Linking: Rodrik on industrialisation

One of the leading scholars on the topic of industrialization is Prof Dani Rodrik. Two of his recent blogposts are relevant for the readers of my blog.

The most recent post by Prof Rodrik is titled “Premature deindustrialization in the developing world“. In this article he explains that industrialization is affecting the developing world more than the industrial world. This is a brilliant read. The full NBER paper that his blog post is based on can be found here.

Another recent post by Prof Rodrik is about services, manufacturing and new growth strategies. In a presentation that he mentions in this post he argues that many developing countries are focusing too much on unproductive small enterprises that face high costs, but that these same small enterprises often absorb low skilled labour. If I say anything more I will most likely mess up his argument, so take a look for yourself!

Looking back at my 2014 blog posts

Before I get blogging in 2015 I want to take a moment to reflect on the most commented on, quoted or disagreed on blog articles that I have written in the last year.

Actually, one of my most popular posts was written in August 2013. The article is called “Is aid systematically unsystemic“. On this particular article I received the most e-mailed comments and also the most links from other bloggers. Since the publication of this post I have received at least 30 emails with people saying that they agreed, disagreed, or wanted to just exchange experience.

During 2014 I had three blog posts that also generated many comments and e-mails. The theme of last years posts was all about industrial policy and especially taking a bottom up perspective of how to improve the industrial system.

The three posts that stood out in terms of comments and controversy were:

Industry development under conditions of complexity – In this post I argued that under conditions of complexity, the best approach is to diagnose through intervention, which means that there is no real separation between diagnosis and intervention. While many readers working with private sector development directly agreed with my statements, some experts working on advising national governments on industrial policy strongly disagreed with my statements that you do not gain insight into a complex system through analysis.

Another post that performed well on referalls and comments was posted in the middle of the year. In Industrial policy is different at local and national levels I argued that at the local level industrial policy (or locational policy) must be much more focused on the private sector as it performs now, and what can be done by various market and non-market organizations to support the development and expansion of the private sector. Again, many economists did not agree that at the local level the approach requires a more qualitative approach that involves actually speaking to business people and representatives of development organizations on a regular basis.

In a post that had little to do with industrial policy on the surface, but more about my ongoing research into complexity thinking, I discussed the implications of recognizing competing hypothesis as an indicator of complexity. The idea of Recognizing competing hypothesis as complex  had its origins in a training session conducted by Prof Dave Snowden on complexity and decision making under conditions of uncertainty. I did not mean a hypothesis like those formulated by a PhD student or researcher, rather a hypothesis as statement of a coherent argument that seems plausable. The link between this article and the rest of my posts is that in industrial policy there are often preferred or desired hypothesis or states that are being pursued, even if they are not explicitly formulated as hypotheses. I tried to argue that keeping our options open or even purposefully increasing the range of options is desirable.

Thank you to all my readers who have written comments and have e-mailed me about these posts. I understand that registering an account on WordPress may be a bit of a hassle, so I thank those that have gone through the trouble and that are always willing to comment and contribute. Thank you also for the tweets and reposts on other sites. Prof Tim Kastelle, my favorite guru on innovation deserves as special mention.

I will start writing again within the next few days. I just wanted to take a moment and reflect on the most influential past posts. Let me know which posts you printed, saved or forwarded, and which posts you think should be expanded upon!

 

 

The oblique search for new industrial opportunities

Industrial policy is typically set at national level. It is often aspirational and attempting to “stretch” an economy into new kinds of production and value addition. Programmes are designed, targets are set such as doubling manufacturing contribution of x% within 7 years. Therefore it is sometimes disconnected from the present as it seeks a new Status Quo, a different structure of production.

Yet the natural process under which new production activities are created is complex. It is not as simple as finding a market opportunity, finding the right production process, securing funding and launching a business. The economic context, the political climate, the entrepreneurs with the right levels of experience, backing and confidence are all needed. And don’t forget individuals with a desire to expand, take risks and try new things.

Danni Rodrik argues that Industrial Policy should be a search and learning process. Many centrally planned industrial policies even cite Rodrik as they then commence with outlining with great certainty what must be done, by whom, with which resources and to which effect. This logic completely ignores the importance of what exists, and what is possible from here. It ignores that fact that the past matters, and that the current structures are the result of a series of evolutionary steps. Complexity science teach us that these plans ignore the fitness landscape, a landscape that is dynamic and constantly changing. Any attempt to extend the horison further than what is within reach should be treated with great caution. One of the greatest obstacles is the attide towards risk and the optimism of enterprises. I don’t think Rodrik meant the ministers officials must do the search, rather, industry must do the search or at least be actively involved in the search in partnership with government and institutions.

But the search is not about answering a simple question. A more oblique approach is called for (see John Kay, Obliquity). Which means we should set aside targets and indicators, and focus on creating small experiments to introduce more variety and options into the system. It means that finding out that something is not possible is as valueble as figuring out that something else is indeed possible. Taking Rodrik literally, it would mean also giving much more attention to what entrepreneurs are searching for and experimenting with in the background. It requires that we recognise that the current economy is creating what is viable under the current dynamic circumstances, and that only strategies that recognise where we are and what is certainly within reach from here is in fact viable. The challenge for developing economies is that what is possible is typically limited and further constrained by strong ideological bias as to what is possible or desirable. For instance, many South African business owners are trying to shift out of price sensitive markets competing on a basis of low cost skills. Entrepreneurs are moving into knowledge and capital intensive production, with more focus on service and integration. Government is searching for a way to employ people with low skills because its own social programmes and service delivery is not a viable fall back for people with insufficient skills.

The search is not about analysis
Complexity describes a situation where the patterns of what exactly is going on is unclear or shifting. We cannot entirely figure out what is leading to what and what is reinforcing what. Due to the dynamism, we cannot really understand the situation better through analysis. Another way of explaining this, is that a situation is complex when more than one competing hypothesis can with some probability explain what is going on. The only way to make sense of complexity is to try something, actually, try many things. And then see what seems to work better. It means that we start with what we have and who we know (and can trust), and then try a range of things with the simple purpose of seeing what is possible within the current constraints of the economic system. Steps must be taken to reduce risks (for instance by ensuring that the costs of failure are small, or that the experiments try different ways of solving the same problem), but then this whole approach in itself must be recognised to be politically risky.

This is where donors and development partners come in. By assisting developing countries to conduct low key experiments in order to create variety is essential, as development partners can reduce the political risks of their counterparts. This approach will furthermore require the abondenment of targets and indicators as an attempt to measure accountability and progress. A more subjective approach that sets indicators that monitors the overall health or dynamism is needed so that the experimentors can sense when they are indeed making progress. Thus the indicators does not measure success, nor input.

Perhaps then a skunkwork approach to a more complexity sensitive industrial policy approach is needed. Let the normal industrial policy targets and rigmarole be there. Politicions and bureacrats like this sense of certainty and purpose. But allow for some experimentation on the side under the heading “industrial policy research”. Allow this team to work with private sector partners to conduct small experiments to try new business models in an incremental way. For instance, do incubation to try new ways of mineral beneficiation, but without investing in large buildings or expensive equipment. Use what is existing as far as possible, even if it means having the manufacturing done on a contract basis elsewhere in order to test if local demand for the outputs exist.

Landing spacecraft on a comet but still not enough development

Since the landing of the Philae lander on comet 67P/Churyumov-Gerasimenko yesterday I have been asked a few times times by readers and friends why this is possible, yet we struggle with development, inequality, racial transformation in RSA, exclusion, inclusion and poverty.

I have two answers. Firstly, landing a spacecraft is by now no longer complex, it has become merely very complicated. I mean that once you can figure out the questions, many answers are self-evident. For the parts that are not self-evident you can conduct research and development, and choose between alternatives based on the results. You still have a huge problem with sequencing, but this is why these space missions are so expensive. That is why India can put a spacecraft in space (it is complicated) but still struggle with gender and poverty issues (it is complex). That is why South Africa can host the SKA (very complicated) but not deliver water to communities (complex, as it involves politics, competencies and competing priorities). We just don’t know in which order to solve all the problems in our developing countries, and everything seems to affect everything. Development is complex. In fact, we are not even allowed to fully unpack or discuss the problems because we have become overly sensitive, making things even more complex.

The second answer is about science. For space, we have science. Once politicians allocate funding to an agency, technocrats and scientists take over. We have scientists and engineers arguing about principles, about facts and about theorems. Experiments are conducted. Tests are run. Data recorded, processed and compared. There is a lot of debate allowed and even if criticism is never nice to receive, it helps to refine results, arguments and propositions. There is also the scientific method which means that even if I think I am right I must still convince editors and reviewers and funders with evidence in the form of data and experiments that can be repeated.

Our problem in development is that we do not appreciate criticism, never-mind not relying on proper research. If I question black economic empowerment policies in South Africa I am labelled a racist, even if I believe some kind of redress is needed but I am merely questioning the current modalities. If I question the way donors select value chains based on preferential impact on women I am described as being against inclusion and social justice. If I question the focus on low skilled jobs I am thrown out as a market liberal or capitalist. We just don’t allow sufficient debate backed by proper research. In many countries where I work criticism is not welcomed or appreciated.

I am afraid that the same can be said of the climate change debate, where any person that questions the prevailing consensus is quickly dismissed as a a person in denial.

Extrinsic and intrinsic rewards

Anybody that has read any recent articles on management, HR, strategy and leadership will know that you cannot use extrinsic rewards to motivate people extrinsically. Yet, in my daily experience of working between ordinary employees and management, in firms and in public institutions, it seems that this logic rules supreme.

The most acute form I find in South African universities, where journal articles are counted by everyone but the good lecturers and researchers. The good researchers and thinkers in our universities don’t need the recognition (nor the little research grant they are paid in South Africa). Our best lecturers and researchers are driven by an intrinsic motivation. They love teaching, they enjoy their work. They love the technology they get to work with. If they were driven by extrinsic rewards they would have been working in the private sector, or running their own company.

I find the same crazy logic in economic development. Except the extrinsic rewards is not so aimed at the individual, but a team or a programme. Perhaps it is not even a reward, it is a unspoken threat. However, our best practitioners out in the field don’t need specific measures (treated as targets), they know what resources they have, and what is not working as it should. Perhaps they don’t know exactly how to fix the complicated nest of interrelated problems, but they are intrinsically motivated to find solutions. To try and try again. Of course the extrinsic reward of a good income matters, but it is not the highest priority. I often think the targets and indicators set for development is a better indication of a lack of trust by donors in their employees (and their counterparts) than it is about making sure activities are leading in the right direction.

The future aint what it seems

I have written many times before about my serendipitous journey into the topic of complexity. One of the important insights for me is that we cannot really predict much of the future under conditions of uncertainty. While there are many things for which we know what the consequences are, we have to acknowledge that there are many situations where we simply don’t know how things will turn out.

As I became more sensitive of the consequence of the insight about unpredictability I realized how much of my work hinged on assisting customers to somehow plot and engineer a specific future path. I moderate at least one strategy session for some or other developmentally minded organization every month, sometimes many times more. All these organizations want to set their portfolio of interventions into motion, and want to make sure their plans are foolproof and environment proof – meaning that failure can be avoided somehow.

Recently I started following Dave Snowden’s advice, assisting customers to have much deeper conversations about what is going on NOW, and what is possible NOW. We’ve been using the 3 Criteria for Quick Wins for a while (see note below), but now I emphasize living in the NOW. At first I felt a bit insecure to insist that we stop trying to focus on the ideal future, but now my confidence has grown. The amazing thing is that many of my customers are responding positively to this focus on what is possible now. Maybe it is more intuitive to work from the current. Maybe South Africa has become so complex that we can actually not afford to spend much time in the future.

I must add, we do still look at the future. I am not promoting a junkie style of optimizing the current without a view of the future. There are some things that we know about the future. For instance, if a University decides to increase investment in post graduate research, they know they will increase revenue, increase research outputs, without necessarily increasing fixed overheads. But we don’t set a high goal, set milestones and lunge into action. We start by saying “how does post graduate research work now?”. We explore the options, the possibilities and the obstacles. We also look at what we’ve tried in the past and whether the context has changed so that we can try a small experiment again. Then we develop a portfolio of small low risk interventions that can be executed simultaneously.

I have been following this approach for just a few months and I must admit that I am pleasantly surprised by the outcomes. Of course it felt weird in the start to leave customers with a portfolio of experiments instead of a clearly developed log-frame like project plan full of milestones, champions, indicators and deliverables. But I can see how my customers’ organizations have become a more healthy, balanced and perhaps even more naturally innovative.

The future ain’t what it seems because we have so many things we can do in the present. It takes real leadership to work with what we have and it takes real courage to break with the typical management-style of detailed project plans, log frames, project charts and the like.

 

Note about quick wins.

As Mesopartner, we define a quick win activity as one where:

1) The resources are within our control. This includes funding, but also key resources, key people and willing champions

2) The results are easy to communicate. Preferably the results are visual so that the benefits of change are disseminated easily to others.

3) We can take the first steps of implementation very soon, within days or weeks.

 

 

Industry development under conditions of complexity

Most economic development projects have a tendency to separate analysis from intervention or implementation. This follows on an engineering approach where you must first understand a problem or issue before you can design interventions which is then logically followed by implementation and later on evaluation. I will not now go off on why this logic is questionable as I have written about this before and we have dedicated the Systemic-Insight.com website to this topic.

But complexity thinking is challenging this norm of separating analysis and intervention.

Auwhere to gothors such as Snowden argues that under conditions of complexity, the best approach is to diagnose through intervention, which means that there is no real separation between diagnosis and intervention. Practically, you might have to spend some days and a little bit of effort to analyze who is interested in a particular issue so that you know where to start, but you have to recognize that even asking some simple questions is in itself already an intervention. Furthermore, the objective of working under conditions of complexity is to introduce more variety so that different approaches to overcoming constraints can be tried out simultaneously. This means that small portfolios of experiments must be developed and supported, trying many different ways to solve a problem. Many of these are guaranteed to fail, but new novelty will also arise. The health of a system depends on more options being proven viable. Strong alignment of interests, priorities and interventions are actually unhealthy for a system in the long run.

I’ve had this discussion many times with fellow practitioners in the last years and usually at some point somebody would say “but not everything is complex”. I agree. They would argue that there are definite casual relations between for instance education and economic development. Well, this may be true in some places. However, whenever a government (or a donor) decides that a particular sector or industry requires support it should assume that the issue is much more complex than it may appear, otherwise the industry actors and supporting organizations and demanding clients would have sorted things out by themselves.

The idea that diagnosis takes place during intervention has many detractors, despite the fact that many strong economic development organizations intuitively follows this process logic of working with diverse stakeholders in an ongoing process. Here is a short list of some of the detractors and their main reason for resisting such a process approach:

  • Large consulting firms: They would fight this approach as processes are much more difficult to quote and manage than a clearly defined project. Furthermore, this kind of approach depends on more expensive multidisciplinary experts that require a combination of technical, facilitation, change and business skills. The number of people that can support such a process are few and far in between.
  • The public sector: To overcome constraints created by complexity requires that dissent be nurtured and premature alignment be avoided. This is also risky for the public sector as things may not be so neat nor supportive of past policies and decisions. Furthermore, when more options are created it is not certain which firms will really take up the solutions – meaning that in a country like South Africa with strong benefit bias this is too risky, as preferred candidates might not be the beneficiaries of public support.
  • Donors and development organizations: Simple cause and effect interventions that depends on controlling certain inputs in order to benefit specific target groups still dominate the logic of donors. Therefore a process that is not specific, and that explores different alternatives may not be appealing to donors. Furthermore, donors are expected to be able to very precisely report not only in inputs, but also on impact. A process that has multiple shifting goal posts makes planning and resource management very difficult. However, many examples exist of donor supported projects that are very open to this approach, but this is mainly the prerogative of the programme managers deployed into the field – it is not systemic.
  • The private sector: Yes, even firms may resist an open ended and exploratory approach. One reason is that firms try to push the problems experienced in the private sector back onto the public sector (blame and responsibility shifting). An exploratory approach puts much more onus on the private sector to not only contribute, but to be open for alternatives and to then actively pursue opportunities that arise. Secondly, the incumbents in the private sector sometimes profits from a disorderly system. Many existing firms will resist newcomers trying different things and trying to create new markets, as this disrupts the way things are done at the moment. In a complexity sensitive approach we have to on purpose introduce novelty into the existing structures, and this means challenging some of the dominant views and agreements about what is going on, what must be done and why nothing has changed. This is very unsettling for the existing actors.
  • Top management in an organization: Management science in itself assumes many casual relations. For instance, strategy development typically starts with defining a vision and objectives, and then making sure that everyone is aligned and committed to these goals. As one of my favorite strategy David Maister argued  “strategy means saying no”. This means that resources are dedicated to a few specific areas in the belief that addressing these would have predictable and desirable effects.

Now I must state that in more ordered domains, where there is less complexity, many of the arguments outlined above are valid. In a small organization with limited resources priorities must be set. Governments cannot help everyone, so somehow a selection must be made. However, I believe that industry development is in many cases complex also because it is so hard to see how unpredictable effects will affect an industry.

I am grateful that I work with organizations that are willing to embark on industry development or institutional development processes that are more complexity sensitive. I believe that such an approach is particularly important for innovation systems promotion and for industrial policy. I am surprised at how many manufacturers and universities have agreed to embrace a more complexity sensitive approach to development, strategy formation and developing new services/products. All involved have been amazed at the early results this far, as these processes typically unleash a lot of energy and creativity by different stakeholders that in the past were more than willing to just observe from a distance what was going on.

Lego for serious girls – my children will be ecstatic

Raising girls and trying to convince them that they can become anything they want to is not so easy. There are so many stereotypes, and even gender empowerment has become stereotypical.

Lego seems to be responding to this slowly. Take a look at this report on CNN. In the meantime, I will try to order my girls a LEGO girls scientist set!

Promoting sectoral innovation systems

I am receiving more requests for support to diagnose and improve innovation systems than ever before. It’s just been a few years since I have decided to focus all my attention at working with the upgrading of regions and industries from an innovation systems perspective and I am pleased that this decision is working out.

The most popular demand is for support to promote sectoral innovation systems. However, people confuse the “sectoral” with a classical sector driven approach. In a purely sector driven approach the focus is on a broad group of firms that falls within a broad industry classification. This may to some extent include some suppliers and key customers, but even a sector-based approach can still be too broad to tell us much about the patterns of innovation, how knowledge is used, and how institutions respond to the typical market failures in that sector.

A sectoral innovation system is more about how different groups that uses a common knowledge and technological domains work, how knowledge flows and how technology (which includes knowledge) evolves. To quote my own work (Cunningham, 2012)

“According to Malerba (2005), the emphasis of sectoral innovation systems is on a group of firms that develop and manufacture the products for a specific sector and that generate and utilize the technologies of that sector. The boundary of the system is drawn around a technological paradigm that is formed by a knowledge base, specific technologies and inputs, the different actors and networks that are systemically interacting, and the institutions supporting a specific industry. This is an important difference from value chain analysis, where the logic of the chain is determined by the system surrounding the conversion of a raw material into a product for a market. “

What I am trying to say is that instead of looking at the manufacturers based on similar inputs (raw materials, equipment, skills) and outputs (products and services), in an sectoral innovation system approach we look more at the common technological or knowledge domain that brings various firms and institutions together. This knowledge domain could spread over several industrial sectors, linking different value chain actors together. In fact, many industrial clusters often emerge around a particular group of complimentary knowledge bases. For example, aluminium die casters, aluminium casting equipment manufacturers, and their key customers in the automotive and aerospace industries would make in interesting sectoral innovation system to investigate. On the surface, automotive and aerospace companies don’t seem to belong together, but from a knowledge and technological domain around aluminum processing and its applications it makes sense.

The second part that people get wrong about a sectoral innovation system is that it goes way beyond innovation at the level of the firms. While the physical results of innovation is often easy to see at the levels of firms, this is just the tip of the iceberg. The innovation system describes how knowledge gets created, shared, forgotten and the dynamic relations between them. Furthermore, in any innovation system approach attention must be given to how policies and rules create incentives to innovate (or not to innovate).

Sectoral innovation systems researchers distinguish between high R&D-intensive sectors (such as electronics or drugs) and low R&D-intensive sectors (such as textiles or shoes). These systems change over time as the different elements co-evolve. This means that within a traditional economic sector like the foundry sector (using standard industrial classification schema) there could be areas that are more R & D intensive (such as aluminium) and other parts where the R & D is mainly done either by equipment suppliers or customers. Each of these different intensity R & D systems within the foundry sector would constitute the starting point of a sectoral innovation system. Another example is the machine tooling sector. In some knowledge domains, tooling is developed by the customer of the toolmaker that is developing a new product. In other knowledge domains, the toolmaker is responsible for assisting a customer to come up with a tooling design. Yet in another area, equipment manufacturers push toolmakers to adopt new ways of making tools. For me these are all different sectoral innovation systems. Lastly, these sectoral innovation systems can also be very different within a country like South Africa. Some regions may be dominated by downstream industries like packaging, while other regions might be influenced more by the availability of high quality infrastructure, market density and logistics.

Let me stop here to keep the post short. In conclusion, a sectoral innovation system approach is more about the knowledge and common technological domains than it is about standard classifications of industries and sub sectors. Within an economic sub sector (like tooling or foundries or food processing) there could be several sectoral innovation systems. To make matters more confusing, several different sectors or links in a value chain could be brought together within a particular sectoral innovation system around specific knowledge or technology domains.

I am looking forward to your questions and comments to this post.

 

Sources:

CUNNINGHAM, S. 2012. 2012.  The fundamentals of innovation system promotion for development practitioners. Leveraging a bottom up understanding for better systemic interventions in innovation systems. Mesopartner Monograph 5. Mesopartner.

MALERBA, F. 2005. Sectoral Systems. How and why innovation differs across sectors. In The Oxford handbook of innovation. Fagerberg, J., Mowery, D.C. & Nelson, R.R. (Eds.), Oxford ; New York: Oxford University Press.

 

Industrial policy is different at local and national levels

Industrial policy at the national, provincial or sub-national and local levels is different. While at the national level, industrial policy is often focused on coordinating public resources around certain priority areas, local industrial policy is almost completely focused on the pressing issues of the private sector and organizing the public sector around these needs. While at the national level, selecting opportunities for investment is often difficult and focused on the future, at the local level industrial policy might get trapped into grappling with “what is” and the legacy of the past.
At all the levels policy makers will be grappling with balancing “what we have now” with “what is desirable”. All too often “what is possible or within reach with what we have” is not asked enough of public and private actors. These questions are much harder to ask and to answer at the higher levels, because the industries are further away or maybe not even entirely visible, and emerging competencies in public and private actors may still be hidden.
At the local level, business is more visible. Unfortunately, at the local level past relations and power struggles between various actors still shape the current dialogue and possibilities for future collaboration. Therefore, industrial policy implementation at a local level must have a strong process element that attempts to reconfigure stakeholder relations around areas of common potential or concern. In our practical experience we know that at the local level it is easier to mobilize the private sector around problems (such as skills shortages and inadequate infrastructure) than around opportunities. However, it requires a certain confidence and maturity of local government and local public agencies to engage with the private sector when they know that they will be dealing with complaining business people. The one thing both the local private sector and the local agencies of the public sector have in common is limited resources. Perhaps local industrial policy then should focus on making the best of the existing limited resources. The focus should be to find opportunities for collaboration that can be exploited in a process approach, not focused on large projects or a grand vision dominated by the public sector, but on a process of finding small opportunities to make better use of local competencies, local knowledge and local capacity in both the public and the private sector. I am not arguing that local industrial policy must be completely inward looking, as the relation between local firms and external markets are an important resource. However, I am arguing that local industrial policy must start with the current reality while mindful of the past and focused on what is called the adjacent possible. The adjacent possible means opportunities or solutions that are within reach by combining, recombining and maybe adding a little to what we have now.
I conclude by stating that at the local level, industrial policy is not so much about the public sector supporting structural change or achieving a vision of new industries. At a local level, industrial policy needs to be entrepreneurial in that it should focus on exploiting existing resources, knowledge and competencies to the fullest. Local industrial policy must have a process approach that does not get trapped into existing stakeholder and sectoral interests, but that strive to unlock the potential of the different knowledge bases and competencies in the locality to solve existing problems in innovative ways, while searching in an ongoing basis for opportunities for collaboration.