 Let me briefly introduce myself. My name is Dimitri Stillinga. I work for Latrite, which is an economics and social research firm based in Kigali. And we're now expanding as well to Malawi and Burundi. How did we get involved in the Learning to Compete program? Well, we worked with IGC, and we got selected to do two pieces of work. One was a book on Rwanda's agribusiness and manufacturing sectors, a bit like the enterprise mapping work John Sutton was talking about this morning. And the second piece of work was a study on Rwanda's export sector, its structure. And we also looked at we tried to explore a bit the learning by exporting hypothesis. Now, my objective for today is to do three things. One, I want to give you a brief overview of the structure of Rwanda's manufacturing sector, what it looks like, what drives it. Second, I want to show some highlights about how exporting works at the firm level and what that means for learning by exporting. And lastly, hopefully, I hope to spark some new ideas about the learning by exporting hypothesis. Now, let me give you a brief overview about Rwanda's export sector. First of all, as you can see in this graph, which depicts merchandise exports over GDP, it's very clear that merchandise exports or the exports of goods have not been a driver of growth in Rwanda. They've been stuck at about 5% of GDP for the past 30 years. Now, these 5% I divided into two big, broad groups. One is commodity exports, which is tea, coffee, and minerals. And that accounts for 90% of those exports, those merchandise exports. The remaining 10%, 11% is other products and manufactured product exports. So you should think about these two export sectors as completely distinct things. They're two different animals. One way of looking at it is to look at destinations. All mineral and other commodity product exports, coffee and tea, go to international markets, in particular Europe, America, and Asia. All other product exports, manufactured products exports go to neighboring countries, DRC and the broader EAC region. Almost 100%. Now, the commodity sector is a relic of the past. It started in the 30s. And the investment happened between 1930 and 1950, spurred by the Belgian colonial administration, who set up mines in about 200 locations. These are the same locations where mining is done today. They did mandatory, they asked farmers to mandatorily switch their cropping to coffee. And in areas where coffee couldn't grow, they started growing pyrethrum. This was because of the crisis in the 1930s. The Belgian colonial administration needed the colonies to also start exporting because of the economic crisis in the world. Tea came a bit later between the 50s and the 60s and was helped by European investors and then development aid. Now, these sectors, I started in the 30s from 1930 through to 1960, account for 90% of Rwanda's merchandise exports today. Not only that, but they've been the main driver of exports growth recently. So if you look at this, you can see that established product exports, so old products, account for 75% of export growth between 2000 and 2010, let's say, of the past decade. Out of new export discoveries, new products, which account for 20% of growth, 60% are also commodity products. You can see three products alone, fully washed coffee or specialty coffee. And two mineral products, tungsten and chromium, account for 60% of new product exports, again, between 2000 and 2010. Now, the situation in the non-commodity export sector is completely different. The vast majority of exports go to two destinations alone, DRC and Burundi. And there are some small exports as well to other EAC countries, including Kenya, Uganda, and Tanzania. However, it's a sector that's growing the fastest. It's also the sector where the future of Rwanda lies. Because if you look, 90% of Rwanda's exports today sit here at the periphery of Rwanda's product space. Here you have tin, minerals, there you have tea, up there you have coffee. And firms in these sectors are in very sparse areas of the product space. They're not going to diversify. The action is going to be happening right here in the more dense areas, which are the kind of products Rwanda exports to Burundi and DRC, construction materials, plastic tanks, beverages, and food, et cetera. Now, it's a sector that is poised to grow very quickly in the near future. This graph here depicts the number of new manufacturing firms every four or five years. And you can see that the number of new manufacturing firms follows the course of history a bit. You have a dip around here, which corresponds to regional crisis in Burundi, Rwanda, and the Great Lakes area. And there's another big dip, of course, during the genocide. And the period we're in now is the highest pace of industrial growth Rwanda has seen since the 1970s. Since 2005, and in particular since 2007, when Rwanda entered the EAC market, this growth has been coming from regional investors. Regional investors in the flour market, in the meat flour market, in the construction material sector in particular. Now, let me tell you how things happen at the firm level in Rwanda. And maybe this can give some insights on how learning by exporting can be analyzed in countries like Rwanda, which are very small economies with very few exporters. Now, point number one is there are very few exporters in Rwanda. Out of the sampling frame we were working with was basically all taxpayers in the country. There are more than 1,000 firms. Now, if you look at exporters amongst those taxpayers, and you take out commodity exporters, you take out retailers and wholesalers, you're left with somewhere between 17 to 40 firms, depending on how you determine what an exporter is. If you look at firms that export more than zero, you can say there are 43 exporters in the country. Firms that export more than $50,000 we're only talking about 24. And firms who export more than 10% of their sales, we're talking about just 17 companies in the whole country. Not only that, but the firms that do export are not very export oriented. A small share of their sales are actually exported. There's only one company that exports more than 50%. Only one manufacturing firm that exports more than 50% of its sales. It's Société Rondez des Chaussures, which is a plastic shoe manufacturer. And the reason he's exporting, it is a Rwandan. He lived in Burundi and worked, was running the Bata shoe factory. But when there was a crisis in Burundi in the 80s, he fled to Rwanda, came back, and started a shoe company there. The market he knows best though is Burundi, which is why he's exporting. If you look at the next four companies on the list, Bakresa, Grain Mill, Steelra, and the Kigali Sement company, these are all companies that were started in 2011 by East African investors. Take them out of the picture, and you can see that the export orientation of Rwanda's manufacturing sector is extremely low. Now, what further complicates analyzing the learning by exporting hypothesis in a country like Rwanda is that you have no comparatives. Rwanda makes all these products. The manufacturing sector makes candles, wooden furniture, nails, paper products. But they're only one or two producers in each sector. It's like mini monopolies in each sector of the manufacturing economy. So again, that is one issue that complicates the learning by exporting analysis. The second thing is there's little switching. And this is something that John Rand also referred to in his presentation yesterday on Mozambique. And I think it's a case of many African economies. So in the case of Rwanda, we can say two things. Firms that are exporting firms in the commodity sector are born global. All of them are exporters, first of all. And second of all, they're literally global, in that they're owned by large international firms that have networks all over the place. New firms in the non-commodity exporting sector are born regional. Most of them are owned by large, very large, actually, 100 million, 200, 300 million dollar East African groups that are investing in Rwanda with an eye on the Burundi and DRC markets. It's more difficult to export to Burundi and DRC from Kenya, from Tanzania. So they set up shop in Rwanda. The third point is that firms don't always export by choice. It can be by default. So one of the hypotheses in the learning by exporting literature is that exporting is a choice. But does that assumption always hold? Let me give you the example of Pembe Flower in Rwanda. So Pembe Flower is the largest wheat producer in Rwanda. It's a $30 million company. And in 2010, it was the second largest exporter in the manufacturing sector. It exported $2 million worth of wheat bran. Now wheat bran is a residue from the milling industry that is used for animal feed. So usually what a company would do, we would sell it to an animal feed manufacturer, which then produces animal feed. The problem is that the animal feeds processor in Rwanda was bankrupt. And in general, anyway, there is hardly any animal feeds market in Rwanda. So basically, either the firm would just get rid of its residues or export it somewhere, possibly at a loss. So it was exporting to Kenya for the simple reason that there was no market for this in Rwanda. This is the second largest exporter in 2010 in terms of volume. And it so happened to also be one of the most productive firms. So how does that impact your learning by exporting analysis? Is the question. Yes, it's the same. It's a regional group. Point number four, very much in the same vein as what John Sutton was saying today, traders play a very, very important role in exporting in Rwanda. This is basically how exports happen. Traders go to neighboring markets. They go to Burundi. They go to DRC. They talk with distributors. They figure out what the demand is. And they go to the producers in Rwanda. And they say, you know what, guys? I want some beer and I want some wheat flour from you. I need to export it to Burundi and DRC. They're the ones doing all the work. And the reason is because, and this we found from a lot of our interviews while we did this enterprise mapping exercise, the reason is because of trade barriers. It's very, very complicated, inexpensive to export to the DRC. So my question is, if that's really the case, then who is doing the learning? Should we be learning by exporting on a sample of traders? Should we be studying it on a sample of manufacturing firms? Should we be looking at the links between them? So again, it's one point that would be interesting to keep in mind in future analysis of this kind. The fifth point is that we shouldn't forget that even in the formal sector, and I'm talking about big firms, exports can be informal. Now let me give you an example. So Rwanda's informal trade networks are used by large companies to export. And informal exports to the biggest trading partners, which are Burundi and DRC, informal exports are larger than formal exports. 60% of exports to DRC are informal. This is 2012 data from the central bank of Rwanda. 55% of exports to Burundi are informal. Now how does it happen? Let me give you the example of Minimex, which is Rwanda's largest maize flower manufacturer. They are exporters. They export to DRC in pretty large quantities. Basically what happens is Minimex sells in bulk to traders in a market at the border. And then individual traders take one sack each on their back and cross in and out of DRC at no cost. It's good business for them because they make a bit of a markup. It's good business for Minimex because it doesn't have to go through the very difficult procedures of exporting to the DRC, which, by the way, is a market it doesn't know. It doesn't have a trading or exporting department that looks into how to export to a country like DRC. So again, if this is the case, how do we account for that? How do we account for that in the learning by exporting literature? The sixth finding, and I think this is an interesting point also, is that supply is the biggest constraint for firms in a country like Rwanda. Rwanda's land looks, the transport costs are extremely high. And from all our interviews with manufacturing firms who spoke to about 50, 55 of them, supply came out as really the number one issue. Now, basically what the delays and the cost of supplies mean for companies in Rwanda's manufacturing sector is that they have low productivity and it decreases their capacity utilization. So if firms can develop effective international supply networks that are stable, that are diversified, that are based on established relations of trust, then those relationships can have a very big impact on productivity. Not only that, but through their international networks, firms keep up with the international best standards with new technology, et cetera. Now, this, of course, can lead to learning. So my question is, why do we assume that learning on the sales side of things is stronger than learning on the supply side of things? Or why is there not a literature on learning by importing? So yeah, with that I want to leave you. I just want to summarize some of my findings, which is Rwanda are very few exporters in Rwanda. Two, there's very little switching, which makes identifying the learning by exporting difficult. Third, exporting is not always done by choice. It can be by default, because there is no domestic market. And I don't think the case of Pembe flower is an exception. Most of the millers face the same problem in Rwanda, and I'm sure there are other companies as well. Traders are central to the equation. So what does that mean for learning? Point number five, formal firms also export to informal networks. It's quite common. And lastly, supply matters the most. So why not analyze also learning by importing? Thank you.