Can aid agencies help systems fix themselves? The implications of complexity for development cooperation

Owen Barder gave a brilliant lectureowen barder 2 on complexity and development to my LSE students earlier this year. Afterwards, I asked him to dig deeper into the ‘so whats’ for aid agencies. The result is this elegant essay (a bit long for a blog, but who cares?). I will try and get some responses to his arguments from similarly large brains.

If economic development is a property of a complex adaptive system, as I’ve argued elsewhere, then what, if anything, can development agencies and NGOs do to accelerate it?

To be clear what we mean when we say that development is a system property, here’s an example from the animal kingdom. You may have seen recently that ants have recently developed “super colonies” – including one that covers 6,000km along the Mediterranean that is said to be the largest co-operative unit in the animal kingdom. It is natural to talk about the “behaviour” of the colony, even though we understand that we are really talking about the individual actions of hundreds of millions of ants. Each ant responds to its external environment, including the behaviour of other ants. Because all the ants are adjusting to each other, this creates the sense that the colony as a whole is changing its behaviour, and we soon begin to ascribe intent and agency to the colony rather than the individual ants of which it consists.

antsLike any complex adaptive system, an ant colony will tend to go through long periods of stability and then sudden periods of rapid change that come about when ants all adjust their behaviour in response to changes in the behaviour of the ants around them. The emergence of a super-colony did not depend on the ants individually becoming fitter and stronger, learning new skills or becoming more entrepreneurial. They didn’t suddenly have access to better nutrients that made them healthier– nor have the ants benefited from universal education, access to microcredit, or new vaccines.  In fact, the ants haven’t changed at all: the colony’s behaviour can change even if the individual ants have not, because it is a self-organising complex system whose behaviour in aggregate is not simply the sum of its parts: it is determined to a large extent by the way those parts interact with each other.

There are plenty of other examples of complex systems in everyday life, such as the weather, our brains, and traffic. These systems have emergent properties which cannot be reduced to or predicted by characteristics of their component parts. Emergent properties such as thunderstorms, consciousness, and traffic jams are not brought about by changes in the characteristics of the systems components (air molecules, synapses, and cars) but by changes in the way these things interact

a very complex system
a very complex system

with each other. Emergent properties are characteristics of systems, not their components.

So what does this mean for development? An economic, political and social system develops when it evolves a complex, self-organised set of interlocking institutions that provide freedom, prosperity and opportunities to its people. Each of these individuals, groups or institutions acts within the constraints and authorising environment provided by the other parts of the system, and as these different parts of the system adapt to each other, so they co-evolve.

For example, an effective civil society depends not only on the resources of Civil Society Organizations themselves, but on the authorising environment and constraints of courts, media, political system, police, army, churches, business and the public.  Effective firms depend on being able to interact with other firms, the maintenance of contestability and competition, and both the support and constraints of government, the courts, civil society, the workforce, unions, consumers and others. And effective government depends on firms, civil society, political institutions, media, law enforcement, courts and so on.  These institutions simultaneously reflect and shape the distribution of power within society.

As these institutions all evolve, and interact with each other in ways that shape each other’s behaviour, so self-organising complexity begins to generate the freedoms, opportunities and prosperity that we call development. Development is a property of the system, not of its components.  All of these institutions have to evolve together: trying to move any one of them individually into a “developed” state is unlikely to be sustained: that component will simply revert to the trajectory that is consistent with the opportunities and constraints provided by the rest of the system.

If you see development as the sum of its parts, then you might think we should work out which constraint binds the most tightly and invest resources to relax it. But if you see development as property of a complex adaptive system, then targeting a specific constraint is unlikely to “make development happen.”

So what does this mean for those of us who want to see development accelerated?

Let us first say what it doesn’t mean.

First, complexity does not condemn us to fatalism: there is plenty we can do to accelerate and shape the evolution of the system.

Second, better analysis is not the solution to complexity: the way to find out what works is through iteration and improvement, not decade-long research projects to understand it better.

Tempting but wrong
Tempting but wrong

Third, there is little value in identifying missing ingredients or binding constraints (eg “growth diagnostics”). Even if the most important gaps could be reliably identified—and we often get even this step wrong—those gaps are the consequences, not the causes, of the condition of the economic and political system.


adaot and survive
adaot and survive

Fourth, we should avoid creating protected incumbents that become obstacles to change, whether in government (through large, unaccountable government-to-government aid flows), the private sector (through subsidised loans or guarantees, or trade preferences), or civil society (external finance can create “astroturf” organisations that are not accountable to their members).  All these types of well-meaning intervention offer, at best, short-term gains at the expense of long-term damage to the dynamic evolution of the economy and polity.

Fifth, the problem is not “lack of capacity” – or rather, the lack of capacity is itself a consequence of the system and will not be rectified by technical cooperation or capacity building programmes. This might be why large-scale evaluations of technical assistance programmes generally declare them to be a failure. (See for example this World Bank review).

If we want to accelerate development we have to focus on improving the system’s capacity to adapt. As in nature, adaptive efficiency depends on two things: variation and selection.  Interventions which improve or accelerate variation and selection will tend to help the system to generate the self-organising complexity we call development.

Examples of policies that rich countries control that can help to bring this about in developing countries include:

a. Trade. Access to overseas markets increases the returns for successful firms and so increases investment and jobs.  It sharpens competition and increases contestability and so drives improvements in performance. It increases access to imported inputs (such as machinery, business services and knowledge) which increase productivity.  And it helps accelerate the circulation of ideas and skills. And unilateral trade concessions like the African Growth and Opportunities Act, enacted by the US Congress in 2000, show how rich countries can unilaterally grant developing countries better trading opportunities, without requiring an immensely costly, reciprocal multilateral negotiation process at the WTO.

taxb. Tax. At the heart of every successful society is a social contract between the state, citizens and business; and tax plays a key role in enforcing that contract. Tax provides the revenues that enables the government to provide essential infrastructure and public goods, and it fosters the accountability to citizens on which good government depends.  Rich countries should do more than provide technical assistance: they should change the mechanisms for international tax cooperation and information sharing to enable developing countries to start to build a sustainable domestic tax base, including by making sure multinational corporations pay their fair share of taxes in the countries in which they’re due.

c. Migration. Apart from improving the wellbeing of migrants themselves, increases in migration opportunities—even the chance to work temporarily in rich countries– for people from developing countries increases the circulation of ideas (for example, countries whose students go to university in democracies tend to improve governance at home) and increases the returns to skills, so encouraging investment in education, and directly benefits the sending communities through remittances and return migration.

d. Transparency.  Greater openness facilitates accountability.  Outsiders can be transparent about the aid they give and the payments they make to governments and public enterprises, for example for oil and other mineral concessions.  They can through their own behaviour encourage the spread of openness as an international norm, for example open contracting, open data and public registries of the owners of companies, thereby expanding the spread of transparency and domestic accountability across the developing world.

e. Media and “connective action”.  Though mobile phones are not a ‘silver bullet’ for development, there are examples of the way that improved communications and media have helped make  governments more accountable and markets more effective.  For example, there is evidence that when voters know more about the policies of candidates in elections, they are less likely to vote on ethnic lines and politicians are more accountable in office. Use of mobile phones can reduce election fraud (for example, in Ghana in 2004) and crowdsourced reporting of bribery might help to reduce petty corruption.

f. Access to ideas. The ability of developing countries to catch up with rich countries has been hampered by obstacles to the spread of new technologies and knowledge – such as medicines, software, machinery and know-how. Industrialised countries have done little to live up to their treaty commitment to facilitate technology transfer, codified in Article 66.2 of the Trade Related Intellectual Property Rights (TRIPs) Agreement.

g. Innovation.  External donors can promote innovation by increasing the returns to success (for example, through Advance Market Commitments or Development Impact Bonds), and by acting more like venture capitalists and less like retail bankers. The key in such interventions is to avoid picking winners, even though there might appear to be short term benefits to doing so. The goal is to build industries that are successful and which are capable of remaining competitive through a process of innovation and adaptation. This is most likely to include a degree of creative destruction, in which new firms enter the industry, replacing less successful incumbents. Donors should stop giving subsidies to particular companies in ways that make these markets less contestable. For example, a country may be less likely to evolve a thriving, successful microfinance industry of its own if many of the most promising opportunities have been taken by externally-subsidised microcredit firms.

Many development interventions try to imitate the consequences of a successful system instead of thinking about how tocomplex-system enable the system to find successful solutions to its own problems. This difference matters because interventions to “plug the gaps” or “relax constraints”, however well-intentioned, may be harmful for the long-run evolution of the system. We provide aid to governments in ways which can make them less rather than more accountable; we provide humanitarian food aid in ways which can destroy markets for local producers; we subsidise civil society organisations and can make them less accountable to their own communities; we subsidise firms to invest abroad in ways which can make them less vulnerable to competition, and so make industries less contestable and innovative; we limit migration in the name of avoiding a brain drain; and through policy conditionality we try to make particular policies better (at least in our view) but usurp the emergence of contested national politics. All these interventions may undermine the long run evolution of a successful, inclusive economic and political system.

A policy to promote development would focus much more on helping the system to evolve, and much less on filling the gaps which are left by systems that have not yet done so.

The quest for sustainability in foreign aid projects over the last four decades has done untold damage. Projects that might have benefits in the long run have been structured and financed as if they could reasonably achieve success or become self-financing within a few years, and are declared failures when they do not. Health services that should have been free have been required to charge user fees in the name of sustainability, so denying access to the poor. It is understandable that aid donors and aid recipients alike would rather imagine that development projects will catalyse permanent change than accept that they will be permanently engaged in an international system of welfare payments. But despite the rhetoric, few programmes have actually achieved the sustainability that they were required to promise.

Sustainability has been elusive in practice because the aid industry thinks too much about gaps and too little about systems. Interventions that accelerate the evolution of a successful economic and social system can be catalytic; interventions that ignore the complexity of the system and only try to fill the gaps that it leaves or imitate its consequences will work only for as long as the intervention continues.

Improvements in the adaptive efficiency of the system are the new sustainability.

Tomorrow Jean Boulton responds on the politics of complexity

Subscribe to our Newsletter

You can unsubscribe at any time by clicking the link in the footer of our emails. For information about our privacy practices, please see our .

We use MailChimp as our marketing platform. By subscribing, you acknowledge that your information will be transferred to MailChimp for processing. Learn more about MailChimp's privacy practices here.


26 Responses to “Can aid agencies help systems fix themselves? The implications of complexity for development cooperation”
  1. Samuel Cumpsty

    Brilliant as ever from Owen. The complex adaptive systems thesis shimmers with the promise of a solution(s). What I’m left wondering is whether the policy implications are not in the end prescribing the same or similar poverty fixes we’ve had for the last 60 years: “it’s finance” (tax), “it’s capital & manufacturing” (trade), “it’s human capital” (migration), “it’s technology” (media, innovation, IP)? If that simplification is not unfair, is the main thrust of the thesis simply that we need to work on these different areas simultaneously, and complementarily (as far as is possible/knowable)? Catalytic, because holistic? If so, how have these 7 areas been picked and how can they avoid the charge of baking cakes the thesis levels at traditional aid? Owen does usually make my head hurt.

  2. Owen Barder


    Thanks. Yes, you are right. There is a risk that this becomes a new “toolkit” and aid agencies start to work in these areas with the normal array of technical assistance projects and policy influencing.

    My plea is for aid agencies to look for changes that that help the system fix itself, rather than applying our our fixes to the system. That is easier to say at a conceptual level than it is to describe in detail.

    Fortunately there are many improvements we can make to our own policies and behaviours which would have the effect of helping developing country systems adapt more rapidly. These include: opening our markets to imports from developing countries, permitting more migrations, changing our intellectual property rules, improving the governance of international institutions, being transparent about the aid we give and the payments we make for oil and minerals, and enforcing international conventions on arms sales and corruption. If we were to focus on these kinds of policies in the future, that would be very different from the “poverty fixes” we have had for the last 60 years.

    I’m sorry to make your head hurt … 🙂


  3. Patrick Proctor

    I’m not sure what the implications might be for the thrust of Owen’s argument, but I wonder whether the starting point in a systems analysis should be the institutions. Those institutions themselves only exist as expressions of iterative patterns of interaction between individuals. Thinking of individual agency moves us even further from a fatalistic outlook and may be suggestive of other kinds of intervention to those mentioned. Any mileage in this observation?

  4. Ben Taylor

    Exquisite eloquence in the elucidation of our raison d’etre here, Owen. We’d see sustainability as a permanent increase in the adaptability of a system to the benefit of the poor.
    My only contention here, though, is similar to that of Samuel. The seven policies you outline all represent laudable aims which would undoubtedly benefit developing countries if they were to change. As you say, these are within the control of rich countries, but are they within the control or remit of development agencies? Everything operates within a system. The system within which these macro level policies operate is one of power asymmetries, self interest, and protectionism. The relative power of aid in respect of these concerns is minimal. There are things that aid can do but I’d see them as simply scratching the surface of major structural problems that inhibit the capacity for the strengthening systems in developing countries.
    I do, however, see another role for aid in respect of facilitating systemic change in developing countries. These are not, by and large, sweeping policy areas which we have the capacity to address. They involve an approach to intervention which is, in itself, adaptable and context specific, not focusing on ‘binding constraints’ but on-going analysis of how intervention impacts on a system and how it can react to capitalise on that change. The example you give of user fees in health care is interesting. In some cases this has led to a sustainable and pro-poor change – the development of private schools in India and Nigeria for example. But to focus on user fees is to focus on the ends of how a system is configured rather than the means of how it got there. To look at the same issues in a different context may produce a different result. In Kenya, sustainability is being seen through the introduction of user fees through water users groups while in Bolivia, the same result is being seen through an improvement in demand driven media and subsequent increase in the accountability of government to the now louder voice of the poor.
    Donors need to be facilitators of change within systems. This needs to be analysis led and context specific. It needs to be experimental and reactive to bring about sustainable change, but most of all, it needs to have poor people at its heart and any intervention within the system must maintain this focus.

  5. Owen Barder

    Patrick – I think we should be modest about both our capacity to analyse institutions, our capacity to predict change, and our capacity to influence change. I live in a country in which the entire chattering class did not predict a Conservative majority government in a political system which they presumably think they understand pretty well. Yet we seem reluctant to adopt any kind of epistemic humility with respect to developing countries.

    Ben – thanks. I agree that the policies of rich countries are also the product of the system from which they emerge, and cannot be technocratically tweaked. But I am optimistic that they can be changed, over time, in relatively inclusive and accountable political systems. That is what I try to do. I agree that donors could in principle play a role as “facilitators of change within systems”, especially providing investment for experimentation and learning; but I am probably less optimistic than you that they really do, or can, play this role effectively while avoiding the risks that I described that they might jeopardise, rather than enhance, the adaptive efficiency of the system.

  6. Eitan Reich

    Thanks Owen for this insightful piece and Duncan for putting the stage. Can I ask you about the Aid system itself? These are huge changes in conceptual thinking of a complex system, how would you start moving the Aid, Humanitarian and Development complex system to be able to adapt to it’s suggested new role?

  7. Gregg Gonsalves

    Owen is simply applying Darwin to development here, where variation in a trait in the presence of competition leads to selection, with a payoff in reproductive fitness. There is a whole field of experimental evolution where microbes are being pushed to adapt in vitro. Owen’s thesis could be put to the test in vivo; different communities could be randomized to received different interventions that promote variation in the context of competition, but what is the push, the payoff? Selection works in nature only because the endgame is reproductive success, what would drive selection in the context of Owen’s model?

  8. Duncan Green

    Nadeem Haque (@nadeemhaque) heroically wrestled with the comments process, but failed, so sent me this blog (, which summarizes his v interesting views on ‘intellectual dumping’

    The aid establishment has grown on the basis of 2 assumptions
    1. That there is a capital shortage in poor countries and
    2. That these countries lack the ability to make policy either because of knowledge or information shortfalls.

    The world has changed and these 2 assumptions are now untenable. Capital markets are flush with cash and they are eager to push it on to poor countries. The internet and globalization has made knowledge easily accessible to all. Most countries now have all manner of expertise. They are all exporting experts to the west.

    Despite these developments aid continues to grow. Financial flows are small. Now aid establishment is retailing policy advice, capacity building and technical assistance.

    While reports and consultants are surrounding poor country policymakers making them feel good, talent from those countries is being released to do outstanding work in the west. Yet their governments would rather have aid than bring back talent. And aid seems to be set to facilitate flight of human capital.

    The aid establishment is now huge. Bilaterals and multilaterals included we are talking of about 30 offices in Islamabad. All of them have officials who need to justify their presence. For that they have to show an agenda and work. They are playing to their bosses out there and have only a limited interest in the welfare of Pakistan. They have large sums at their disposal and agendas.

    Meanwhile, our government’s human resource policies are such that no competent well-educated Pakistani can find gainful, responsible and respectable employment at home. The only option for such talented people is to leave the country. Those who for personal reasons are compelled to stay in the country, they have to be employed by the aid establishment, reporting to junior aid officials and following their agendas.
    We all agree that governance is one of our biggest problems in Pakistan. Yet this system of aid perpetuates poor governance. Ministers are treated like royalty by the aid establishment. They are addressed as “Excellency” and given a lot of courtly ceremony. They are wined and dined on international trips that can hardly stand the test of scrutiny. They are invited as chief guests to conference/ceremonies. They are given the pulpit at various events to make empty repetitive speeches. In short, ministers are distracted from their real work and rewarded for keeping the status quo.
    The many donor bureaucracies now dwarf country administrations. The latter in any case have been stripped off their human capital and live with outmoded systems and technology, thanks to continued postponement of reform. Consequently, ministries have no ability to develop policy and other initiatives or to react to the various demands that face them. This suits donors who have agendas that need to be served. The less the opposition the better!

    The various aid offices and consultants ply ministers with more reports and proposals than the ministry has capability to understand. Serves both parties well but it does postpone reform and hence keeps our administration on a perpetual decline in terms of their capacity. It is not surprising to see all our governance systems are deteriorating.

    The continuous supply of consultants, technical assistance, training and conferences puts donors fully in charge of the agenda. When they say we need to measure poverty, everyone turns to it. They spend millions of dollars measuring poverty and arranging conferences for discussing poverty. They pivot to Trade with India, the whole country starts discussing that.

    Quite unintendedly, the aid enterprise operates as if it were “dumping” on our intellectual enterprise. Donor advice to cut the deficit while demanding increases in several heads from environment to social safety nets, the government has no money for research in any area. Only donors have liberal funds for research which they use for their agenda. In this environment, the entire intellectual effort of the country works for the donors.

    No money is available for local problem solving. For example, in Pakistan there is a consensus that civil service reform is critical for improving governance. Yet because the donors opposed this, the subject has never had any funding for research or conferences or ministerial presentation. The entrenched civil service resists discussion on this reform and donors wittingly or unwittingly oppose the reform.

    Our think tanks and universities are depleted of serious research and public intellectuals as the aid enterprise only funds low quality agenda based consulting. Such consulting also is stratified where the meaty contracts go to firms overseas that have symbiotic (often questionable) relationship with aid. Local intellectuals and firms are virtually hired as research assistants by the contracting firms sitting in distant capitals collecting huge overheads and fees.

    Perhaps the biggest unintended consequence of the aid enterprise is that the current arrangements are killing the intellectual enterprise in all countries. In this regime, chances of developing a Brookings or a McKinseys in a poor country are minimal. The best option for talent is to migrate.

    Angus Deaton has rightly noted that state capacity which is so critical to development suffers the most with aid. Add to that the rapid deterioration in intellect and problem solving capacity of the country and continuing failure to develop becomes somewhat clearer.

  9. Owen Barder

    Eitan – thanks. I agree that we have to think of the donor and aid institutions as shaped by the system within which they operate. (I am always bewildered that aid agency staff who claim to be able to achieve policy change in other people’s countries simultaneously claim to be impotent at home). Many of these ideas are gradually being adopted in various parts of the development system; but many of the drivers of rapid adaptation (variation and selection) are weak or missing.

    Nadeem – thanks. Your points about the negative effects of aid programmes on state capacity seem to resonate with the argument I was making in the blog post: that aid interventions might bring about temporary gains (at best) at the expense of damaging the long-term adaptive capacity of the system.


  10. Owen Barder

    Gregg – That’s right. Darwin’s “survival of the fittest” is one well known example of how evolution works. There are lots of other examples which we encounter all the time, sometimes consciously making the comparison with biological evolution and sometimes not. In economics the most obvious examples are the evolution of industries (think Hayek) and of stock markets; but this selection also occur in lots of less obvious ways (eg in consumer products, legal institutions etc). And in politics, political parties evolve to adapt to the selection of elections, and there lots of other mechanisms for evolution of institutions. Of course, Richard Dawkins made the point about how ideas evolve in The Selfish Gene (where he coined the term “meme” for this).

    In biological evolution, the successful reproduction of the gene is fitness function. In all these other evolutionary processes, societies can, and should, shape the fitness functions to match their values. So if we want more equal societies, we should consider what kind of political and economic fitness functions would be consistent with that. (There isn’t one fitness function, but many.) Norman Borlaug helped bring about the Green Revolution by overriding the fitness function for wheat to force selection towards high yield varieties.


  11. Alan Hudson

    Thanks for this Owen – very useful.

    This understanding of systems and the role that external actors can play in enhancing their adaptive efficiency, is the basis for Global Integrity’s new strategy.

    Our hypothesis is that we can play a useful role in facilitating country-level (system-level) learning, to strengthen (make more adaptive, more resilient) the system.

    We’ve put together a short survey to get some feedback on our emerging strategy and what it might mean for Global Integrity – working closely with country level partenrs – to strengthen learning loops at country level.

    Also worth connecting this discussion more directly to the Open Governance, Doing Development Differently and Adaptive Development discussions, including ODI’s recent report.

    Finally, provocatively, does this systems-thinking make you sympathetic to Jeffrey Sachs’ Millennium Villages Project? If not, could MVP be tweaked to be more about adaptive systems?

  12. Scott

    I mostly like what I see here, but am puzzled by the dig at growth diagnostics. It first does not appear to me that growth diagnostics is inconsistent with recognizing the complexity of development. It also is not clear to me that a government, or one advising a government, on what policies to adopt should not pursue growth diagnostics. When considering any candidate reform, presumably they should take account of the political constraints and the economic circumstances in which they find themselves, and make those adjustments which might optimize performance given non-ideal conditions. What am I missing here?

  13. Matthew Greenall

    Owen thanks for this, thought provoking as usual. Although I think your points on development are clear, I am feeling a bit short-changed on the question of how aid agencies can fix themselves. I agree with what you’ve said about where aid agencies go wrong. But at the same time I don’t think it is viable for countries or agencies to see catastrophes or crises as a symptom and to take the attitude that rather than respond they should step back and work on strengthening the systems. Of course I don’t think you are arguing that either. But if we agree that there’s an imperative to respond to a humanitarian catastrophe (even while we also look at the big picture), and part of our aid infrastructure is therefore purposed to responding, then it’s not hard to see why the aid sector gets into the difficulties you’ve described. The DEC rightly gets our money because it can save lives in the immediate aftermath of a catastrophe. But why stop at earthquakes? Other ongoing catastrophes like AIDS or gender based violence can and should also be seen through a humanitarian lens. But once a humanitarian approach is seen to be the answer, and once you are dealing with a problem as big as AIDS, it obviously becomes very easy for donors or agencies – and indeed, local responders in the affected countries – to ignore systems and policies and go in and try to back winners and “fix things”. Often but not always effectively, often unsustainably, sometimes against the grain of the countries where aid is being provided, and sometimes with long term negative effects as you state.

    I think part of the challenge is to try to figure out when “plugging the gaps” is appropriate and how, if at all, to bring gap-plugging in line with the need to address systems.

  14. Owen Barder


    Thanks. That’s a great question.

    I am trying to highlight a difference between:

    a. identify the gaps using technically sophisticated tools and then (in the best cases) be politically savvy about how you go about trying to fix those problems.


    b. identify interventions that will make it more likely that the economic and political system will adapt to solve the problems that create these gaps.

    I think of “growth diagnostics” as the first type rather than the second, whereas I think interventions of the second type are more likely to be effective IF the goal is to accelerate development.


  15. Owen Barder


    Thanks. I agree.

    I think “plugging the gaps is hugely important and worthwhile activity – whether in humanitarian crises, or by providing key public services in poor countries. I think we should do much, much more of it.

    But I think

    (a) we should be careful to do it in a way that does not do long term harm to the evolution of the system; and

    (b) we shouldn’t claim that it will catalyze faster or better development, nor structure it as if will.


  16. Kentaro Toyama

    Dear Owen,

    I’ve often enjoyed your critiques of international development, and I think your core idea that the world (and therefore development) should be viewed as a complex system has merits. Among other things, it is yet another way to see why attempts at central planning often fail. I like the general approach to tinker, which is similar to Easterly’s conception of “Searchers.”

    However, the recommendations here were deeply disappointing, and I think this is one of the greatest weaknesses of the theory. It’s hard to see how they differ substantially from standard neo-liberal, Washington consensus recommendations: free trade, transparency, communication, and innovation?!?! Why bother with complexity? We have plenty of mainstream economists telling us the same thing. If anything, you’ve revealed your underlying biases — while seeming to offer a solution different from standard economics, you’ve merely reconfirmed them.

    The idea that we are like an ant colony that requires no change in individuals seems woefully misguided. You seem to suggest that there’s no point in educating each ant, because in the end, what matters is the size of the colony. But, what we’re after in international development isn’t a bigger colony — what we’re after is improved quality of life for each ant. For that, there have to be changes at the individual level, in addition to systemic changes at the colony level.

    Ant colonies, incidentally, don’t change much in complexity despite their size — the same processes are duplicated over and over for every N ants. There’s nothing like the equivalent of globalized trade, which would mean that ants at one end of the colony specialize in something that ants at the other end don’t manufacture but trade for. Instead, even gargantuan colonies are just a archipelago of many self-contained colonies, each not substantially different from their neighbors. If we wanted a world of many subsistence villages sprawling across the planet, that would be the right analogy. But, if we want wise, educated leadership and citizenry, then it seems an unfortunate choice.


    • Owen Barder


      Many thanks for taking the time to comment.

      First, the ants. As I’m sure you know, I am not saying that humans are like an ant colony (nor am I saying that humans are like synapses, or cars, or air molecules). I’m using an ant colony (and the brain, traffic and weather) to illustrate a central characteristic of a complex adaptive system, which is that it can have “emergent properties”.

      Ant colonies are not all alike – for example, there are slave-making ants (, farmer ants ( etc. Nowhere do I suggest that what matters is the size of the colony. In the article I cited ( ants have started to work for queens other than their own. This is an interesting new “behaviour” which has not been previously observed; and it is interesting that we talk about the behaviour of the colony, rather than the ants. The point of talking about ants is to illustrate that these different forms of social organisation are characteristics of the colony, not of the ants. It is not the size of the ant colony that is interesting, but the way that the interaction of the ants, each responding to each other, generates these emergent properties of the system as a whole.

      Second, neo-liberal. If you mean neo-liberal in the technical sense, this isn’t entirely an neo-liberal agenda, but I do accept that there is a whiff of Hayek about it. I’m unrepentant about that. Some of the things that most mainstream economists recommend – such as trade liberalisation, action to prevent abuse of monopoly power, free movement of people, effective contract enforcement, property rights – are excellent advice in most circumstances. You may be shocked to know that I am also in favour of large parts of the Washington Consensus (I reckon seven of the ten principles are generally good advice and the other three often are too.)

      However, I plead not guilty to this being a rehash of the boring old neo-liberal economic prescription in two important respects. First, you’ll see that most of my recommendations are for developed countries rather than developing countries. I’m mainly not in the business of telling developing countries what to do. If rich countries would open their markets to poor countries, allow greater migration, reform intellectual property rights, protect sea lanes, cooperate on tax enforcement, stop paying bribes, stop selling arms etc, that would improve the chances of development happening faster and better in the developing world. I don’t recall telling rich countries to get their house in order being any part of the neoliberal policy prescription. Second, I do not accept a presumption which is often found in neoliberal thinking that “you can’t buck the market” – that is, I don’t share a kind of fatalism which requires us to accept whatever world survival of the economically fittest serves up to us. The second fundamental theorem of welfare economics ( teaches us that there are infinitely many Pareto-efficient equilibria which can be the outcome of a competitive market: which one we end up with depends on the initial endowments. We both can, and must, choose what kind of society we want to be – that is, which of many possible Pareto-efficient outcomes we want to head towards – based on our values and ethics. There is a huge role for society – mainly in the form of government – to shape the fitness function so that selective pressures take us towards it. If we want a more equal society – which I do – then we should work towards that. But we should do that wherever possible in a way that enhances rather than curtails the adaptive efficiency of our system. (For example, it would be generally better to tackle the accumulation of extreme wealth by breaking up monopolies and ensuring healthy competition than by allowing people to accumulate large wealth this way and then taxing them).


  17. Achim Kemmerling

    I like the argument, but, to be honest, I don’t share the premise. Development aid is not, I repeat not about development. Maintaining this idea creates a false normative benchmark against which reality can only look dire. Economists know relatively little about what causes growth. There is this bonmot that modern growth theory can explain everything about growth except growth itself.
    On a national level, ‘structural policies’, sometimes claimed to help regions grow are mainly redistributive instruments to mitigate divergence. The same goes for EU funds.

    Development is (hopefully) about social purposes, broadly defined. Poverty, inequality, insurance, whatever you like. Nobody would expect on the national level that social policy creates tremendous rates of growth. That’d be an added benefit, but that’s not the main rationale of the policy. So, again, thinking about development aid as related to development, is way too ambitious and self defeating.

    I therefore propose the following mantra: development aid is not about development. development aid is not about development. development aid is not about development…

    • Owen Barder


      Thanks. I agree

      (a) much aid is not intended to bring about development (I have set this out at greater length here:

      (b) aid may not be a very good tool to bring about economic growth and development, and we have little idea of whether or how it works.

      But I believe that other policies (ie not just aid) can help bring about economic growth. For example, if rich countries open their markets to developing countries, or allow more migration from developing countries, that could well accelerate economic growth, at least in some circumstances.


      • Achim Kemmerling

        Dear Owen,

        this piece you linked looks very thoughtful and well balanced. I start to think, however, that we should be more radical and eliminate all references to growth from aid. Maybe then the advocacy coalition will shrink, but we could focus more on those goals in aid that are realistic and thereby increase legitimacy.



  18. Kentaro Toyama

    Thanks for your thoughtful response, Owen.

    We disagree on Washington Consensus for developing countries. It seems clear to me that some protectionism is actually good for poorer countries, as I believe China dramatically demonstrated. The standard economic claim that almost all trade is good because it is voluntarily entered into and benefits both parties doesn’t acknowledge the fact that trade between greatly unequal parties easily turns to exploitation and corruption that often harms the poorer nation.

    However, it’s very interesting that you say that your recommendations are primarily for developed countries. I presume, given your focus on international development, that these recommendations are nevertheless meant to help poorer countries, as well. Here, I tend to agree — if rich countries actually took their own Washington Consensus advice, it probably would help poorer ones. I believe this is similar to Ha-Joon Chang’s thesis.

    Finally as to welfare economics and government role in defining the fitness function, I agree completely. Here, though, your recommendations seem to violate some of the claims of complex systems theory. Why should we expect that shaping fitness functions can be done (by human institutions) in a way that aren’t overwhelmed by the chaos of the system? It sounds like you’re suggesting that some human entities can be “outside” of the complex system, and manipulate just the right parts just so. But, isn’t this the same mistake made by those you criticize?

    I agree with your recommendations on this point, however — the disagreement is that the complex systems model, while valid for much of mass behavior, begins to break down when there are powerful bodies in the mix. I majored in physics, where it’s well-known that as soon as you have three bodies in space, you can no longer predict their future course (i.e., it’s a chaotic system). However, it’s also true that if one (or even two) of them is sufficiently massive, you can approximate very well and predict quite well, to boot — as is the case with our solar system. Drawing the analogy to global development, I’d say that powerful entities (like governments and large corporations) can in fact force or manipulate significant, intentional changes within the system. What’s harder is that with other large bodies in the mix, changes by one can lead to unpredictable responses by the others.


    • Owen Barder

      Dear Kentaro

      Thanks. It is useful to have this dialogue, not least because it gives me an opportunity to clarify things about which I am evidently not making myself clear.

      First, I agree that uncontrolled trade liberalisation by developing countries is not necessarily good for developing countries. That’s why #6 is one of the three Washington Consensus prescriptions which I do not believe is universally applicable, and it is not something that I have recommended (though in many situations I would). My recommendation is that +developed+ countries should open their markets to developed countries. I believe this would be unambiguously good for those developing countries which would get increased market access. This is one example of the many respects in which my recommendations are NOT the same as the Washington Consensus.

      Second: yes. My recommendations are mainly about what rich countries should do because those policies would help poor countries. (They would also benefit rich countries, incidentally, but that is not the main reason why I personally am in favour of them.)

      Third, I agree with your point that the policies and behaviours of rich countries are themselves endogenous to an complex adaptive system. But that doesn’t mean that they can’t change (see my reply to Ben above). As I said in my blog post, complexity should not lead to fatalism.

      I thought your use of the term “chaos” was probably confusing (if not to you, since you majored in Physics, then to other readers). Complex Adaptive Systems are not the same as chaotic systems. Roughly speaking, a CAS is where complicated relationships generate simple-seeming, emergent patterns; whereas in deterministic chaos a relatively small number of non-linear interactions create extremely complicated patterns. They are different types of non-linear systems with different characteristics.


  19. Joan Tallada

    Hi Owen. I have to admit that I was a little puzzled when I read your post (late, catching up readings in summer!). Your theories are a (rather simplistic) application of what in sociology is called “structural functionalism”, a theory mostly develped in the 50-60s of the XXth century but that can be traced back to French scholars Comte and Durkheim in the XIXth. As Giddens say: “Functionalist thought, from Comte onwards, has looked particularly towards biology [the ants in your example] as the science providing the closest and most compatible model for social science. Biology has been taken to provide a guide to conceptualizing the structure and the function of social systems and to analyzing processes of evolution via mechanisms of adaptation … functionalism strongly emphasises the pre-eminence of the social world over its individual parts (i.e. its constituent actors, human subjects).” Functionalism was discredited and declined because its neglect to address the role of conflict and inequality in society, as well as unable to explain how social change is produced – not adaptative, but disruptive change, as we have seen in history. All in all ,a rather conservative approach that focus more on fixing the inefficiencies of a potentially perfect system than in questioning the model and its implications and in encouraging alternatives. Joan

  20. But…

    It is possible to spend a good chunk (80 %) of the funding to in quite classic ways as a donor. With little money wasted. To change the system, necessary for long term, sustainable development, you need flexibel system thinking.

    To vaccinate children, which is necessary to save their lives, you need money and follow established protocol (OK with some complex systems sensitivity thrown in the mix). Sometimes you just need a bridge.