 Thank you. I'm going to present on the last of the three substantial topics that we covered in the Learning to Compete project. What I'm going to talk about is our findings in relation to agglomeration and its importance in industrial development. Just by way of introduction, when we talk about agglomeration, we're talking about the potential productivity benefits or productivity spillovers associated with firms being located in a close geographic concentrated area. What we're looking at here is a little bit about what the drivers of agglomeration are and a little bit about whether or not they actually lead to these types of productivity spillovers. Firm generally, we can think about them being drawn together for a variety of different reasons. We can summarize these in the motivation to reduce costs and its costs of transporting goods, costs of transporting people in a way, and also costs of transporting ideas. That's the way we would conceptualize it. There's evidence of clustering in both developed and developing countries. It's something that is thought to contribute to a lot of productivity growth. Understanding what the benefits are of agglomeration and what the driving forces are important for the design of spatial policies. What we find in our studies is that on the one hand, there are productivity spillovers, significant productivity spillovers associated with agglomeration, above and beyond that which you would naturally expect from being closely located in an urbanized environment. We also find that these benefits are not for all types of firms. There is significant heterogeneity in the extent to which these benefits are realized. In addition, the spatial distribution of economic activity is going to have implications for inequality within a country. This is another reason why it's important to think about these issues. There's a lot of evidence on agglomeration, its benefits and its drivers in developed countries, but in developing countries there was a significant gap. This is partly related to the fact that it requires a lot of data in order to be able to look at these issues. In learning to compete, what we try to do is address this gap and answer two particular questions. Number one, what are the drivers of agglomeration in poor countries? Number two, what the impacts are on firm-level productivity? Just to summarize very briefly when we think about the theoretical reasons why we might expect firms to agglomerate or to cluster together, aside from the natural advantages associated with locating in certain places, there are three standard reasons why we would expect firms to locate together. First of all, there is a broader market for input suppliers. Input suppliers themselves might locate where there's a large amount of manufacturing activity. They can benefit from economies of scale and then downstream firms can benefit from lower inventory costs and timely delivery of inputs and even sometimes inputs tailored to their needs. Also, in an industrial cluster, you benefit from a thick labor market, so a pooling of the labor resources that you need and it can facilitate a better matching of workers to jobs and a better matching of hiring and firing more generally. Then, of course, we've got another agglomerative force is that of technology transfers and knowledge transfers. That's spillovers that happen between firms because of their geographical concentration or their location near each other. In learning to compete, we looked at four different country cases where we had data to explore this issue. In each of the country cases, we had Vietnam, Cambodia, Ethiopia and Tunisia. We used firm level panel data with the exception of Cambodia, which was just across section. We first of all examined the pattern of clustering to see whether or not there was evidence that there was a clustering of different activities. We found that in all cases, firms are highly clustered. It depends really on the infrastructures to the extent to which that clustering occurs. For example, in Cambodia, which is a poor country relatively to Vietnam, poor infrastructure and with higher transport costs, we found a significant amount of clustering and a significant amount of clustering of the very small firms that due to these other constraints couldn't locate anywhere else. We had similar findings in Ethiopia and Tunisia. Those were the four cases that we were looking at. Then we tried to dig a bit deeper into each of these cases to see whether or not we could uncover some other insights for developing countries. From this, we've got three key lessons to raise today. The first lesson from what we have learned is that measurement matters. In fact, how you conceptualize agglomeration in developing countries also matters. Thank you. What we found in the case of Vietnam is that the source of the agglomeration economies and how you think about the types of knowledge transfers and technology spillovers, how you think about them actually happening, really matters. What we found was that a lot of the measures that are standard in the literature think about agglomeration in terms of employees and in terms of employment and that a lot of the spillovers, technology spillovers happen through that mechanism, workers moving jobs and so on and so forth. What we found in Vietnam is that in fact it's the entrepreneur that matters. The measurement and how you conceptualize thinking about entrepreneurs being in clusters and detecting these agglomerative forces needs to be given consideration when looking at agglomeration in developing countries. The second lesson was that agglomeration increases firm-level productivity. Now, this is very challenging. There's a lot of identification challenges associated with finding these productivity spillovers or identifying these productivity spillovers econometrically. But we did our best to do that within the different studies. What we did find that obviously unsurprisingly, urbanization economies are important, but controlling for these, we found evidence of further productivity spillovers. In Vietnam, the evidence was that the small firms are the ones who gain more and there were significant spillovers for foreign-owned firms in fact. In Tunisia, they also found evidence of a transmission of ideas between firms that are located close to one another. In Cambodia, there was weaker evidence of net productivity gains overall, but there was some evidence for informal firms in particular which was quite I think a novel addition to the literature, the fact that we managed to look at informal firms. In Ethiopia, the agglomeration firms had higher productivity, but only if they produced similar products. The interesting thing about these findings as well was that in both Cambodia and Ethiopia, there was evidence of strong negative effects on prices. This is competitive effects, which is what you might expect. This brings us to the third lesson, which is that clustering was not the best option for all firms, that the increased competitive pressures in localized markets will drive down the price of goods and services as we would expect, but poor infrastructure, lack of regulation, and other constraints will lead to firms having no choice really but to locate close to each other. This is problematic for very small industries or very small firms who don't have the opportunity, if you like, to grow, expand, and invest. These competitive pressures, if you like, are a countervailing force when we look at agglomeration economies. There are productivity spillovers for a lot of firms, but a lot of the small firms do not benefit on net. Of course, consumers will benefit from lower prices, but in terms of growing industry and growing small enterprises, that is problematic. Our conclusion there was that clustering is not necessarily the best option for all firms. Just to conclude on some of the broad policy implications, what we found was that there are potentially significant benefits from agglomeration. Of course, this points to the possibility that a policy of attracting firms and encouraging firms to locate near each other through SES, for example, might be a wise policy. I'm not sure, but maybe John will also speak about that in a little while. But of course, you have to bear in mind that a lot of the evidence from the SES that's out there, and not something we study specifically in our work, but the other evidence that's out there has suggested that the SES in Africa have not been that successful, and particularly not near as successful as they have been in the rest of Asia. The evidence shows that there's low levels of investment, there's limited exports, not much job creation in general. They haven't been as successful as what was originally expected. But recently, there has been a renewed commitment in that amongst sub-Saharan African countries, and a lot of this is driven by China, where they are actively establishing a lot of SES. I think it will be interesting that a lot more further research is needed in order to see whether or not this has the desired effect. But the other policy implication is that not all firms are benefiting from clustering, and a lot of this is due to the fact that they're constrained by their location choice, and they're constrained by the infrastructure, they're constrained by regulation, and other factors. So giving more flexibility in that location choice is important, and that will require investment in infrastructure, and in making the market more flexible in general.