Transforming Ethiopia: the only way is up and the only time is now
Ethiopia is not a mineral-rich country. Ethiopia’s greatest endowments are its people, its waters and its fertile lands1. Repeated colonial incursions into Ethiopia’s sphere of influence by the ferenjis – the Europeans that had colonised almost all of the rest of Africa – brought about a realisation among Ethiopia’s late 19th and 20th century’s rulers. If Ethiopia was going to hold on to its vaunted independence, then its military, its technology and its economy would have to catch up with those of the colonialists. Starting from Haile Selassie’s time – the 1940s – Ethiopia started looking for ways to industrialise its economy and achieve structural transformation – economist jargon for the shift of large proportions of labour and other inputs from low-productivity, mostly agricultural uses, to modern high-productivity sectors and activities.
Today, the urgency of structural transformation is even more pressing. Yes, Ethiopia’s economy has grown at an impressive rate in the last 20 years, lifting millions out of poverty. But the pace of growth is slowing and challenges loom large. Millions of youth enter a stagnating job market every year offering fewer and fewer opportunities in smallholder farming – where most Ethiopians work today – as average farm sizes shrink. Ethiopia’s foreign-currency revenues are stagnant and highly concentrated between coffee exports and airline fares2. Without reliable and substantial foreign currency inflows, the country cannot finance the infrastructure investment it needs to flourish. To escape this trap, the current Government seeks to prime structural transformation by nurturing and scaling up a labour-intensive light manufacturing sector focused on exports, particularly of apparel. It has worked before, for example in the cases of China and Viet Nam, and there has been encouraging progress in Ethiopia too. But several hard obstacles keep the sector’s development in check: expensive logistics, low labour productivity and a restrictive foreign currency regime – to name the elephants in the room. The rest of this article unpacks some of the history, opportunities and challenges of structural transformation in Ethiopia, sketches out approaches to tackling some challenges, and points to an accelerating trend that makes the whole endeavour a lot more urgent: automation in light manufacturing.
From Stitches to Riches
Successive governments of Ethiopia have left different marks on the country’s economic geography, in their attempt to trigger widespread industrialisation. The textile mill built by the fascist Italian occupation in Dire Dawa, the imperial sugar mills of Wonji, Shoa and Metehara, fueled by the foreign investment that Emperor Haile Selassie had successfully wooed; the quasi-Communist Derg’s nationalised industries with the Soviet-bloc-funded tractor assembly plant of Adama; and finally, the Hawassa Industrial Park, embodiment of the industrial strategy of the incumbent government, the Ethiopian People’s Revolutionary Democratic Front, which swept the Derg out of power in 1991 and has been ruling the country since.
Today’s Wonji-Shoa sugar factory, which has replaced the obsolescing sugar mills of Wonji and Shoa. Photo Credit: Ezega News.
The Ethiopian economy is undoubtedly sizzling with potential: IMF forecasts put Ethiopia’s real GDP growth at 8.5% in the next 2 yearsii, the highest in Sub-Saharan Africa, and the country is routinely ranked among the top-5 most attractive destinations for FDI in Africa. Its record in reducing consumption poverty is impressive, from almost 56% in 2000, to just under 27% in 2016. But Ethiopia faces a tall order in pivoting the export-economy towards manufactures. The Government has rightly been touting underexploited opportunities in different embryonic value chains, from old time favourites sugar cane and cotton to pharmaceuticals and – it was once tweeted – business process outsourcing. Thus far, there has only been a lot of engine-revving: manufacturing still accounts for just about 5% of GDP.
Like their Chinese and Vietnamese predecessors, Ethiopian policy-makers are betting that light export-focused manufacturing will be the locomotive in the country’s journey towards becoming a modern industrialised society. The specific sectors designed to provide the draft here are apparel as well as – to a lesser extent – leather product exports. The country boasts vast tracts of agricultural land suitable for cotton production and Ethiopia’s cattle herds are Africa’s most numerous. Using manufacturing demand to fuel agricultural development in the production of inputs, such as skins and export-grade cotton is a great idea for at least two reasons. It saves forex-and-time-costly imports of inputs and it creates jobs and provides income upstream, in agriculture, where the poverty-reducing-impacts are highest. Sadly, Ethiopia’s recent history is rife with failed attempts at doing just this. The first textile plant in the country was set up by the fascist Italian occupants in Dire Dawa in 1939. Successive governments have tried in different ways to overhaul both the textile and leather supply chains. But spells of political stability have been rare since the 1960s, and the political economy of land in areas of high agricultural potential and lowland grazing regions has perhaps been the most consistent source of conflict in the country: not the most fertile ground for investment in inputs to manufacturing. So insufficient investment throughout the value chain, from cotton seeds and animal veterinary care to industrial workforce skills, have so far stifled the dream of Ethiopia stitching its way out of dependency on agricultural commodity exports.
Konso Tribal Land, Ethiopia. Photo Credit: Flickr User Rod Waddington.
The current government has taken an ostensibly modern tack to developing these darling industries, inspired by success in China’s Special Economic Zones and Vietnam’s Industrial Parks and Export Processing Zones. The idea is to aggressively crowd international private sector investment and expertise by dotting the country with light-manufacturing industrial parks5. Picture these parks as physical and regulatory bubbles, where the government provides high-quality infrastructure (roads, power, water, wastewater management systems, high-quality shed space etc.), attractive land lease prices, tax holidays, simplified procedures and regulatory one-stop-shops to shield investors from the difficulty of doing business in the ordinary Ethiopian economy. Low wage rates and energy prices prevailing in the country represent powerful draws for foreign manufacturers.
The Hawassa Industrial Park. Photo Credit: YouTube User Rick Steves’ Europe.
There are now six active government-sponsored industrial parks in operation and up to nine more are said to be in the pipeline. So far, all active parks focus on the textile and garments industry and are exclusively dedicated to export markets. The flagship park, Hawassa, sent out its first significant clothing exports shipment in 2016. In mid-2018, export revenues from the first two parks overtook those going to the rest of the Ethiopian garments sector, showing that these enclaves really do have the potential to sustain export growth at scale. But the total sector’s export earnings in FY 2018 were less than USD 100 million, i.e. less than a twelfth of what was envisaged in the government’s Growth and Transformation Plan II (GTP II)1. It remains to be seen how much four new textile and apparel industrial parks inaugurated in 2019 will add to that number.
Wait for it…
While the GTP II targets proved too optimistic, it is too early to judge whether the government’s strategy is yielding fruit. Ambitious targets are needed to give direction and impetus to an ambitious project: such is the structural transformation of Ethiopia’s economy. There are positive signs and real reasons for hope. Consider the sustained – if not stellar – growth in leather product and clothing exports seen in the last two years. The recent, unplanned appearance of a mobile-phone export sector also holds promise: electronics was certainly a driving sector in both China and Viet Nam’s export-driven booms. The sector has been growing fast thanks to investment by Chinese affordable mobile-phone manufacturer Transsion. The company aims to cater to both the domestic market and the burgeoning demand for affordable smartphones in Africa, leveraging Ethiopian Airlines’ continental footprint for shipments.
In both China and Viet Nam, it was only six or seven years after the inauguration of market-opening reforms that clothing and footwear began consistently driving more than 10% of export growth. In both cases, another six years later or so, the electronics exports sector would also reach a similar level. Ethiopia is only three years into its latest stab at industrialisation, is at a far lower stage of development, and faces rather different challenges than the two Asian economies did in the 1980s. It is a landlocked country and lacks a significant industrial tradition, whether colonial or collectivist, and so has to start from scratch the process of coaxing a largely agrarian workforce into very different jobs and lifestyles. It would be unfair and unrealistic to require that Ethiopia follow the same breakneck growth trajectory of its two Asian models.
Source: Own calculations on Atlas of Economic Complexity data. Note data for 1998 is missing in the source dataset, and so interpolated in both charts.
International garment firms are flocking to industrial parks, and the previous government’s overtures have generated real buzz in the sector. Yet, operations have run into several hard obstacles, that risk undermining the momentum built up so far. We’ll look at two structural ones and one policy-based: the cost and timing both of outbound and inbound logistics through Djibouti, the low factory productivity of an agrarian workforce, and the shortage of foreign currency to pay for imported inputs.
Jammed in Djibouti
Essentially all of Ethiopia’s seaborne trade goes through the port of Djibouti. The cost and sloth of Ethiopian outbound and, especially, inbound logistics through Djibouti is a well-worn subject in Ethiopian trade policy circles. Many costs and constraints feed in to the issue. To illustrate one: clearing a container through the Port of Djibouti’s container terminal costs between two and four times as much in fees as regional comparator ports like Mombasa in Kenya or Damietta in Egypt. But the currently viable alternatives of Port Sudan and Mombasa are simply too far to be competitive.
Port of Djibouti. Photo Credit: Wikipedia.
Ethiopia has been working to reduce its reliance on the Port of Djibouti as the sole inlet and outlet for seaborne trade. Abiy Ahmed, the peace-building and market-minded new Prime Minister, has ended a state of war with Eritrea that lasted for 20 years. This came with a declaration of friendship, the opening of land and air borders, and a promise to rehabilitate the historical trade route between Addis Ababa and the Eritrean port of Asseb. The previous administration had also acquired a 19% stake in Berbera port, in neighbouring Somaliland. But progress in linking to Asseb has stalled. Both the port and the Asseb-Addis Ababa road link require investment to become a viable alternative to the Djibouti corridor. Worryingly, transit through the Ethio-Eritrean border post on the road to Asseb was halted by the Eritrean Government in April this year. All land borders between Ethiopia and Eritrea are, once again, closed.
On the domestic side, Abiy has committed Ethiopia to partial liberalisation of its external logistics sector, until now dominated by a very profitable state monopoly. If processing times and costs along the Djibouti corridor stay as high as they are, the government should push harder to develop these alternative routes. This will put pressure on Djibouti to offer Ethiopian freight forwarders lower rates or risk losing business. The Djiboutian government will want to avoid the latter: direct revenues from the port make up 20-25% of its tax revenue, and Ethiopia is the source of 85% of traffic through its port3. Letting go of the Government’s grip on international freight should be done carefully, ensuring the sector remains compact in negotiating down Djibouti’s rates: it is easier to bargain with your single, irreplaceable vendor if you’re their single, irreplaceable customer.
With 2 million Ethiopians entering the workforce every year for the next decade6, the number of children per mother falling slowly but still very high, at around 4.6, and the average farmland inheritance prospects shrinking in many places to the limits of sustainability, the Ethiopian youth urgently need the government’s industrial strategy to work. But the youth will have to be willing and able to work in factories – and making the transition from traditional economic activities into capitalist factory labour, often in tough urban working and living conditions, comes with new forms of hardship and social disruption. The rate of churn (hiring and firing) experienced by industrial park firms was reportedly very high in 2018, as difficulties in finding housing near plants, low salaries and long working hours drove wave after wave of recruits away. This of course harms productivity. Some firms allege factory output per worker is as low as half of what their factories achieve in neighbouring Kenya, linking it to the difficulty in retaining workers as well as high absence rates.
Ethiopian factory workers making shoes at the Huajian shoe factory in the Eastern Industrial Zone, Ethiopia. Photo Credit: Flickr User UNIDO.
Both the churn rate and productivity gap were reported to be coming down in later 2018, however, in a sign that workforce training and search efforts are yielding fruit. Momentum ought to be sustained, predictable obstacles like provision of housing for workers should be addressed promptly, and the upskilling of the workforce and development of an Ethiopian middle and top management in industry should be pushed. These jobs are fewer but crucial: they can increase productivity by increasing the motivation of Ethiopian workers, and allow technological transfer into the domestic sector to take root in the know-how of the country’s workforce too.
The foreign-currency crunch is perhaps the toughest economic challenge facing Ethiopia now. The policy of the National Bank of Ethiopia (NBE) – Ethiopia’s central bank – is to let the Ethiopian birr slowly lose value against the dollar. Keeping a relatively stable peg to the dollar has its advantages, by offering currency stability to traders and investors and keeping the foreign-currency-denominated debt burden in check. But because the NBE also wants to sell its own monetary policy and control domestic prices it also has to enforce strict controls on capital flows. The economic law governing this dynamic has, fittingly, one of the more sinister monikers in the field: the impossible trinity5.
The result is that the birr is chronically and heavily overvalued, meaning that the market is signalling that buying the birr should be much cheaper. The black market has been buying dollars for 30% more birr than the going official rate. Import-exporters reportedly sell coffee at a loss on international markets, just to get their hands on dollars to finance their much more lucrative import business. The country is forever cash-strapped, with hard foreign currency in the financial system’s reserves worth the equivalent of less than two months of Ethiopia’s expenditure on imports, a worryingly small buffer by international standards. A complex forex rationing system ties down the dollars already in the country and prioritises their use according to GTP II’s goals. In the context of forex shortages, this has a sensible motivation given the large basic needs of the country that are met by imports (wheat, medicines and cooking oil, to name a few). But it also means that clothing exporters are legally required to relinquish 70% of their foreign exchange earnings from exports to the NBE, and so cannot use them to buy inputs or capital goods from abroad to ramp up production7. This would be a great stimulus for the local industry if only there were a sufficiently ample and good domestic supply of inputs like cotton, synthetic fabrics and accessories, but that is not the case yet. Almost all of it needs to be imported8 – of course via Djibouti: recall how painful that can be.
Some form of drastic and permanent downward realignment of the currency’s price is necessary: it is undermining the core of the government’s industrial strategy in too many ways. But any devaluation, however well-orchestrated, automatically increases the government’s forex-denominated debt and the import bill, and can have a knock-on effect on inflation, including in food items, if not appropriately sterilised through restrictive monetary policy, which heaps up its own negative economic consequences. This is why the government is reluctant to devalue. It will generate certain short-term pain in the face of uncertain medium-term gain. However, Ethiopia is already heavily invested in its own modernisation, and so must see the gamble through. The ruling party, which has been in power for the last 28 years, can afford the long-termist outlook required to do so.
What’s that Metallic Whirring Sound…?
When it comes to the long term, though, time is not on Ethiopia’s side. The economic opportunity and imperative of finding work for the millions of Ethiopian youth that enter the workforce every year is here, now. But there is another looming threat, one that Ethiopia has no control over: the rise of the robots. Advances in the application of artificial intelligence (AI) to automatised manufacturing threaten the viability of the light-manufacturing-driven industrialisation model, placing a ticking timer on Ethiopia’s hopes of riding that model to prosperity.
A paper published by Carl Benedikt Frey and Michael Osborne back in 20139 rocked this particular boat hard in the developed world. They built a dataset that characterised jobs based on the type of tasks they involve and then used the opinions of experts in AI to determine whether each subset of 70 jobs was “computerisable” or not. They used this to model the risk of automation facing a broader set of 702 jobs, covering 97% of the US workforce10. Their model predicted that 47% of jobs in the US are exposed to automation11. A salvo of papers followed disputing that headline. But what interests us here is the fate of one particular job identified by SOC code 51-6031 of the US Bureau of Labor Statistics: the sewing machine operator. This code encompasses a large chunk of jobs created by any textile industry. Frey and Osborne’s AI-experts give it a 89% probability of being automated, putting it in the top 30% of the most automatable jobs. Several more jobs in the textile industry get an even higher rating of automation risk. This of course does not tell us how long it will take for automation to be able to effectively accomplish all the tasks of these operators, or when it will be economically viable to do so. Right now, sewing even a relatively simple garment requires far too much finesse of movement, adaptability to fabric’s complex physics, visual intelligence, and coordinated, precision movements for AI-powered machines to tackle effectively. But progress is incredibly fast these days. AI is already capable of sewing t-shirts economically12, and is making its first inroads into the quality-control department of garment factories13. The question is whether Ethiopia’s competitive edge in textiles, which is grounded in the industry’s intense need for labour and Ethiopia’s low labour costs, may still qualify as a launchpad for structural transformation in a dozen years or so. China, Viet Nam and Bangladesh are still milking their hard-earned primacy in garment exports for jobs and forex, two or three decades after orchestrating their great pivot toward these sectors. If they don’t move fast, Ethiopia and other later industrialisers may not have the same luxury.
Ethiopia is fitfully gearing itself up for economic transformation, as the new administration sizes up the dizzying array of challenges it faces. Some are time-honoured headaches for poor countries, like the forex shortage, the seismic lifestyle shifts required of an industrialising workforce, and the breaking down of barriers to international trade. Some, like impending automation, are more distant and unfamiliar, but no less real. Still, others look like pure cosmic irony. Just as the government begins to uncurl its iron fist and make strides toward domestic democratic discourse, ethnic strife tugs at the country’s unity, and high-profile assassinations and a failed regional coup attempt to rock the peace. Even as PM Abiy preaches and practices regional peace, violent conflict rages on across the borders with South Sudan, Sudan and Somalia. In the latter two countries, bloodshed is on the rise. Though initial euphoria for Abiy’s government has died down, hopes remain at least as high as the obstacles. Ethiopians love sayings and the one I hear the most these days is: linega sil yich’ellimal. It is darkest before the dawn. This is, of course, untrue in astronomical fact, but the political history of Ethiopia is rife with both failures due to denied facts and successes due to defied expectations. Observers and inhabitants of the region will be watching, eager to find out which of the two Ethiopia’s economic transformation turns out to be.
Rero Robots. Photo Credit: Wikimedia Commons.
 Yes, there are gold deposits, and the biblical account of the Queen of Sheba’s visit to King Solomon, which is often believed to have been Ethiopian, has her bringing him about 4 tonnes of gold in gifts. But, the most quantitatively impressive gift she brought was a never-before-seen payload of spices. Solomon, for his part, was far more interested in her than in her riches, according to the Ethiopian account of the legend, the K’bre Negest. The GTP II is written in terms of Ethiopian Fiscal Years, which run from July to June. GTP II’s horizon is better stated as FY 2015/16-2019/20, but simplify notation here and in what follows, referring only to the second Gregorian-calendar year covered by the fiscal year. That means that when we say FY 2016 we mean FY 2015/16, etc.
 Ethiopian Airlines and coffee account respectively for roughly 3 and 1 in every 10 dollars that the country makes from exports. Following a 15% devaluation of the national currency, the birr, against the dollar in October 2017, FY 2018 saw increasingly expensive food items driving overall inflation to a peak of 15% in January 2018. (World Bank 2019, Seventh Economic Update). These days, it is even higher.
 World Bank, 2013, ”Transport and logisitics in Djibouti: contribution to job creation and economic diversification” The latest estimate by the World Bank puts GDP growth at 7.7% for the 2018 fiscal year. http://documents.worldbank.org/curated/en/432421554200542956/pdf/Special-Topic-Poverty-and-Household-Welfare-in-Ethiopia-2011-2016.pdf
 CSA, 2016, Ethiopia Demographic and Health Survey, p.78, Figure 5.1. Exports of manufactured goods should’ve accounted for 1.2 and 1.6% of GDP in 2016 and then 2017: the actual figure stood instead at 0.5% across the two years, the same as in the GTP I 5-year period. (See. Arkebe (2018) “The Structure and Performance of the Ethiopian Manufacturing Sector”)
 The impossible trinity is the impossibility of keeping exchange rates fixed, while pursuing independent monetary policy and letting international capital flow in and out of the country freely. The Ethiopian Government is also working on the development of agro-industrial parks, focused more on food and drink products emerging from agricultural value chains. There are also private industrial parks, mostly Chinese-owned and focusing largely on leather products and construction. In what follows, Industrial parks should be intended to exclude agro-industrial parks and private industrial parks.
 World Bank, 2019, “Ethiopia Economic Update 7. Special topic: Poverty & household welfare in Ethiopia, 2011-16”
 Directive No. FXD/48/2017 ‘Directive for Amendment of Retention and Utilization of Export Earnings and Inward Remittances’ (Government of Ethiopia, 2017) https://www.nbe.gov.et/pdf/directives/forex/Retention%20Directives.pdf
 This leaves clothes’ exporters three options, none of them attractive: financing imports with working capital forex injections from abroad and then wait for the system to let them repatriate forex profits (effectively investing massively and long-term in the birr, a steadily depreciating currency); delaying production until forex is made available by the rationing system and imports get through Djibouti; financing imports through an NBE-approved supplier’s credit scheme in foreign currency, so paying interest on their purchase of raw materials for as long as it takes to assemble them into garments and export them.
 Frey and Osborne, “The Future of Employment: How Susceptible Are Jobs to Computerisation?”
 Not without irony, Frey and Osborne used an AI-algorithm to do some of the modelling.
 The number is 85% for Ethiopia, according to the World Bank (2016 World Development Report), due to the high automation potential in agriculture.
 See SoftWear Automation’s LOWRY robot (e.g. at https://qz.com/1064679/a-new-t-shirt-sewing-robot-can-make-as-many-shirts-per-hour-as-17-factory-workers/
 See e.g. PolyU’s WiseEye algorithm. https://www.innovationintextiles.com/new-aipowered-system-to-automate-quality-control-process/
Alfredo Manfredini Böhm is an economic policy professional. He has advised the UK and Ethiopian government broadly on economic policy, including in the domains of trade policy, transport and logistics, energy, social protection and labour. Alfredo holds a Masters in Economic Research from the University of Cambridge and a degree in Philosophy from the University of Pavia. Ethiopian economic development, moral philosophy, data science and languages rank highest among his favourite topics.