More Tales of Two Tails

The following post is by Dennis Whittle, co-founder of GlobalGiving. Dennis blogs at Pulling for the Underdog.

An eloquent 3 year-old would have been better asking "What the dickens are you talking about?  Who is defining success?  Who says failure is bad, anyway?" - Joe

Earlier I blogged about aid cheerleaders and critics. Each camp argues about the mean outcome of aid rather than the distribution of impact among projects. Both camps agree that some projects have positive results and others negative. So why not try to figure out which projects work and focus our resources on them?

I got some great and insightful comments and a few nice aid distribution graphs from readers.  Here are some key themes:

1. The mean *does* matter if the distribution is random. In other words, if we can't predict in advance what types of projects will succeed, we should only spend more resources if the mean outcome is positive.
2. Many people believe that on average the biggest positive returns come from investment in health projects.
3. We should also look at the distribution of impact even within successful projects, because even projects that are successful on average can have negative impacts on poorer or more vulnerable people.
4. Given the difficulty in predicting ex-ante what will work, a lot of experimentation is necessary.  But do we believe that existing evaluation systems provide the feedback loops necessary to shift aid resources toward successful initiatives?
5. "Joe," the commenter above, argues that in any case traditional evaluators (aid experts) are not in the best position to decide what works and what doesn't.

Petr Jansky sent a paper he is working on with colleagues at Oxford about cocoa farmers in Ghana.  The local trade association was upset that they could not get pervasive adoption of a new package of fertilizer and other inputs designed to increase yields.  According to their models, the benefits to farmers should be very high.  The study found that - on average - that was true, but that the package of inputs has negative returns to farmers with certain types of soil or other constraints.  Farmers with zero or negative returns were simply opting out.

At first glance, these findings seem obvious and trivial.  But they are profound, in at least two ways.  First, retention rates are an implicit and easily observable proxy for net returns to farmers.  We don't need expensive outside evaluations to tell us whether the overall project is working or not.  And second, permitting farmers to decide acknowledges differential impacts on different people even within a single project.

What other ways could we design aid projects to allow the beneficiaries themselves to evaluate the impact and opt in or out depending on the impact for them personally?  And how would it change the life of aid workers if their projects were evaluated not by outside experts and formal analyses but by beneficiaries themselves speaking through the proxy of adoption?

A Tale of Two Tails

The following post by Dennis Whittle is cross-posted from his blog Pulling for the Underdog. Dennis is co-founder of GlobalGiving. This past weekend I took my three and a half year old son to Princeton to a colloquium on foreign aid.  Speaking were senior people from both the aid industry (including Raj Shah, Administrator of USAID) and academia (including Angus Deaton, one of the best professors I have ever had).  There was a spirited discussion of whether aid "works."

Afterwards, my son asked "Dad, doesn't the distribution matter as much as the mean?"

"Yes," I replied.  "It does. In fact, the distribution may be more important than the mean. Professor Deaton would be proud of you for pointing that out."

Let's assume that aid impact can be measured on a scale of -4 (horrendously harmful) to +4 (miraculously wonderful).  Figure 1 shows the implicit belief of most aid cheerleaders.  The average impact is +1, with most of the impact greater than zero.  The cheerleaders say "Yes, there is a small part of aid in the shaded area under the curve that has negative effects, but those examples get too much publicity.  We really need to do a better job of publicizing and explaining the large area under the curve that represents positive impact."

By contrast, the critics feel that Figure 2 is more accurate.  They believe that the average impact is -1, with the vast majority of projects (the non-shaded area under the curve) having an impact less than zero.  The impact of some projects even approaches the nightmare of -4.  Most critics will concede that there are some projects (the shaded area) that have a positive impact, and if pressed they will offer some personal examples.  (Professor Deaton offered certain health projects, for example.)

The important question is not whether aid as a whole "works," which has been the subject of a large number of papers in recent years.  The real question is what the distribution of impact is.

So, readers, here is your homework assignment: 1) Do you consider yourself a cheerleader or critic? 2) Please download a blank version of the graph here, fill in your own guess at the distribution and email it to me.  3) Describe what types of projects you feel fall into the category of effective.  I will post a follow up with selected responses and insights.

The Answer

....that no single key, no formula can, in principle, solve the problems of individuals or societies; that general solutions are not solutions, universal ends are never real ends.... ...that liberty--of actual individuals, in specific times and places--is an absolute value; that a minimum area of free action is a moral necessity for all men, not to be suppressed in the name of abstractions or general principles so freely bandied about by the great thinkers of this or any age, such as ... humanity, or progress...names invoked to justify acts of detestable cruelty and despotism, magic formulas designed to stifle the voices of human feeling and consience.

This is Isaiah Berlin describing the views of the great Russian thinker Alexander Herzen (1812-1870), although I think he was also describing the views of the great Russian thinker Isaiah Berlin. It's from one of my all-lifetime-favorite books, Russian Thinkers. (HT Dennis Whittle for reminding me of the Herzen chapter.)

Berlin presents the lite version of Herzen in a direct quote from his autobiography:

'Human life is a great social duty,' [said Louis Blanc], 'man must constantly sacrifice himself for society.'

'Why?' I asked suddenly.

'How do you mean "Why?"-but surely the whole purpose and mission of man is the well-being of society?' [said Blanc]

'But it will never be attained if everybody makes sacrifices and nobody enjoys himself.'

Another version of this I heard a long time ago, not sure where:

The purpose of life is to live for others.

Then what should the others live for?

Aid is not just complicated; it’s complex

One of the points that we try to make on this blog is that aid, planned from an ultra high level and driven to alleviate just the symptoms of poverty, doesn't realistically address the complex problems of international development. We understand that our own economies are complex and require complex allocation mechanisms (i.e. markets; see also "failure of the U.S.S.R.") but this thinking doesn't hold when it comes to helping the poor. So consequently we come up with overly simple solutions to far more difficult puzzles. Ben Ramalingam, author of the blog (and forthcoming book) Aid on the Edge of Chaos, explains this another way in an interview with Dennis Whittle:

[I]nternational aid has been built on a very particular way of looking at the world, and this continues to dog its efforts. As a senior USAID colleague put it, because of our urgency to end poverty, we act as if development is a construction, a matter of planning and engineering, rather the complex and often opaque set of interactions that we know it to be.

...The whole system disguises rather than navigates complexity, and it does so at various levels – in developing countries and within the aid system. This maintains a series of collective illusions and overly simplistic assumptions about the nature of systems, about the nature of change, and about the nature of human actors.

So the end result of all of this is that poverty, vulnerability, disease are all treated as if are simple puzzles. Aid, and aid agencies are then presented as the missing pieces to complete the puzzle. This not only gives aid a greater importance than perhaps it is due, but it also misrepresents the nature of the problems we face, and the also presents aid flow as very simple.

Instead of engaging with complexity, it is dismissed, or relegated to an afterthought, and the tools and techniques we employ make it easy for us to do this. We treat complex things as if they were merely complicated.

What is the difference? As Ben goes on to explain, complicated systems can be modeled mathematically, but complex systems cannot.

[For complex systems,] there is no mathematical model which can say, if X is the situation then do Y. Sustainability, healthy communities, raising families have all been given as examples of such complex systems and processes. Peacebuilding would be another, women’s empowerment, natural resource management, capacity building initiatives, innovation systems, the list goes on and on. Complexity science pulls back the curtain on these processes and it can force you to think about the world you live in in a different way.

Thanks to Dennis for this pointer to Ben's work. (See also Nancy Birdsall's blog post about Dennis on the occasion of his retirement from GlobalGiving.)

The Androids are coming, is aid ready?

This post is the second in a series by Dennis Whittle. Dennis is the CEO of GlobalGiving, an international marketplace for philanthropy. In my last post, I argued that the "operating system" used by the current international aid agencies is stuck using IBM punch cards while the rest of the world has moved on to cell phones, laptops, and iPhones.

In the old system, you had to type programs into a stack of hundreds of punch cards, walk them down to the computer center, hand them to an operator, and wait in line for them to be processed. The lower you were in seniority, the longer you had to wait.

Contrast that with today, when hundreds of millions of people have their own computers, and several billion people have access to cell phones. iPhone and Android platforms now provide a way for hundreds of thousands of individuals to create and distribute new software. That software may or may not succeed in the marketplace, but it can be downloaded and used, and the most popular new apps get flagged to other users.

What would an analogous distributed operating system look like for the aid business?  The design of any such operating system has to address five questions:

1. Who decides which problems aid should address?
2. Who comes up with the solutions?
3. Who gives the funding?
4. Who competes to implement the solutions?
5. Who gives feedback on how well the solutions work?

The existing system is closed. It assumes these questions will be answered by experts within one of the "mainframe" organizations like the World Bank, the UNDP, USAID, or one of the big NGOs.

Unlike 60 years ago, the expertise, resources and technology now exist to make a new decentralized and distributed aid system possible. Many more people—and not just experts—have relevant developing country experience (for example, the Peace Corps alone has over 250,000 alumni). Regular Americans give over $25 billion each year to charitable causes abroad—about the same amount as the US official aid budget. And PCs, internet, cell phones and related technologies now make it possible to connect all of these people and resources directly to the people who need help.

Marketplaces that directly connect funders to projects include GlobalGiving, DonorsChoose, MissionFish, GiveIndia, HelpArgentina, Conexion Columbia, Betterplace (Germany), GreaterGood South Africa, Net4Kids (Netherlands).  Markets need quality information, and groups such as Guidestar, New Philanthropy Capital, GiveWell, Great NonProfits, Philanthropedia, Keystone Accountability, Charity Navigator, and the BBB Wise Giving Alliance have begun to operate mechanisms that collect and make available data from a wide variety of sources. Technology startups such as Frontline SMS and Ushahidi are making it possible for beneficiaries to have direct input into what they need ex-ante (schools? health clinics? microcredit?) and how well projects are being implemented once underway.  Other initiatives (for example, Jumo) are in the planning stages.

Not all of these new initiatives will succeed.  But within a few years, it will be possible for any person or group in the world to help answer any of the five questions above that form the core of the aid operating system.  If the existing agencies that run closed systems don't adapt to this new set of conditions, they will become as irrelevant as old IBM mainframes.

--

Related post: Is aid stuck using IBM punch cards?

Is aid stuck using IBM punch cards?

This post was written by Dennis Whittle. Dennis is the CEO of GlobalGiving, an international marketplace for philanthropy.

When I went to college in the late 1970s we used punch cards like the one pictured here to put information into the (mainframe) computer. Over the next thirty years, competition in computer technology led to rapid innovation. Over those same thirty years, the aid industry followed a decidedly different pattern of development.

In the early 1980s, the DOS operating system was a huge leap forward since it allowed people to use PCs on their own desktops. Ten years later, Microsoft released the Windows operating systems, which provided a graphical interface - another big advance.

All along, Apple had been selling its own graphical interface while struggling to avoid bankruptcy. But once Apple released its beautiful Mac OSX operating system, it rapidly gained market share, forcing Microsoft to develop (after several failures) a powerful new release of its Windows system. At the same time, Apple released the iPhone, a real game changer, especially when combined with Apple's App Store, which provided a distribution platform for hundreds of thousands of small, independent developers to release software applications. As a result of its competitive success, Apple's market value overtook Microsoft’s in May 2010. Some people think Apple’s new iPad could revolutionize the industry again.

The competition between Microsoft and Apple was based on what the customer wanted and liked. Both companies experimented with many products, some of which failed and some of which succeeded. They got rid (mostly) of the failed products and continually and incrementally enhanced the successful ones. Every once in a while (as with the graphical interface and the iPhone), they made giant leaps.

Contrast this with the aid industry, which, ironically, is managed in a centrally planned way even as it promotes market-based solutions to developing countries. The big aid agencies get very little feedback from their ultimate beneficiaries - the people they are trying to help. There has even been a recent trend around "partnerships and collaboration," whereby agencies agree to divide up their business and not compete.  For example, the World Bank might agree to concentrate on telecoms in a set of countries, while the ADB handles health.  This further insulates them from competition. The result has been little innovation over the past decades.

The operating system example is a good metaphor in itself. What operating system is the aid industry using right now? Unfortunately, over the past sixty years, it hasn’t progressed much beyond punch cards.  While there have been improvements to the various processes, the aid business is still based largely on a "mainframe" model, with a small number of mainframes such as the World Bank, ADB, UN, and bilaterals such as USAID, MCC, DfiD, and AusAID dominating the market.  What will it take to move toward a distributed "desktop" model of aid, and to stimulate the parallel creation of the DOS, Windows, Mac OSX, and even iPhone operating systems?

What would a competitive distributed operating system for aid look like? Put your ideas in the comments, and I’ll write a follow-up post.