broadly effective) MDB procurement rules might be one way to achieve that.