Lant Pritchett on reducing learning poverty: targeted interventions vs systemic change

“… when the whole system is producing weak results for nearly every child, then “inclusion” is a false premise. In this situation, it is necessary to fix the whole system and increase performance across the board in order to reduce the number of children stranded in low performance.

Lant Pritchett – Tackling Education poverty with system-wide improvements

Lant Pritchett is very good at what he does. If you’re interested in poverty reduction – whether poverty in general or learning poverty in particular – I highly recommend this article.

Here’s another highlight (it’s really worth clicking straight to the original to get everything, including the graphs).

The Two Approaches to Poverty Reduction

Pretty much every definition of “poverty” posits some element of human wellbeing (income, access to health care, water, nutrition, education) and a level below which there is “unacceptable deprivation” (which is ultimately a social decision). Given that poverty line, there are analytically two ways of reducing the number of people in poverty (in any dimension):

– Shift the whole distribution up, more or less equi-proportionately, and that shift pushes people over the line, which I call “systemic” change.

– Focus on bringing up those below the line with efforts or activities that are specifically targeted, usually “programmatic.”

Around the World Bank’s World Development Report 1990 that launched the measures of “dollar a day” poverty, there was a big debate about whether the strategy for reducing poverty was a “two-legged” stool (labor-intensive economic growth plus broad based investments in human capital) or a “three-legged” stool of those two plus “social protection” programs targeting the poor, and the compromise was a “two-and-a-half-legged stool”—which then led to laughter all around about how a good metaphor turned goofy: who wants a two-and-a-half-legged stool?

30 years later, we have good cross-section and time series data. The debate is over. Poverty reduction relies on growth in the overall distribution of income/consumption—and pretty much nothing else explains poverty across countries or over time.  … Over time the median of the distribution explains roughly all (an R2 of .98 in levels, .93 in changes, so both very near 1) of the variation across countries in headcount poverty. Improvements in the median consumption are an empirically necessary and sufficient condition for large scale poverty reduction.

This is not to deny that there might be cost-effective programs for addressing poverty, just that variations across countries in the scale and scope and effectiveness at which anti-poverty programs have been deployed to “kink” the bottom of the distribution up, do not appear to explain much (if any) of the differences in poverty across countries or over time. This is not an argument against poverty programs: they can be tactics. But the facts are a compelling obstacle to imagining that targeted poverty programs (those that reduce poverty without altering the median) can be, in and of themselves, an anti-poverty strategy (or even more than a small part of one).

Lant Pritchett – Tackling Education Poverty with system-wide improvements

One Reply to “Lant Pritchett on reducing learning poverty: targeted interventions vs systemic change”

I'd love to hear your thoughts and recommended resources...