The container arrives at Dar es Salaam and sits. Goods destined for Zambia, Uganda, and Rwanda spend an average of 25 days at this port \u2014 five days longer than cargo moving to countries with their own coastlines. It is the opening charge on what economists call the landlocked penalty: the systematic additional cost imposed on countries that cannot reach the ocean without crossing someone else’s territory first.<\/p>\n\n\n\n
From Dar es Salaam the route to Lusaka runs 1,800 kilometres overland. Every kilometre of it passes through Tanzania \u2014 its roads, its customs posts, its politics. A Zambian importer pays not just for freight but for the quality of Tanzanian infrastructure, the reliability of Tanzanian border officials, and whatever the current state of bilateral relations happens to be. None of those variables are within Zambia’s control. All of them determine what Zambia pays.<\/p>\n\n\n\n
Luxembourg is 400 kilometres from the nearest ocean port. Its GDP per capita sits at approximately $135,000 \u2014 making it, by most measures, the wealthiest country on earth. It is landlocked. Switzerland is landlocked. Austria is landlocked. And yet for all three, the landlocked penalty \u2014 the systematic cost structure that makes Zambia’s importers pay more for the same goods \u2014 effectively does not exist.<\/p>\n\n\n\n
Same geographic category. Radically different outcomes. The explanation is not geography.<\/p>\n\n\n\n
The world has 32 landlocked developing countries, home to more than 500 million people. Sixteen of them are classified by the United Nations as among the world’s least developed nations. That clustering is not coincidence.<\/p>\n\n\n\n
Alexander Moore’s 2018 structural gravity analysis of trade data from 2005 to 2014, published in the Journal of Quantitative Economics, found that landlocked countries export between 27 and 41 percent less than otherwise comparable coastal countries \u2014 and that among developing countries specifically, the penalty exceeded 40 percent. “Otherwise comparable” is doing real work: the analysis controls for income, distance to trading partners, country size, and a range of other variables. The gap is attributable to the landlocked condition itself.<\/p>\n\n\n\n
The investment picture is bleaker. Landlocked developing countries collectively attract less than one percent of global foreign direct investment flows. Supply chains routed through multiple countries with unpredictable border times are chains that global manufacturers avoid. An assembly plant in a landlocked developing country that depends on transit through two neighbours with variable governance is a liability that simply does not appear in a coastal competitor’s cost model.<\/p>\n\n\n\n
Sachs, Gallup, and Mellinger established in 1999 that temperate zones proximate to navigable water account for just 8 percent of the world’s inhabited land and 23 percent of its population \u2014 yet 53 percent of world GDP. The economic geography of the planet is not neutral. Faye, McArthur, Sachs, and Snow’s 2004 analysis of landlocked developing countries in the Journal of Human Development provided the granular account: the mechanisms through which geographic isolation translates into persistent poverty traps, and why the trap is so difficult to escape without specific structural advantages.<\/p>\n\n\n\n
The landlocked penalty is concentrated almost entirely in developing countries. High-income landlocked countries \u2014 Luxembourg, Switzerland, Austria, Liechtenstein \u2014 show essentially no penalty at all. The geographic condition is identical. The economic outcomes are not. Which means geography is not the variable that matters.<\/p>\n\n\n\n
The geography-vs-institutions debate<\/strong>\n\nJeffrey Sachs argues that geography directly shapes economic outcomes through trade costs, disease burden, and agricultural productivity \u2014 that the map is, quite literally, destiny. Daron Acemoglu and James Robinson argue in Why Nations Fail that geography matters mainly by shaping colonial institutions, which then persist and determine outcomes long after colonialism ends. The two positions imply different policy responses, and the dispute has generated decades of empirical jousting.\n\nThe evidence in this article is consistent with both \u2014 in a way neither camp fully captures. Transit dependency is a geographic fact: landlocked countries cannot reach the ocean without crossing other countries' territory. But whether that fact produces poverty depends almost entirely on institutions \u2014 specifically the institutions of transit neighbours and regional bodies, not just the landlocked country's own. Geography is the constraint. Institutional capacity determines whether the constraint is binding. Switzerland and Chad are both landlocked. What converts one into a development trap and leaves the other unaffected is not geological.<\/code><\/pre>\n<\/div><\/div>\n\n\n\nThe Stack \u2014 How Transit Dependency Actually Works<\/h3>\n\n\n\n
“Coastal countries trade with the world,” Paul Collier wrote in The Bottom Billion<\/em> in 2007. “Landlocked countries trade with their neighbours.” This is the right diagnosis, but it understates the problem. Landlocked countries don’t merely trade with their neighbours \u2014 they depend on their neighbours’ cooperation for access to every other market on earth. That dependence has four layers, and the layers compound.<\/p>\n\n\n\nThe first is physical. Overland freight costs vastly more per tonne-kilometre than sea freight. A cross-border truck in Central Asia pays approximately $700 to transit a single Uzbek border, with roughly a quarter of that figure going to unofficial payments \u2014 bribes, facilitation fees, whatever the local term happens to be. In West Africa, roadblocks staffed by officials who expect payment appear roughly every 30 kilometres and add around ten percent to total haulage costs. A Zambian trucking operator pays roughly 50 percent more for fuel than regional counterparts, because Zambia’s fuel supply chain runs through multiple transit countries, each adding its own margin. These figures come from World Bank analysis of landlocked transport costs \u2014 they are not worst-case illustrations, they are documented averages.<\/p>\n\n\n\n
The second layer is time unpredictability. For export manufacturing, logistics predictability matters as much as logistics cost. Just-in-time production is impossible when transit time varies between three weeks and six, depending on factors the manufacturer cannot control. World Bank data from 2008 documented that goods destined for Uganda, Rwanda, and Burundi spent an average of 25 days at Dar es Salaam, compared to 20 days for coastal-bound cargo. Five days sounds manageable \u2014 until it becomes a range of five to thirty, and the range is determined by other countries’ port management decisions.<\/p>\n\n\n\n
The third layer is political. A landlocked country’s access to global markets depends on the political relationship between itself and every country it transits through. When that relationship deteriorates, trade stops \u2014 not because of anything the landlocked country did, but because its transit neighbour closed the road. Uzbekistan’s border restrictions in the 1990s and 2000s severely disrupted Kyrgyz and Tajik commerce even when those countries’ own policies were functioning. The landlocked country did nothing wrong. The road closed anyway.<\/p>\n\n\n\n
The fourth layer is compound poverty \u2014 the one Collier’s formulation captures best. Landlocked developing countries are almost always surrounded by other developing countries. Their transit neighbours have underfunded roads, underpaid customs officials, and political instability of their own. The penalty is not just about being landlocked. It’s about being landlocked in a poor region. Switzerland’s transit neighbours are Germany, France, Austria, and Italy. Chad’s transit neighbour is Libya.<\/p>\n\n\n\n
These four layers don’t add \u2014 they multiply. Poor infrastructure extends transit times. Unpredictable transit times kill investment. Political volatility in transit countries can shut access entirely. And all of it is amplified by the poverty of the surrounding region, which makes each layer worse.<\/p>\n\n\n\n
The starkest illustration is the doubly landlocked case. Uzbekistan and Liechtenstein are the only two countries in the world that must cross at least two countries to reach any ocean port. Their economic situations could not be more different. Liechtenstein, embedded in Switzerland and effectively operating within the EU single market, carries a per capita income of $207,974 \u2014 the highest of any landlocked country on earth. Uzbekistan, ringed by Central Asian states with their own transit barriers, faces the stacked mechanism twice over. The decisive variable is not the geographic condition. It is the identity of the neighbours within which that condition is embedded.<\/p>\n\n\n\n
The European Dissolve \u2014 When Geography Stops Mattering<\/h3>\n\n\n\n
Luxembourg’s $135,000 GDP per capita requires an explanation, and the right one starts not with Luxembourg’s policies but with what Luxembourg’s trucks do not have to do. They do not stop at customs posts. They do not pay unofficial fees to move between Luxembourg and Antwerp. They are not subject to inspection at national borders that have been, for legal purposes, abolished.<\/p>\n\n\n\n
The EU single market, for a landlocked country, means one thing above all: transit friction has been designed away. Luxembourg is landlocked in the same geographic sense Zambia is landlocked. It is not landlocked in any economically meaningful sense, because the transit problem that defines landlockedness has been removed by fifty years of deliberate institutional construction.<\/p>\n\n\n\n
Switzerland is not an EU member, but it has concluded more than 120 bilateral agreements with the EU, including on land transport, that give it substantially similar transit access. Trucks move between Switzerland, Germany, France, and Italy on essentially the same terms as within the single market. The legal structure differs; the practical result is nearly identical.<\/p>\n\n\n\n
What both cases share is not “good institutions” in the abstract \u2014 that phrase, in development economics, tends to mean nothing specific. What they share is collective institutional quality: they are embedded in a regional integration framework among wealthy, capable states that eliminates transit friction at the border. That framework required neighbours who were willing, economically capable, and stable enough to participate in building it over decades. It is not the product of one country’s good decisions.<\/p>\n\n\n\n
Luxembourg’s wealth also reflects a structural feature of its economy that makes geography secondary: its primary exports are financial instruments, regulatory access, and administrative functions. The European Court of Justice sits in Luxembourg. Major investment funds are domiciled there. Banking and wealth management generate most of its economic output. A port is irrelevant to all of this. The geographic constraint \u2014 always a constraint on physical trade \u2014 dissolves when physical trade is not your primary business.<\/p>\n\n\n\n
Switzerland demonstrates the same logic at larger scale. Roche and Novartis, both headquartered in Basel, anchor a chemical and pharmaceutical complex that accounts for roughly half of Switzerland’s total exports. Financial services in Zurich and Geneva manage assets worth trillions of dollars. MSC, the world’s largest container shipping company, has been headquartered in Geneva since 1978 \u2014 its business is coordinating the movement of physical goods around the world, but the value it creates is informational. It doesn’t need a port. It needs a legal system and a phone line. The absence of a Swiss coastline is operationally irrelevant to a company that owns the ships.<\/p>\n\n\n\n
Austria adds a geographic advantage on top of the institutional solution: the Danube, linked to the Rhine via the Main-Danube Canal, gives it effective river access to both the Black Sea and the North Sea. But this is a bonus layered on EU membership, not a substitute for it. The essential mechanism is political, not topographic.<\/p>\n\n\n\n
The key claim in all three cases: none of them could have achieved this outcome alone, through better domestic policy. The solution required the cooperation of wealthy neighbours willing to build, over half a century, an integration framework that abolished the transit problem. That framework does not exist in sub-Saharan Africa, Central Asia, or South Asia at comparable depth or income level.<\/p>\n\n\n\n
\nThe double landlocked paradox<\/strong>\n\nLiechtenstein \u2014 population 39,000, area 160 square kilometres \u2014 is simultaneously the world's richest landlocked country and one of only two doubly landlocked countries on earth: it must cross through Switzerland before it can cross through any other country to reach the sea. By the logic of transit dependency, it should be trapped twice over.\n\nInstead, it has the highest GDP per capita of any landlocked nation: $207,974 in 2023, according to the World Bank. Liechtenstein operates within a currency union with Switzerland and uses Swiss customs controls at its external frontiers, but accesses the EU single market through its own EEA membership, held since 1995. Switzerland is not in the EEA \u2014 Swiss voters rejected it in a 1992 referendum, and Switzerland instead relies on bilateral agreements with Brussels. Its economy runs on precision manufacturing, financial services, and a corporate registration industry. It is doubly landlocked in the geographic sense. It is not landlocked in any economic sense. The map tells you almost nothing.<\/code><\/pre>\n<\/div><\/div>\n\n\n\nBotswana and the Limits of the Lesson<\/h3>\n\n\n\n
In September 1966, Botswana became independent. It was among the six poorest countries in the world. It had 22 university graduates, seven miles of paved roads, and an average income below $80 per capita. It was landlocked, bordered by apartheid South Africa, Rhodesia before it became Zimbabwe, and South West Africa before it became Namibia \u2014 three territories whose political stability and commercial reliability could not be assumed.<\/p>\n\n\n\n
Within a year, diamond-bearing kimberlite pipes were identified near Orapa. The Orapa mine opened in 1971. Jwaneng, which would become one of the richest diamond mines on earth, was located in 1972. By the time the diamond sector was fully operational, Botswana faced the decision that landlocked resource-rich countries face repeatedly: who captures the rents?<\/p>\n\n\n\n
The answer, in Botswana’s case, was the state. The Minerals Act of 1967 vested mineral rights in the national government rather than allowing tribal or individual capture. The partnership with De Beers \u2014 Debswana, structured as a 50\/50 joint venture \u2014 channelled royalties through the national budget rather than allowing raw wealth to be exported. Acemoglu, Johnson, and Robinson’s 2002 paper “An African Success Story: Botswana” documented the institutional foundation that made this work: pre-colonial kgotla assemblies that constrained chiefly authority, a tradition of inclusive decision-making that survived into the post-independence state. Diamonds provided the rents. Capable institutions ensured those rents funded infrastructure, education, and health rather than private accounts elsewhere.<\/p>\n\n\n\n
Today diamonds account for roughly a quarter of Botswana’s GDP and around 80 percent of export income.<\/p>\n\n\n\n
That concentration is where the story turns uncomfortable. In the second quarter of 2025, Botswana’s diamond output fell 43 percent \u2014 the steepest quarterly decline in the country’s modern mining history. Lab-grown diamonds, once a novelty, now constitute roughly 20 percent of global diamond sales and sell for up to 80 percent less than natural stones, according to reporting in Fortune in late 2025. The World Bank projects the economy to contract for the second consecutive year.<\/p>\n\n\n\n
The Botswana institutions story is true, as far as it goes. The Minerals Act was a sound decision. The Debswana structure was well-designed. The kgotla tradition was real. But the institutions story has always been, at its core, about how well Botswana managed a windfall. It says nothing about whether the windfall itself was replicable.<\/p>\n\n\n\n
It wasn’t. Botswana happened to sit on one of the world’s highest concentrations of gem-quality diamonds in a geological formation that made large-scale extraction economically viable. Zambia had copper. The DRC had coltan, cobalt, and gold. Neither produced the same outcome \u2014 partly for institutional reasons, but also because of commodity price volatility, colonial-era extraction structures, and post-independence political dynamics that Botswana’s particular history happened to avoid. British colonial neglect of Botswana, which left its pre-colonial institutions largely intact, is not a model for replication.<\/p>\n\n\n\n
And even Botswana’s success has a ceiling that rarely gets stated plainly. After fifty years of diamond revenues and genuine institutional quality, the country has not diversified out of resource dependence, has not become a manufacturing exporter, and faces the landlocked constraint in agriculture and industry as acutely as any comparable economy. The diamonds deferred the penalty in sectors where rent could substitute for trade competitiveness. They did not remove the geographic constraint. In manufacturing and agriculture, the constraint has never lifted.<\/p>\n\n\n\n
What Botswana bought was time. And time, as the lab-grown diamond market is now demonstrating, runs out.<\/p>\n\n\n\n
\nZambia: the counterfactual<\/strong>\n\nIn 1966, Zambia's copper reserves were more valuable than Botswana's diamonds. Both countries were landlocked at independence, surrounded by unstable neighbours, with devastated infrastructure and minimal human capital. The divergence since is one of the starkest natural experiments in development economics.\n\nZambia nationalised its copper industry in stages in 1969 and 1973, creating Zambia Consolidated Copper Mines \u2014 which was subsequently mismanaged, underinvested, and exposed to the full force of the 1970s commodity price collapse with no reserve cushion. Economic decline continued through the 1990s, when per capita income was lower than at independence. Recovery has been partial.\n\nAtsushi Iimi's 2006 IMF Working Paper (WP\/06\/138) compared the two countries' resource management trajectories directly. His conclusion: institutional quality in the management of mineral rents, not the rents themselves, determined which country grew and which didn't. The Botswana advantage was not the geology. It was what happened to the money after extraction \u2014 and the institutional capacity that ensured it happened that way.<\/code><\/pre>\n<\/div><\/div>\n\n\n\nWhat Can Actually Be Done<\/h3>\n\n\n\n
The European solution \u2014 deep regional integration among wealthy, capable neighbours \u2014 is the only intervention that genuinely eliminates the transit dependency problem rather than mitigating it at the margin. The analogues in Africa exist on paper. SADC, COMESA, and the African Continental Free Trade Area have produced real progress: simplified transit documents, harmonised customs procedures, some reduction in the number of stamps required to move a truck across a border. These gains are real.<\/p>\n\n\n\n
But the constraint the European case reveals is not the legal framework. It is the income level of the neighbours. Swiss transit works because Germany and France have excellent roads, functioning customs technology, and stable governance. A transit corridor between Zambia and Tanzania is only as good as the weakest point in the chain \u2014 in practice, roads that wash out in the wet season, customs systems that run on underpaid officials with discretionary power, and political relationships that have historically been variable. The problem is not that African integration is worse-designed than European integration. It is that European integration was built on infrastructure and income levels that African partners do not yet have, and which took Europe decades to accumulate before the legal framework had any practical effect.<\/p>\n\n\n\n
The Middle Corridor \u2014 the multimodal route from China through Kazakhstan, across the Caspian to Azerbaijan and Turkey \u2014 represents genuine infrastructure investment in Central Asian transit. Kazakhstan’s oil revenues have funded real logistics improvements, and Russia’s war in Ukraine has redirected significant cargo volume through this alternative. But the corridor primarily benefits resource exporters, not landlocked manufacturing or agricultural economies, and its geopolitical foundations remain fragile.<\/p>\n\n\n\n
Institutional reform \u2014 reducing corruption at borders, streamlining customs documentation, improving trade facilitation \u2014 is demonstrably effective within limits. Research on landlocked development costs finds that institutional improvements can meaningfully reduce the transit penalty. But the structural ceiling on this approach is visible: meaningful reform requires improvements in transit countries, not just the landlocked country itself. A government that transforms its own customs procedures gains nothing if the bottleneck is two borders away, where different officials, different incentives, and different political accountability apply. The collective action problem cannot be solved unilaterally.<\/p>\n\n\n\n
This is the verdict the evidence produces. The forty-four landlocked countries in the world divide into roughly two groups. The first \u2014 the European cases \u2014 operates within a regional integration framework of wealthy, capable neighbours that effectively dissolves the transit problem. The second \u2014 Zambia, Bolivia, Kyrgyzstan, Chad, Mongolia, Nepal, and roughly forty others \u2014 does not. For the second group, the available interventions reduce the penalty at the margin. None of them remove it.<\/p>\n\n\n\n
The landlocked trap for developing countries is not primarily a policy problem that a landlocked government can solve. It is a structural problem created by the income level and institutional quality of a country’s neighbours, the regional infrastructure stock, and the degree of regional economic integration \u2014 all of which are either very slow to change or largely outside the landlocked country’s control.<\/p>\n\n\n\n
That container at Dar es Salaam is not waiting because of Zambia’s policies \u2014 it is waiting because access to the ocean depends on governance decisions made by officials in countries that are not Zambia. The European solution to this problem required fifty years of integration-building among neighbours wealthy enough and willing enough to cooperate; Botswana’s required the geological accident of extraordinary diamond deposits discovered within a year of independence. Neither is available to most of the countries that remain stuck. For those countries \u2014 the majority of the landlocked world \u2014 the question is not what they should do but what the global trading system owes to nations whose structural disadvantage it did not create and has shown no serious interest in fixing.<\/p>\n\n\n\n
That question has no flattering answer.<\/p>\n\n\n\n