Thank you to @ianbirrell for taking the trouble to reply to my thread, disagreeing with the conclusion he drew from DFID's engagement with the Millennium Villages Project in Northern Ghana.

1/13

We agree that the project turned out to be poor value for money. @ianbirrell cites the evidence for this (in his tweets 5-15) from the @ItadLtd evaluation from the project.

So far so good.

2/13

Where we diverge is on the conclusion to draw from this experience.

@ianbirrell said in his article that DFID spent £11m "to test if aid makes a difference". That isn't true.

DFID's goal was to support these villages and test whether the MVP approach works.

3/13

Many people - including me - were sceptical at the time that the MVP approach would work. That uncertainty was why DFID refused to back MVPs until Jeff Sachs agreed to include a rigorous evaluation.

4/13

I rarely disagree with @m_clem, but I don't agree that it was a "misstep" for DFID to support an MVP project in Ghana, provided that there was an independent and transparent evaluation.

It was important to find out if the proponents of this approach were right.

5/13

But it is ludicrous for @ianbirrell to pretend that the point of DFID's support of MVP was "to prove aid works"

Imagine a drug trial to test the efficacy of a particular medicine.

You would not call this an attempt to "prove that medicine works."

6/13

A successful drug trial would show that *some* medicines work. An unsuccessful drug trial would show that *some* medicines do not work.

So it is with development programmes. Some work well, some work a bit, some not at all.

7/13

So @ianbirrell has set up a straw man. He wants you to believe that this was organised as a landmark test by DFID of whether aid works, which then failed. It was no such thing.

If DFID had wanted evidence that aid works generally, they would not have bet on MVP!

8/13

There are plenty of great examples of aid making a huge positive difference, demonstrated by proper evidence, achieving enormous value for money.

Supporters of aid, like me, should not claim that these show that all aid works. But they do show that some aid works.

9/13

Conversely, opponents of aid should not seize on projects that do not work and use them to claim that no aid works.

Instead they should congratulate the people running the project for having the integrity to conduct open and transparent evaluations, and for learning.

10/13

The reason I am taking up @ianbirrell's valuable time with this discussion is that I think it is corrosive to good public policy for journalists to misinterpret the evidence this way.

11/13

We want open, transparent, rigorous evaluation of public policies. Some policies will turn out to be ineffective, or poor value for money. Finding this out is how we improve and advance. (And how we stop doing things that don't work.)

12/13

If journalists opportunistically seize on negative findings about particular programmes as evidence that all policies in that field are a failure, or use them to criticize the public body that ran the test for wasting public money on failed projects, we will never learn.

13/13