Guide to the Real Economy · Home

8  Chapter 6 — Territory, Agglomeration, and Global Value Chains

Spatial economics and the geography of production

“The secrets of industry are in the air.” — Alfred Marshall, Principles of Economics (1890)

“Ask where a product is made and you ask the wrong question. Ask instead who captures the value.” — after Gary Gereffi

The guiding thread of this chapter
  1. Economic activity concentrates because concentration is productive. Marshall’s trinity of shared inputs, labour market pooling, and knowledge spillovers gives firms in the same place advantages that isolated firms cannot replicate. The question is not whether agglomeration is beneficial but whether its costs (congestion, land rent, inequality) ultimately outrun its gains.
  2. The New Economic Geography showed that concentration can emerge from nothing. Krugman’s model (increasing returns, transport costs, mobile factors) generates core-periphery structures without any natural advantage. The corollary: geography is partially arbitrary, and active policy can, in principle, shift the centre of gravity.
  3. Myrdal’s cumulative causation corrects the optimism. Left to themselves, markets do not equalise regions. They diverge. Backwash effects, specifically the metropolis absorbing capital, youth, and talent from the periphery, dominate spread effects in poor, weakly integrated economies. Morocco’s coastal macrocephaly is not an accident of geography; it is a product of colonial construction and cumulative reinforcement.
  4. Clusters are not agglomerations: they are systems. A cluster adds to co-location the organised collective learning, tacit knowledge spillovers, social embeddedness, and input-output linkages that turn proximity into productivity. Not every industrial park is a cluster. The test is Hirschman’s: what happens when the lead firm leaves?
  5. Global value chains fragment production and concentrate power. The smile curve shows that value accrues at the ends (design, brands, after-sales) and not in the assembly trough. Inserting into a chain is not the same as capturing value from it. Morocco assembles; it is beginning, unevenly, to capture.
  6. Morocco is a semi-periphery in motion. In Wallerstein’s vocabulary: too industrialised and too integrated to be a pure commodity periphery, not yet autonomous enough in technology and brands to approach the centre. The window of nearshoring and renewable energy offers a repositioning opportunity, provided absorbed capability and not just assembly capacity is built.

The Map

The preceding chapters built the microeconomics of firms and markets in abstract space. This chapter gives that space back its geography. Why does half of Morocco’s GDP concentrate along a thin coastal strip between Kénitra and El Jadida? Why do automotive suppliers cluster around Tanger-Med and aeronautics firms around Casablanca-Nouaceur? Why does a chef trained in a Marrakech kitchen stay in Marrakech even when she changes restaurant? And why, despite all of this concentration, do the workers in Morocco’s export assembly zones capture so little of the value they create?

This chapter addresses these questions in three ascending movements. First, the economics of why activity concentrates: agglomeration forces from Marshall through Krugman, and the heterodox counterpoint of Myrdal and Kaldor. Second, the meso-organisation of clusters and industrial districts: how firms organise collectively in a territory, and what separates genuine production systems from mere enclave assemblies. Third, the architecture of global value chains and Morocco’s position within them: the smile curve, governance structures, upgrading strategies, and the long shadow of the structuralist critique.


9 Part I — Why Activity Concentrates: From Marshall to Krugman and Myrdal

9.1 The founding intuition: land rent and bid-rent curves

The first economist to take space seriously was a German landowner. In The Isolated State (1826), Johann Heinrich von Thünen imagined a single city in the centre of a uniform fertile plain with no rivers or roads, and asked how agricultural activities would distribute themselves around it. His answer founds all spatial economics.

Each farmer arbitrates between the price obtained in the city and the cost of transporting production there. Heavy, perishable, or costly-to-move goods, fresh vegetables, dairy, must be produced close to the centre. Goods that tolerate a long journey, grain, timber, can be produced further out. The result is a set of concentric rings determined by a simple logic: bid rent, the maximum rent a farmer can offer for land at a given distance while still earning a normal profit. The bid-rent curve for market gardening falls steeply with distance (because transport costs per kilogram are high); the bid-rent curve for cereals falls more gently. At each point in space, the activity willing to pay the highest rent wins the land. Rings emerge from competition for soil, not from central planning.

Von Thünen’s mechanism is the skeleton of all that follows: the value of a location depends on what it economises in transport and what markets it accesses. This logic, now expressed in terms of market potential and freight costs rather than horse-drawn carts, still determines the spatial distribution of Morocco’s agriculture (intensive export horticulture near Agadir, rainfed cereals in the interior plains) and the location premium for warehousing in Tanger-Med’s industrial zone relative to inland sites.

The market potential concept, formalised by Chauncy Harris in 1954 and embedded in the gravity models of trade, extends this logic to industrial location. A location’s attractiveness depends on the sum of markets it can reach, weighted by their size and the inverse of their distance. Morocco’s most fundamental geographic asset is proximity to the European Union, the world’s largest single market for manufactured goods, at a distance that allows just-in-time automotive supply. This market potential advantage, not any domestic comparative cost advantage, is the primary explanation for why the Renault-Tanger and Stellantis-Kénitra investments materialised in Morocco rather than in cheaper labour markets further away. The distance premium is a gravity effect: proximity reduces transport costs, enables frequent shipments and lower inventory, and allows quality supervision by European engineers who can visit rather than video conference. Morocco’s industrial strategy has been, in essence, a strategy of monetising market potential: converting geographic proximity into logistical and contractual relationships that generate durable productive coupling.

9.2 Marshall’s trinity: the three agglomeration forces

Alfred Marshall, describing the industrial concentrations of 19th-century England, Sheffield cutlery, Lancashire textiles, identified three mechanisms that make firms productive by being near each other, mechanisms that Gilles Duranton and Diego Puga formalised a century later as sharing, matching, and learning.

Sharing is the first mechanism. A firm in an industrial district shares with its neighbours infrastructure, specialised services, and thick input markets that none of them could individually afford. A specialised machine tool maintenance firm can only survive if there are enough users nearby. A port terminal is viable only if traffic is dense enough. A specialised legal firm serving the automotive supply chain only justifies its existence in a region with a substantial automotive industry. All of these services are accessible to a firm in the cluster but inaccessible to an identical firm operating in isolation. The firm inside the cluster uses them; the firm outside the cluster must do without or travel for them at high cost.

The secrets of the industry are in the air

Marshall’s most celebrated phrase, “the secrets of industry are in the air”, describes a mechanism observable in any specialised district. A young machinist in Morocco’s aeronautics cluster around Nouaceur learns from conversations at lunch, by watching a colleague solve a problem he has not yet encountered, by observing how a senior technician handles a tolerance deviation. This knowledge circulates without being bought or sold. It is not in a manual. It passes between bodies in the same physical space. A machinist who moves to a different city, however skilled, loses access to this invisible infrastructure of shared knowledge. The cluster is not only a place to work; it is a learning environment that produces knowledge faster than any of its members could generate individually. This is what Marshall heard in the hum of the Sheffield steel district, and what a Moroccan industrial park will generate if, and only if, genuine worker mobility and interaction are allowed to develop within it.

Matching is the second mechanism. A thick local labour market allows workers to find jobs well matched to their specific skills, and employers to find candidates closely fitted to their precise needs. A specialist in automotive wiring harness assembly has limited options if there is only one assembler in her city; she has multiple options and real bargaining power if the regional labour market contains ten. For the firm, a thick labour pool reduces recruitment time and the probability of hiring a misfit. For the worker, multiple employers reduce the risk of job loss and enable career progression without geographic uprooting. Both sides gain from the density: the district produces better matches than either isolated labour market or single large employer could.

Learning is the third and most consequential mechanism. Knowledge circulates in clusters in ways it cannot in dispersed markets. Worker mobility carries tacit knowledge from firm to firm. Informal contacts between engineers at competing firms generate information exchanges that formal technology transfer arrangements cannot replicate. The knowledge does not have to be purchased or contracted; it flows through the social fabric of a specialised community. Marshall called this knowledge “in the air”, Duranton and Puga formalise it as the learning externality, but the underlying observation is the same. Doubling the density of skilled workers in a location does not just double the labour supply; it more than doubles the rate of knowledge circulation.

The chef who changes restaurant but not city

A talented chef trained in Marrakech’s best kitchens changes restaurants every few years. Each time, she carries with her techniques, flavour memories, presentation ideas, and kitchen discipline that she redistributes in her new establishment. Her former colleagues do the same across the city. Within a decade, a body of culinary knowledge born in one kitchen has diffused through an entire restaurant community. None of this knowledge passed through a manual or a patent. It travelled in hands and heads through the local mobility of people. This is territorial learning made visible. And it explains why a chef who emigrates alone to a city without a culinary tradition loses much of what made her exceptional: not her individual skill, but the atmospheric infrastructure that sharpened it. The buzz of a cluster is not metaphor, it is a knowledge production mechanism that depends on proximity.

These three mechanisms generate a quantifiable productivity premium. Econometric estimates consistently find that doubling employment density in a city raises productivity by 3–8%. This may seem modest in annual terms. Cumulated over decades, it accounts for a substantial share of the income gap between metropolis and interior. Casablanca’s productivity advantage over comparable-sized cities in Morocco’s interior is not entirely explainable by the quality of individual workers it attracts. A significant part is the agglomeration itself, the density of interaction, the thick markets, the knowledge spillovers that no amount of land and capital can replicate outside the cluster.

9.3 Dispersion forces and the limits of concentration

Agglomeration is not unlimited. Against the centripetal forces pull centrifugal ones: land rent rises as the city grows, pricing out less productive activities and households; congestion mounts, traffic, network saturation, administrative bottlenecks; pollution and environmental degradation erode quality of life; and in semi-arid countries, pressure on scarce water resources ultimately constrains how large any concentration can grow sustainably.

Casablanca illustrates the two faces of this medal. Its density generates the productivity premium that makes it Morocco’s economic capital. The same density generates land prices that have expelled lower-income households to a diffuse peri-urban ring of remote new towns (Tamesna, Tamansourt), long commute times that absorb several hours of every working day, water stress that will intensify as climate change reduces rainfall in the Atlantic coast watershed, and acute traffic congestion that already imposes measurable economic costs.

The critical Moroccan specificity is water. The coastal urban conurbation from Kénitra to Casablanca concentrates millions of inhabitants and water-intensive industries on a strip where aquifer recharge is limited and increasingly unreliable under climate variability. Rather than self-correcting through market signals, as agglomeration theory would predict if land and water were properly priced, the concentration is maintained by massive hydraulic investment: inter-basin water transfers, desalination plants at Agadir and planned for Casablanca, groundwater extraction at rates that exceed recharge. The constraint does not disperse the activity; it displaces the problem from geography to public finance and ecological futures.

9.4 Krugman’s New Economic Geography: concentration from nothing

Why does Morocco’s economic activity concentrate in one part of the territory rather than another? Not primarily because of natural endowments, Casablanca has no coal, no exceptional port geography that Safi or Nador could not have provided, no mineral resource. It concentrated because of historical contingencies amplified by economic mechanisms that Paul Krugman formalised in the New Economic Geography.

Krugman’s model (1991), developed with Masahisa Fujita and Anthony Venables, combines three ingredients: increasing returns at the firm level (firms are more efficient at larger scale), transport costs between regions, and the mobility of some factors (workers). From this combination, a centre-periphery structure emerges spontaneously, without any initial natural advantage being required. The circular causation is tight. Firms want to locate where the market is large, to save transport costs and exploit increasing returns. But the market is large where many workers and consumers are. Workers want to locate where many firms are, because they find employment, higher wages, and greater product variety there. The home market effect: a slightly larger region attracts more than proportionally the activities with increasing returns, because they can serve the large local market without transport cost while still exporting to smaller markets. The price index effect: more local producers mean lower consumer prices (less transport cost), which raises real wages in the region, attracting more workers. Once a small initial advantage tilts the balance toward a region, both effects self-reinforce and concentration accelerates.

The model produces a counter-intuitive result about transport costs: the bell-shaped curve. When transport costs are very high, each region must produce locally what it consumes, activity is dispersed. When transport costs are very low, location becomes irrelevant, activity can again disperse. It is in the intermediate range, when transport is costly enough to make proximity to the large market valuable but cheap enough to allow export to smaller markets, that concentration is strongest. The policy implication is not obvious: improving infrastructure, highways, high-speed rail, does not automatically disperse activity toward inland regions. It may instead amplify the metropolis by opening regional markets to it. The tunnel effect, infrastructure that connects without irrigating, is a real risk in Morocco, where highway and TGV investments link Casablanca-Rabat to the national network without necessarily generating new productive activity in the interior rather than accelerating migration toward the coast.

9.5 Myrdal and cumulative causation: the heterodox correction

The New Economic Geography describes a mechanism. It is largely silent about whether the outcome is desirable or correctable. The heterodox tradition, led by Gunnar Myrdal and Nicholas Kaldor, starts from the same empirical observation, concentration is cumulative, but draws darker normative and political conclusions.

Myrdal’s circular cumulative causation (1957) describes a self-reinforcing dynamic of both success and failure. A region that attracts an enterprise attracts workers, whose spending supports commerce, whose taxes finance schools and roads, which attract new enterprises. Success begets success. Symmetrically, a region losing activities loses its youth, its shops close, its public services decline, which discourages further investment. Failure begets failure. Far from the neoclassical convergence, where poor regions should attract capital from rich ones because wages are lower, Myrdal sees self-sustaining divergence.

Myrdal distinguishes backwash effects, the metropolis pulling capital, skills, and dynamism from the periphery, from spread effects, prosperity creating demand for peripheral products and diffusing technical progress outward. His crucial empirical claim: in poor, weakly integrated economies, backwash effects systematically dominate spread effects. The market, left to itself, worsens territorial inequality rather than correcting it.

Nicholas Kaldor adds the productive foundation: his extension of Verdoorn’s Law shows that industrial activity exhibits increasing returns to scale over time through learning-by-doing. A region that gains an early lead in manufacturing sees its productivity grow faster, reinforcing its competitive advantage, drawing more production, and accelerating productivity gains further. The type of productive specialisation determines whether a territory accumulates cumulatively or stagnates. Manufacturing cumulates; primary production and agriculture typically do not. The free market, therefore, and even free trade, tend to lock in and amplify initial differences in productive structure rather than equalising them across regions or nations.

Morocco’s coastal macrocephaly is constructed, not natural

Morocco’s economic macrocephaly, half of GDP on a thin coastal strip, is presented in some analyses as an expression of geographic efficiency: the coast is simply more productive. This reading neglects the historical construction behind the current distribution. The colonial distinction between Maroc utile and Maroc inutile, useful Morocco and useless Morocco, was not a geographic discovery. It was a political choice: to invest in the coast, extract from the interior, and systematically underdevelop inland territories. The railways connected coastal ports to interior mineral and agricultural zones for export, not to integrate a national economy. The cities that grew were those the colonial administration chose to develop, including Casablanca, which became the administrative and commercial capital partly through infrastructure investment that was a deliberate reorientation of the pre-colonial urban hierarchy.

Post-independence development policy largely reproduced this geography, concentrating public investment and industrial infrastructure on the coast. Cumulative causation then did the rest: the coast attracted talent because it had jobs; it had jobs because it had infrastructure; it had infrastructure because investment followed where jobs were. Territorial inequality in Morocco is not the efficient outcome of comparative advantage. It is the sedimented product of colonial construction amplified by decades of cumulative reinforcement. Understanding this is the precondition for any policy that seeks to reduce it, and for recognising that reducing it will require deliberately fighting against the centripetal forces the analysis identifies, not merely improving the general business climate.

Andrés Rodríguez-Pose has given this analysis a contemporary political dimension: the territorial concentration that maximises aggregate efficiency also produces what he calls “the revenge of the places that don’t matter.” Regions left behind feed a resentment that translates into electoral behaviour, the Brexit vote, populist surges in de-industrialised regions of France and the United States, that ultimately threatens the stability of the institutions enabling efficiency in the first place. Morocco’s territorial inequalities, if left unaddressed, carry the same political risk: a coastal modernity that a majority of the population experiences only from the outside breeds disconnection and eventually instability.


10 Part II — Clusters and Industrial Districts: When Proximity Becomes System

10.1 The industrial district: from Marshall to Becattini’s Third Italy

Proximity is necessary for agglomeration economies to operate. It is not sufficient for them to produce genuine collective learning and innovation. What separates an industrial park where firms happen to be near each other from a true productive system is the subject of the cluster economics tradition.

Giacomo Becattini’s rediscovery of the Marshallian industrial district in the late 1970s started from an empirical anomaly: while theory predicted the crushing of small enterprises by large integrated firms, entire regions of central and north-eastern Italy, the “Third Italy”, were thriving through networks of small firms. Prato in textiles, Sassuolo in ceramics, Carpi in knitwear: territories where hundreds of highly specialised firms formed together a productive machine of remarkable flexibility and quality, exporting to the world. Becattini understood that these districts were not reducible to simple geographic agglomeration. They were socio-territorial entities where a community of people, with shared values, norms of reciprocity, and familial transmission of skills, overlaid a population of enterprises linked by cooperation as much as competition. A workshop overwhelmed with orders subcontracts to its neighbour, knowing the favour will be returned. This combination of competitive rivalry and operational cooperation, coopetition, is the district’s strength. Coordination passes not through the hierarchy of a large firm but through a network of small units connected by markets and social conventions.

Ann Markusen extended this typology to four ideal types. The Marshallian district is the original: small firms, local capital, dense input-output linkages, strong social embeddedness, as in the Third Italy. The hub-and-spoke district is organised around one or a few large anchor firms with local suppliers gravitating around them, the structure of many PAI ecosystems in Morocco. The satellite platform is a set of branch plants or subsidiaries of external firms with limited local linkages, assembled by public incentives but not generating local autonomous capability, the risk profile of much of Morocco’s export processing zones. The state-anchored district is built around a large public enterprise that organises local supply chains and secures the district against capital flight, the OCP complex in Khouribga-Jorf Lasfar is the clearest Moroccan example.

Michael Porter’s cluster diamond adds a managerial and competitive intelligence dimension: competitive advantage in a cluster depends on the interaction of four forces, factor conditions (skilled labour, specialised infrastructure), demand conditions (local customers who are sophisticated and demanding), related and supporting industries (a density of suppliers and complementary activities), and firm strategy and rivalry (local competition that forces continuous improvement). Porter’s framework is more normative than Becattini’s, designed to guide investment rather than simply describe organic emergence, and has influenced explicitly the design of Morocco’s PAI ecosystems.

10.2 Tacit knowledge, absorptive capacity, and the limits of technology transfer

Why does geographic proximity remain decisive for innovation in an age of instant global information flows? Meric Gertler’s answer is a distinction between two types of knowledge. Codified knowledge, what can be written in a manual, a patent, a technical specification, travels freely and at negligible cost. Tacit knowledge, the hands-on judgment, the intuitive diagnosis, the craft knowledge that resists verbalization, transfers only through shared practice, direct observation, and mentorship. It is precisely tacit knowledge that provides competitive advantage, because it is difficult to imitate and therefore to relocate. It remains attached to territory. This is why, despite the internet, innovation continues to concentrate in a handful of places. A production technique cannot be learned by video conference.

10.3 The SECI spiral: how tacit knowledge becomes shared knowledge

Ikujiro Nonaka’s model of knowledge creation, developed in the 1990s, gives the most concrete account of how tacit knowledge circulates within a cluster and gradually becomes collective capability. Knowledge creation, he shows, proceeds through a spiral of four movements: socialisation, externalisation, combination, and internalisation (SECI).

Socialisation transmits tacit knowledge through tacit knowledge: the apprentice learns the master’s technique by watching and imitating, without a single word being written. This is the predominant mode in artisanal districts, in surgical teams, in machine tool workshops, and in Morocco’s aeronautics cluster at Nouaceur, where junior technicians shadow senior ones on composite lay-up procedures that no manual can fully specify. Externalisation converts tacit knowledge into explicit: the experienced engineer manages to describe in a procedure, a diagram, or a standard operating instruction something that was previously pure intuition. This conversion is the hardest step and produces the most value, because explicit knowledge can be stored, transmitted, and combined. Combination assembles existing explicit knowledge to produce new explicit knowledge: crossing datasets, integrating specifications, building on existing codified procedures. Internalisation reconverts explicit knowledge back into tacit: after enough practice, a written procedure becomes second nature, a new embodied competence, ready to restart the spiral.

The cluster matters because socialisation and externalisation, the two moments where tacit knowledge is at stake, require physical proximity, direct observation, shared context, and trust. These conditions are produced by co-location, stable professional communities, worker mobility within the district, and the informal social relations that Marshall heard in the buzz of Sheffield and Becattini documented in Prato. A Moroccan industrial zone that installs firms in adjacent buildings but produces no worker mobility, no shared professional community, no informal knowledge exchange between firms, no ba in Nonaka’s term for the enabling space — will run the SECI spiral slowly or not at all. The distinction between a genuine district and an assembly enclave is precisely this: whether the spiral is turning.

Michael Storper’s formulation captures what this implies at the territorial scale: untraded interdependencies, shared conventions, trust networks, informal circulation of persons and ideas, reputations, are the glue of an innovative territory. These are not market transactions. They are the invisible infrastructure that makes markets and formal contracts work at low cost. They are also the reason why a successful territorial system is so difficult to replicate elsewhere: building the walls is fast; creating the trust and tacit knowledge that inhabit them takes a generation.

10.4 Boschma’s five proximities: too close can be as bad as too far

Ron Boschma’s framework (2005) refines the analysis of proximity by identifying five distinct dimensions that facilitate learning, while warning that excess proximity in any dimension can be as damaging as its absence.

Geographic proximity is the most obvious: physical co-location reduces transport and communication costs and enables the face-to-face contact that tacit knowledge requires. But geographic proximity alone is not sufficient, and can be misleading: firms in the same industrial park may be geographically close while being cognitively or organisationally distant. Cognitive proximity, sharing a knowledge base, a technological paradigm, a professional vocabulary, is what allows partners to understand each other and recombine their knowledge productively. Too little cognitive proximity and communication produces noise; too much and the parties bring no new knowledge to each other. Organisational proximity, belonging to the same network, supply chain, or institutional relationship, creates stable channels for knowledge exchange. Social proximity, interpersonal trust built through repeated interaction and shared social context, reduces transaction costs and makes informal knowledge sharing safe. Institutional proximity, shared rules, norms, standards, and regulatory frameworks, reduces uncertainty and makes coordination predictable.

The critical insight is the optimal proximity hypothesis: learning and innovation require neither too little nor too much proximity along any dimension. A cluster that achieves high geographic, cognitive, and social proximity but closes itself to external knowledge (the lock-in described by Grabher) stagnates in technological space precisely because it is too internally coherent. The Ruhr coal-steel complex failed to anticipate the collapse of its industries not from laziness but from a surfeit of internal coherence that made heterodox signals invisible. Morocco’s industrial ecosystems face the same risk: the PAI automotive cluster around Tanger is building geographic and organisational proximity rapidly, but cognitive proximity to the global automotive engineering frontier (the electric vehicle transition) requires deliberate pipeline investment through partnerships with European engineering schools, Moroccan university research programmes, and co-development agreements that bring frontier knowledge into the cluster rather than waiting for it to diffuse organically.

The concept of absorptive capacity (Cohen and Levinthal, 1990) is crucial for developing economies. To assimilate technology from abroad, it is not enough to have access to it; one must already possess enough knowledge and competence to understand, adapt, and eventually improve upon it. Paradoxically, you need to already know a great deal to be able to learn more. A territory lacking absorptive capacity can host the most modern factories without learning anything from them, the technology remains a black box that is operated but not understood. A territory that has invested in education, applied research, and local engineering capability, by contrast, transforms the presence of foreign firms into genuine learning and eventually autonomous innovation. This is exactly what separates East Asian catching-up trajectories, where assembly served as a springboard to technological mastery, from blocked trajectories where assembly remained a dead end. Absorptive capacity is the missing link between attracting foreign investment and actually developing.

Harald Bathelt and colleagues give this duality a memorable formulation: local buzz and global pipelines. An innovative territory combines the buzz of informal proximity-based exchanges with deliberately constructed channels to distant sources of knowledge. A cluster living only on its internal buzz eventually turns in circles; it needs pipelines to the global frontier to remain competitive. Clusters that are closed systems, high social density, low external connectivity, risk the lock-in that Gernot Grabher documented in the Ruhr’s coal-steel complex: a territory so well-adapted to a particular technological paradigm that it cannot absorb the signals heralding its obsolescence. The district’s strength becomes the source of its inability to transform.

10.5 Perroux, Hirschman, and the test of linkages

The economic sociology of clusters reaches its most operational formulation in two political economy traditions that are essential for evaluating any development project.

François Perroux’s growth poles theory (1955) identifies the propulsive firm, a dynamic, innovative, fast-growing activity of sufficient size, that exercises domination effects on surrounding industries, drawing them into its wake through purchases and sales. But Perroux issued a capital warning, validated by France’s own failed experiment with “métropoles d’équilibre” in the 1960s: a growth pole is primarily an abstraction of economic relationships, not necessarily a geographic concentration. Installing large factories in backward regions does not automatically trigger poles. Too often the factory came but the expected linkage effects did not follow, for lack of a local fabric capable of absorbing them. The implanted enterprise became an enclave: connected to the outside but disconnected from its immediate environment. This warning is directly relevant to evaluating Morocco’s large industrial projects.

Albert Hirschman sharpens the criterion further in The Strategy of Economic Development (1958). An activity creates several types of linkages. Backward linkages stem from its purchases: an assembly plant creates demand for components, potentially generating local suppliers. Forward linkages stem from its production: a steel plant supplies input to a downstream transformation industry. Fiscal linkages generate tax revenues financing further public investment. Consumption linkages arise when distributed incomes create markets for consumer goods.

Hirschman’s test for Morocco

To know whether an industrial concentration is a genuine productive system or an enclave, ask Hirschman’s question: what happens when the lead firm leaves? If its departure leaves behind a fabric of competent suppliers, reusable skills, embedded savoir-faire, and local innovation capacity, then a territorial productive system has constituted itself. If its departure leaves an empty building and unemployed workers, the district was a satellite platform, dependent and removable.

Morocco’s automotive industry is moving toward the first case but has not fully arrived. Morocco’s textile and offshoring sectors remain closer to the second. The phosphate complex, because it rests on a non-relocatable national resource and an integrated operator that controls the chain from extraction to chemicals, is the most solidly anchored, and it is also where local capability upgrading is most advanced. The hierarchy of robustness follows exactly Markusen’s typology: the state-anchored district is most resilient, the hub-and-spoke is promising but dependent, the satellite platform is the most fragile.


11 Part III — Global Value Chains: Who Governs, Who Captures

11.1 From national products to fragmented tasks

For most of economic history, international trade was conceived as the exchange of finished goods between nations: Portugal sold wine, England sold cloth. This Ricardian framework assumed that each good was entirely produced in one country before being exchanged. This is no longer how the world economy works. A smartphone is designed in California, its components manufactured in South Korea, Japan, and Taiwan, assembled in China, and sold everywhere. It has no nationality; it is the product of a chain.

Gary Gereffi’s concept of the global value chain (1994) provides the analytical vocabulary. His central question is not “what is produced where” but “who governs the chain.” A value chain is not an anonymous market where everyone trades freely. It is an organised structure, typically piloted by a dominant lead firm that sets the rules, standards, prices, and profit distribution across the chain. Gereffi distinguishes two broad governance archetypes: producer-driven chains, where large manufacturers in capital-intensive industries (automotive, aeronautics) structure the chain through their technological mastery and heavy investment; and buyer-driven chains, typical of clothing, toys, and agri-food, where large retailers and brand-holders without factories command, they design and sell but source from a worldwide network of subcontractors they place in competition.

The distinction matters enormously for a developing-country supplier. In a producer-driven chain, technological capability determines position. In a buyer-driven chain, cost, flexibility, and compliance with the buyer’s specifications determine position, and the buyer, holding the brand and the customer relationship, can switch suppliers relatively easily.

11.2 The smile curve and the paradox of impoverishing growth

The most powerful image in the economics of global value chains is the smile curve, attributed to Acer’s Stan Shih and formalised by Ram Mudambi. Plotting value added per stage along the production chain, the curve dips in the middle, assembly, and rises at both ends: R&D, design, and intellectual property at the upstream extreme, and branding, distribution, marketing, and after-sales at the downstream extreme. The high-value stages are knowledge-intensive, IP-protected, and typically monopolised by firms in rich countries. Assembly is labour-intensive, technologically standardised, and contestable by any country offering lower costs.

The paradox of immiserising growth follows directly. A country that successfully expands its assembly capacity, attracts more factories, hires more workers, exports more vehicles or garments, can simultaneously see its terms of trade worsen if increased competition from other low-cost assemblers drives down the price of the assembly service. Growth in volume does not guarantee growth in income if the value captured per unit falls. This is the trap Morocco must navigate in textiles and in the commodity segments of automotive assembly: growing activity does not automatically mean growing value capture.

The Moroccan tomato and the European supermarket

A Moroccan farmer grows tomatoes that reach a German supermarket shelf. Who captures the value? The German retailer captures the brand margin and the consumer relationship. The European logistics chain captures transport. Quality inspection, cold chain, traceability certification, all demanded by European standards, are largely provided by service firms in Morocco but to specifications designed in Europe. The Moroccan farmer receives a price at the farm gate that covers production costs but captures little of the final retail margin. To move up the value chain, Morocco would need its own brands recognised by European consumers, its own distribution relationships with retailers, its own logistics capability to deliver just-in-time to multiple markets. Each of these capabilities is being built, slowly and unevenly, in Morocco’s agri-food export sector. Until they exist, the farmer produces the physical substrate; the value accrues elsewhere.

11.3 Five governance structures and their upgrading implications

Gereffi, Humphrey, and Sturgeon (2005) refined the typology to five governance structures, ranging from market (standardised products, easy supplier switching, price-based competition) through modular (complex transactions governed by codified specifications that allow suppliers to manage their own activities), relational (complex interactions requiring mutual adaptation and trust), captive (small suppliers dependent on dominant buyers with limited capabilities) to hierarchy (full vertical integration by the lead firm). Position in this spectrum determines a supplier’s bargaining power, its exposure to switching, and its upgrading trajectory.

Morocco’s automotive sector has evolved from captive toward relational and modular governance as local supplier capabilities have deepened. Tier-2 and tier-3 Moroccan firms that meet international quality standards (IATF 16949), that can manage their own production scheduling, and that have invested in engineering capability are treated as modular suppliers with meaningful autonomy. Firms that remain purely assembly subcontractors under tight supervision remain in captive governance, which means limited learning, limited bargaining power, and high switching risk.

Upgrading is the process of moving to more valuable activities within or across chains. It takes four forms: process upgrading (doing the same task more efficiently); product upgrading (shifting to higher-value products); functional upgrading (taking on higher-value functions, design, marketing, that were previously controlled upstream or downstream); and intersectoral upgrading (applying capabilities built in one chain to a new, higher-value chain). Morocco’s trajectory in phosphate, from raw rock to processed fertiliser to customised agri-nutrition to battery cathode materials, traces a near-textbook sequence of all four forms.

11.4 Coe, Yeung, and strategic coupling

Neil Coe and Henry Yeung’s Global Production Networks framework adds a crucial relational dimension: a territory’s insertion in global chains is not an exogenous fact determined by comparative advantage. It is the outcome of a strategic coupling between a territory’s specific assets (skills, infrastructure, proximity, institutions) and the strategic needs of global lead firms at a particular moment. This coupling is constructed through deliberate policy, active promotion, and institutional complementarities, and it is reversible. A firm that coupled with Morocco because of proximity to Europe and low assembly costs may recouple with a different territory if Morocco’s cost advantage erodes or if automation reduces the importance of low-wage assembly.

Tanger-Med is the paradigmatic case of successful strategic coupling. Morocco transformed geographic proximity to the Strait of Gibraltar from a natural fact into a constructed competitive advantage through infrastructure investment of global standard, a free zone regime, trade agreement networks, and targeted FDI promotion. The port complex has made Morocco a node in Mediterranean logistics networks that generate transhipment revenues and anchorage for industrial zones. Its value is not the geography alone, it is the institutional and logistical capability built around the geography that makes it competitive.


12 Part IV — The Structuralist Critique: Periphery, Dependence, and the Question of Power

12.1 Prebisch-Singer and the terms of trade

The analytical tools developed above describe mechanisms. They are largely silent about the political economy of who benefits. The structuralist tradition, rooted in Latin American development economics, provides the missing dimension.

Raúl Prebisch and Hans Singer, independently, formulated in 1950 the thesis of the secular deterioration of the terms of trade for primary commodity exporters. The reasoning: demand for primary commodities grows less rapidly than income, while demand for manufactured goods accelerates (Engel’s Law applied to trade). Productivity gains in the industrial centre translate into higher wages and profits, because organised labour markets and concentrated market structures allow workers and firms to capture the gains, while in the periphery, productivity gains translate into lower prices, because competitive commodity markets and weak labour organisations prevent capture. Consequence: to buy the same machine, the periphery must sell progressively more sacks of coffee or tonnes of ore. Trade, far from equalising living standards as Ricardo promised, structurally transfers value from periphery to centre.

The empirical record is mixed over different periods and products. But the core insight has proved durable: primary commodity specialisation exposes to terms-of-trade deterioration and ruinous price volatility, and the project of industrialisation as a path out is well-founded.

Wallerstein’s world-system analysis adds the category that was missing: between the dominant centre and the dominated periphery sits the semi-periphery, countries combining both roles, exploited by the centre and exploiting more peripheral economies in turn, and capable (in periods of world-system recomposition) of moving upward toward the centre or downward toward deeper peripheralisation. The semi-periphery is the zone of mobility in the world system.

Morocco is an almost ideal-typical semi-periphery. Too industrialised and too integrated in European production networks to be a pure primary commodity exporter. Not yet autonomous enough in technology, brand ownership, and financial control to belong to the centre. Fernando Henrique Cardoso’s associated dependent development names this third position precisely: genuine growth and industrialisation are possible within dependency, but they remain subordinated, their dynamics, technologies, and markets commanded partly from outside. Knowing Morocco is a semi-periphery is not pessimistic. It is analytically precise, and it identifies the degrees of freedom that exist.

12.2 The smile of phosphate and the frown of textiles

Applied to Morocco’s current position, the full analytical framework generates a differentiated verdict, not a single judgment but a map of positions across sectors.

OCP and phosphate represent Morocco’s most successful value capture strategy, precisely because they rest on a non-relocatable natural endowment. Rather than exporting raw rock, the pure periphery option, OCP systematically moved up the chain: phosphoric acid, diammonium phosphate, customised NPK fertilisers for specific soil profiles, and now positioning for battery cathode materials as EV adoption accelerates. This is a multi-stage functional and intersectoral upgrading trajectory that contradicts the most rigid structuralist pessimism: a resource-based firm from the semi-periphery becoming a global market-power actor on its segment. The enabling condition was state-supported integration, not private market logic, that maintained the value within a national institution rather than licensing it to foreign processors.

The automotive sector sits in an intermediate and dynamic position. Morocco has successfully attracted both Renault-Tanger and Stellantis-Kénitra, become Africa’s leading vehicle exporter, and progressively increased the domestic integration rate. The trajectory moves from purely captive assembly (hub-and-spoke district in Markusen’s terms) toward increasing relational governance as Moroccan tier-2 and tier-3 suppliers develop. The electric vehicle transition is both the largest risk (the existing component competencies are internal combustion specific) and the largest potential opportunity (Morocco’s phosphate endowment positions it for battery supply chains if the technological and hydric challenges are resolved).

Aeronautics around Casablanca-Nouaceur is technologically sophisticated but strategically fragile. The dependent-development diagnosis applies most precisely here: real capability in precision machining, cabling, and composites; growing skills; but no conception, no intellectual property, no brand, and a very short list of global decision-makers whose sourcing choices determine the entire sector’s trajectory. Cardoso’s formulation, development real but commanded from outside, is exact.

12.3 Agro-industry: the district that is also a chain

Morocco’s agri-food export sector offers an instructive intermediate case that escapes easy classification. The Souss-Massa basin around Agadir, the Berkane citrus zone in the Oriental, the soft fruit districts of the Atlantic coast: these territories exhibit genuine Marshallian district characteristics. Hundreds of producers, conditioned by generations of specialised practice, share input markets (irrigation infrastructure, packaging suppliers, refrigerated transport networks), a thick regional labour market for agricultural and logistics workers, and tacit knowledge of crop cycles, pest management, and export timing that circulates through farmer associations, cooperatives, and the dense informal networks of producers who have grown up in the same valleys.

But these districts are simultaneously inserted in buyer-driven global value chains piloted from Brussels, Paris, and London. The governance mechanism is not price: it is certification and standards. European retailers (Carrefour, Lidl, Intermarché) and their logistics intermediaries impose compliance with GlobalGAP (Good Agricultural Practices), BRCGS food safety standards, BIO certifications, and bespoke traceability systems that connect a tomato in a Rotterdam supermarket to a specific plot in Ait Melloul. This governance by norms is the dominant mechanism of value chain coordination in agri-food: the lead firm sets the rules; the supplier absorbs the compliance cost; the consumer pays the quality premium; and the lead firm captures the brand value and the customer relationship.

The welfare implications are double-edged. Morocco’s agri-food producers have been genuinely upgraded by European standards: quality management systems, cold chain discipline, traceability infrastructure, and export logistics have transformed Moroccan horticulture in two decades. These are real process and product upgrades, visible in the ONSSA certification capacity and the logistics hubs that handle chilled exports through Tanger-Med. But functional upgrading, owning the brand, controlling distribution, managing the consumer relationship, remains minimal. And the standards regime creates a paradox: meeting European food safety requirements is expensive and technically demanding, which creates barriers that exclude smaller producers who lack the capital and technical capacity for compliance, concentrating export capacity in larger commercial farms while marginalising the smallholders that make up most of the basin’s social fabric. The Plan Maroc Vert and Génération Green strategies have sought to aggregate small producers through cooperatives and aggregator structures, with real but uneven results. The agro-industry case shows that governance by norms is not neutral: it upgrades those who can comply and excludes those who cannot, restructuring the district’s social geography even as it raises its aggregate technical level.

12.4 The territorial policy dilemma: efficiency versus equity

The analytical framework this chapter has built does not end in description. It generates a fundamental policy question that any government managing a polarised territory must face: if cumulative causation generates self-reinforcing spatial inequality, and if concentration is productive in aggregate terms, should public investment reinforce the metropolis or compensate the periphery?

The efficiency argument for the metropolis is clear: agglomeration externalities make investment in dense areas more productive per dirham. A road that reduces congestion in Casablanca generates more economic activity than the same investment in a region with sparse productive activity. Education investment near a cluster of high-skill employers produces higher private and social returns than equivalent investment in a zone without a labour market thick enough to absorb graduates.

The equity argument for the periphery is also clear, and it is reinforced by Rodríguez-Pose’s political economy dimension: territorial inequality that concentrates prosperity in the coastal corridor while the interior loses its youth, its public services, and its productive dynamism creates social resentment and political instability that eventually imposes costs on the entire national economy.

Morocco’s Régionalisation Avancée (2015) represented a structural response to this dilemma: devolving economic governance capacity to the twelve regions, giving regional councils budget authority and economic planning mandates, and creating Regional Investment Centres (CRIs) as one-stop shops for investors in each region. The ambition was to create multiple centres of economic dynamism that could partially break the coastal monopoly on growth. The results have been uneven: regions with existing productive assets and strong regional elites (Souss-Massa, Tanger-Tétouan-Al Hoceima, Fès-Meknès) have used the framework actively; interior regions with weaker institutional endowments have struggled to translate formal devolution into genuine economic dynamism. The lesson from the agglomeration theory is sobering: devolution of formal authority does not automatically trigger the cumulative success spiral. What triggers it is a critical mass of productive activity generating the sharing, matching, and learning externalities that pull further investment. Reaching that critical mass in peripheral regions requires more than administrative decentralisation; it requires deliberate productive investment sufficient to cross the tipping point, sustained long enough that the cumulative mechanism engages.

12.5 The nearshoring window and its conditions

A structural shift in global supply chain thinking after the COVID-19 disruptions of 2020-2021 has created a window of opportunity for Morocco that did not exist in the same form before. European manufacturers and retailers, having experienced catastrophic supply chain disruptions when Asian manufacturing halted, have developed a systematic preference for geographic diversification toward nearby suppliers. This nearshoring impulse is reinforced by three additional drivers: the EU’s Carbon Border Adjustment Mechanism, which will raise the cost of high-carbon production in distant locations relative to low-carbon production nearby; rising shipping costs and longer lead times that erode the price advantage of distant assembly; and growing consumer and regulatory demand for supply chain transparency and accountability that is more easily satisfied with suppliers at shorter distance.

For Morocco, this window is real but conditional. The sectors that benefit are those where proximity provides genuine operational advantage: just-in-time automotive components, short-cycle fashion textiles, agri-food requiring fresh supply, and business process outsourcing in French and Arabic. For each of these, the window opens specifically if Morocco develops capabilities that generic low-cost assembly cannot provide: design capability, flexible production at small quantities, digital supply chain management, quality certification. Capturing the nearshoring window is therefore not a matter of cost competition but of capability positioning.

Textiles and garments face the clearest smile-curve trap. Inserted in buyer-driven chains where European retailers control the brand and the consumer relationship, Moroccan producers operate in the assembly trough. Nearshoring, the European demand for short-lead-time proximity suppliers after COVID supply chain disruptions, offers a genuine repositioning window: Moroccan suppliers who can deliver fast, small-series, higher-fashion-cycle orders capture value that Asian distance suppliers cannot provide. But capturing this window requires upgrading to design capability, brand, and direct retailer relationships, precisely the functional upgrading that moves producers out of the trough.


13 Conclusion: The Micro That Makes the Macro

This chapter has built three nested arguments that together constitute the economics of where and how development happens.

Territory is not a passive container. Agglomeration forces, sharing, matching, learning, make concentrated activity more productive than dispersed activity. This productivity premium justifies spatial concentration. But concentration also generates cumulative divergence, Myrdal’s backwash effects dominating spread effects, that the market left to itself will deepen rather than correct. Morocco’s coastal macrocephaly is not efficient geography; it is constructed inequality. Addressing it requires deliberate counter-force, not merely improving the general climate.

Clusters are not parks. Co-location generates agglomeration economies. Genuine productive systems generate learning, tacit knowledge transmission, Hirschman linkages, and embeddedness that make them resilient when external shocks hit. The critical test is what survives if the lead firm leaves. Morocco’s PAI ecosystems range from genuinely systemic (OCP-Khouribga) to potentially systemic (automotive if local suppliers deepen) to currently enclave (much of offshoring and lower-tier textile assembly).

Global value chains distribute power unequally. The smile curve is not a metaphor, it is a quantitative fact about where value accrues. Being inserted in a chain is not the same as capturing value from it. Morocco is a semi-periphery in motion: it assembles more than it captures in most sectors, captures substantially in phosphate, and is accumulating the absorptive capacity in automotive and aeronautics that upgrading requires. The window, nearshoring, the EV transition, renewable energy positioning, is real. Whether Morocco passes through it or finds it closing behind it is a question this book returns to in Chapters 11, 12, and the Conclusion.

The critical connective thread: the microeconomic decisions of firms, Coase’s make-or-buy, Williamson’s asset specificity, the lead firm’s governance choice, aggregate into the macroeconomic structure of where value accrues globally. The regional economist who maps cluster density and the trade economist who measures terms of trade are both reading the same accumulated record of millions of firm-level decisions. This is how the micro makes the macro.

What this chapter should leave with the reader

Agglomeration generates productivity through sharing, matching, and learning. Doubling employment density raises productivity by 3–8%. Concentration is partly self-generated: Krugman’s model shows it can emerge without natural advantage.

Cumulative causation (Myrdal, Kaldor) means left-to-market territorial dynamics diverge rather than converge. Backwash effects dominate spread effects in weakly integrated economies. Morocco’s coastal concentration is historically constructed, not naturally determined.

Clusters are socio-territorial systems, not just co-located firms. Becattini’s industrial districts, Porter’s competitive diamonds, Markusen’s four types, and Hirschman’s linkage test all ask the same question: are the firms producing a collective system or merely occupying adjacent buildings?

Absorptive capacity determines whether foreign investment generates learning or remains a black box. Building it requires prior investment in engineering education, applied research, and worker skill, not just attractive industrial zones.

Global value chains are governed by lead firms that capture value at the smile curve’s ends (design, brands) while assigning assembly, the trough, to suppliers they can switch. Upgrading means moving along the smile: process, product, functional, intersectoral.

Morocco is a semi-periphery in motion (Wallerstein/Cardoso). It assembles more than it captures, with OCP as the principal exception. The nearshoring window, the EV transition, and renewable energy positioning offer real upgrading opportunities, but only if absorptive capacity, not just assembly capacity, is built.


14 Further Reading

Marshall, A. (1890). Principles of Economics, Book IV. The foundational analysis of industrial districts and knowledge spillovers. Still readable and rich.

Krugman, P. (1991). Geography and Trade. MIT Press. The originating text of the New Economic Geography. Accessible and written with Krugman’s characteristic clarity.

Myrdal, G. (1957). Economic Theory and Underdeveloped Regions. Duckworth. The most important heterodox contribution to economic geography. The concept of circular cumulative causation provides the essential counterpoint to Krugman.

Hirschman, A.O. (1958). The Strategy of Economic Development. Yale University Press. The linkages framework. One of development economics’ most original and operationally useful texts.

Becattini, G. (1990). “The Marshallian industrial district as a socio-economic notion.” The founding paper of the Italian school. Essential for distinguishing genuine districts from mere industrial parks.

Gereffi, G., Humphrey, J. and Sturgeon, T. (2005). “The governance of global value chains.” Review of International Political Economy. The authoritative taxonomy of GVC governance structures and their upgrading implications.

Baldwin, R. (2016). The Great Convergence. Harvard University Press. The most lucid account of how trade shifted from goods exchange to task fragmentation, and what this means for developing countries.

Cardoso, F.H. and Faletto, E. (1969). Dependency and Development in Latin America. The most nuanced statement of the dependent-development thesis. Applied to Morocco, it is more illuminating than either optimistic insertion narratives or pessimistic dependency paralysis.

Rodríguez-Pose, A. (2018). “The revenge of the places that don’t matter.” Cambridge Journal of Regions, Economy and Society. The political economy of territorial inequality. Makes the connection between spatial economics and political instability that every development economist should know.