Tag Archives: DFID

Amoral hazards

In the last post we saw that humanitarian aid agencies are not gathering data on the outcomes of their work (such as years of life saved, average weight increases etc).  Not doing so has some dangerous ramifications, but before looking at these I want to characterize the interplay of motivations that leads to this situation.

Lord Ashdown is clear: donors are as allergic to measuring outcomes as the aid agencies themselves.  In fact, once a contract is determined there is no reason to distinguish between the two tiers; agencies are putting DFID priorities into action.  The distinction between the two tiers should be important when DFID or AusAID or USAID are choosing which agency to fund during a particular crisis.  How do they choose?  They use the process evaluations that the agencies are using to monitor themselves.  For example, who can shift the most food? Who can buy the most tents?

Donors like these types of ‘outcome’ because they are easy to interpret and sell to the public.  They don’t offer budgets for proper monitoring and evaluation and so the agency doesn’t have a choice but to continue with the status quo.  Maybe that’s all, but these agencies have clout (Oxfam, Save the Children) and the British public rarely scrutinize DFID reports to see how that money is being spent.  And yet the agencies don’t say anything, and they don’t put sufficient pressure on DFID to think differently. It is a closed loop: agency and donor benefit when costs are not compared to returns.  The risk of a poor strategy, like wasted food or worse, is borne by the recipients of aid.

This is a moral hazard.  A moral hazard is an economics term for when an agent (person or company etc) behaves differently because they do not bear the consequence of risk taking.

The situation is a bit like the banking crisis.  Aid agencies are ‘too big to fail’, inasmuch as we need them for when the next crisis comes along.  In the banking crisis, it was tax-payers’ money that went via the government to the banks.  With aid agencies, the aid recipients do not themselves contribute money (they’re normally too poor after all) but money is spent on their behalf by governments so that that aid can be given to them (the ‘process’ output often measured).  In both cases there is very little thought to whether or not the bearer of the ultimate risk (tax-payer in the UK, aid recipient in Somalia) is getting a good deal, that couldn’t be better supplied by someone else.  The life-blood of these agencies is recipients in their camps, they need them, but all of the risk is pushed onto the  same recipients.  For example, Oxfam won’t be outcompeted by Save the Children even if Save the Children could actually help 3 times as many people per pound spent in a crisis; we wouldn’t know.  All we know is what they are spending: how many tents? How much food? How many aid workers?  It is the people in Oxfam’s camps (who else!?) will bear the brunt of this inefficiency, alone.

So, no-one in the system need care about the quality/efficiency of services because the risk is passed on to the recipients of aid.  And, like the banks, they probably are too big to fail, indicating the need for tight, independent regulation.  Otherwise there is a run-away feedback of motivation: DFID spend more and the agencies spend more.  Perhaps the ‘good-will’ of the agency members is expected to stop this but, as Lord Ashdown says, the not insubstantial costs of doing so are not covered by DFID grants.  We cannot blame the aid workers themselves because the system has closed in.  However, if they felt the true risk of bad aid delivery then perhaps they’d be complaining with a louder voice.


Filed under Uncategorized


Outcomes that require contact with the people being helped are rarely measured during or after the delivery of humanitarian aid.

Consider this from the opening paragraph on ‘Impact’ in Lord Ashdown’s Humanitarian Emergency Response Review (HERR):

The emergency response sector does not routinely assess impact. What this means is that for some operations it is almost impossible to say how the assistance helped people, or even whether it helped them.

After noting that measuring anything in a crisis is indeed more difficult than normal, he suggests that much of the resistance to proper monitoring and evaluation is because,

it is convenient not to have to do it; the need for high visibility action, shaped by donors and the media; the competition between agencies and the lack of serious consultation with the people in need; the failures in leadership. Perhaps most important, impact assessments require time and resourcesthatare not trivial, and to date few donors have been willing to support these efforts on the humanitarian side of the system

And finally, just to complete the picture, he says,

…evaluations tend to evaluate activities rather than impact and do not often enough include a participatory element so that disaster survivors can hold agencies to account.

Which really takes us to the crux of the issue.  Where outcomes are measured, they are in terms of output – food, medicine, shelter – and not results.  Imagine that you are running a sales business and you have to decide which of your sales people to give the most resources.  They are allowed to negotiate on prices and make a hard sale.  Carefully, you count up the number of units each of your 20 salespeople have sold; and look at that, there’s quite a range!  So you give more and more of the best products to those whose sales figures are the highest (and even a little bonus).  Then you sit back and wait to get rich…

But you don’t get rich, and in this context the reason is obvious: revenue.  Your salespeople are driving down prices to shift resources beyond the point of breaking even.  But who can blame them, after all, shifting resources is what you measured and used as the basis of reward.  You’re bankrupted.

The banality of this example is most of the point; outcomes should be measured on a combination of revenue and costs (the not so mysterious idea of profit) and not on either alone.  But it goes further.  Businesses do not choose to suffer under the tyranny of the revenue/costs balance; that balance is the definition of the proper use of resources.  It is not something cold and ruthless, it is the simplest rational measurement of outcome that we have.

Perhaps humanitarian agencies think that they are measuring results: food delivered, check.  Shelter built, check.  They’re not.  And we know this because the outcomes measured do not require actual contact with the people the agencies are there to help.  Weight-gain? Health? Well-being? Alive? These are the outcomes we should be interested in, not processes, not our costs.

The next stage is to explore why things are like this, why we’re not doing things properly.  Lord Ashdown has gone some of the way to explaining it but I believe that it is a lot more complicated.

Leave a comment

Filed under Uncategorized