When Rules Are Wrong: Time to Rethink How MCC Identifies Partner Countries

The Millennium Challenge Corporation (MCC) has officially kicked off its FY2017 “selection cycle.” The big news won’t happen until December when MCC’s board will pick the countries that will be eligible for the agency’s flagship compact and threshold programs, but last week’s release of the “Candidate Country Report” marks the first official action. Normally this is a pretty pro forma step that’s hardly blogworthy, but this year’s report showcases the inadequacy of the current rules that determine MCC’s candidate pool and the unnecessary instability they create—this year three countries that formerly “graduated” are suddenly back in the pool. Hopefully this will serve as added impetus for MCC and members of Congress to make some small but fundamental adjustments to how MCC defines candidate countries, like the ones Nancy Birdsall, Anna Diofasi, and I recently explored.

A simple proposition with two big problems

MCC was established to provide large-scale grant funding to poor, relatively well-governed countries. The first step in picking these partners is to identify which countries are “poor” enough to be considered. MCC’s authorizing legislation addresses this question by defining candidates as the set of countries categorized by the World Bank as low income (LIC) or lower middle income (LMIC) based on GNI per capita in a given year.

There are two main problems with this definition. 1) Graduation is sudden and unpredictable, creating unnecessary hardships for both MCC and the graduating country and 2) GNI per capita alone is an inadequate measure of development need.

The trouble with an inflexible graduation rule

Under existing rules, once a country crosses the threshold from LMIC to upper middle income (UMIC), it can no longer be a candidate for MCC assistance. But in reality, the World Bank’s income classifications are based on arbitrary lines. There is no reason to believe that a country’s move to UMIC status is accompanied by real change in the country’s poverty profile or need for development support. Yet from one year to the next a country can go from being a perfectly reasonable country for MCC to consider financing to too rich to even contemplate. Unpredictability with respect to the timing of a partner country’s income category transitions also unnecessarily complicates MCC’s medium-term budgeting and creates needless ambiguity in the bilateral relationship.

It gets even weirder when countries start to fluctuate around the cutoff. Historically, this has been rare; most countries that transition to UMIC never return to LMIC. But this year, three countries that formerly graduated are now back in the candidate pool.

Three countries that formerly graduated are now back in the MCC candidate pool

 
Note on source: MCC relies on World Bank estimates of GNI to create its candidate pool. The historical data presented here reflect the figures MCC used in the given fiscal year; they are not the revised estimates currently available from the World Bank.

 
Mongolia is particularly noteworthy because of all the hoopla surrounding its graduation to UMIC last year, just one year after MCC selected it to start developing a second compact. Because Mongolia had suddenly become “too rich” to be a candidate country, MCC could not make them eligible for FY2016 funds.  The agency nonetheless came forward with a statement saying it would continue to work with Mongolia.

Some of MCC’s overseers pushed back against the move, arguing that a UMIC was an inappropriate MCC partner and suggesting the agency was trying to skirt around its own rules. Others countered that MCC had made the right choice by recognizing that despite an increase in GNI, Mongolia’s need had changed little. The fact that Mongolia is once again an LMIC underscores the arbitrariness of the income categorization threshold and calls into question how a country, in the span of just two years (in the absence of major governance shifts), could be considered alternately appropriate, then inappropriate, then appropriate again as a partner for MCC. The fact is, Mongolia continues to be a relatively poor country, despite experiencing extremely fast (though now waning), mining-fueled growth. Its boost in GNI did not, and could not be expected to, instantly translate into the institutional transformation that will help alleviate poverty among its population.

I’m sympathetic to the concern that MCC was overstepping its bounds by operating outside its rules. It wasn’t in this case because it could use funds from the year Mongolia was eligible, but I appreciate the importance MCC’s stakeholders place on the agency’s use of rule-based decision making in its country selection process. Indeed, these rules are valuable as a buffer against political interference or attempts by the agency to command a larger budget. But when the rules don’t make sense, it’s time for them to change. At a minimum, there should be a more gradual, or “lagged” graduation rule that enables countries to remain in the candidate pool for a designated period of time after transitioning to UMIC.

The shortcomings of GNI

Adjusting the graduation policy, while important, won’t address the fact that GNI per capita is, by itself, an inadequate measure for assessing development need. As CGD President Nancy Birdsall told members of the Senate Foreign Relations Committee late last year, GNI per capita doesn’t provide an clear picture of broad-based well-being, especially in places where there is significant inequality or where growth is driven largely by enclave (usually extractive) sectors. Now that poverty is increasingly concentrated in middle income countries, being able to clearly understand a country’s poverty profile in addition to its national income is all the more important. It may make sense for MCC to be able to work in certain well-governed UMICs with relatively greater development need.   

GNI per capita will almost certainly remain an important component of how MCC identifies candidate countries. After all, a measure of average well-being is informative in terms of the potential availability of the country’s own resources for development.  And let’s not overlook the important fact that GNI per capita is still one of very few indicators that have annually updated data for all countries. However, there are a number of other indicators of development need that are worth exploring as complementary to GNI per capita. As Nancy Birdsall, Anna Diofasi, and I explain, one particularly promising measure is median household income or consumption, which better covers typical material well-being. After all, MCC’s mission is poverty reduction through economic growth. GNI per capita can tell you about growth, but you need something else to tell you about poverty.