Picture the moment. You are a clerk in a Venetian trading house, recording the costs of a caravan from the Black Sea port of Trebizond to Tabriz, then onward into the Mongol-administered roads of Central Asia. The sea is right there, visible from the quay. You ignore it. So do your employers, for generations.
The puzzle is genuine. Venice and Genoa controlled fleets capable of oceanic passage. Coastal shipping was cheaper per ton-mile than overland transport by almost every measure historians have reconstructed. And yet the most profitable commercial networks of the twelfth through fifteenth centuries ran through mountain passes, desert waypoints, and landlocked relay cities, pulling Italian, Catalan, and later Hanseatic capital deep into the continental interior. The obvious route was not the right route. That gap deserves an explanation.
The road that could be taxed was the road that could be trusted
Profitability is not revenue per mile. It is the spread between price at origin and price at destination, minus every cost and risk along the way. Coastal routes were cheap to sail but expensive to secure. The open sea in the medieval Mediterranean and Baltic was ungoverned space, and pirates operated with near-impunity along the Dalmatian coast, around Sicily, and throughout the Aegean. A merchant republic could escort a convoy, but convoy protection required naval assets, coordination, and the cooperation of whatever power controlled the coastline at the next port. None of that was free. None of it was reliable.
Landlocked routes, paradoxically, ran through governed space. The Mongol Pax, which held Central Asian roads from roughly the mid-thirteenth century onward, is the textbook case: the Yuan and Ilkhanate administrations actively taxed caravans, which meant they also had a direct financial incentive to suppress banditry. Think of it less like a highway toll and more like a protection racket with a regulator's budget. A merchant paying at Samarkand was buying something real. The Venetians understood this with the clarity of people who had to make payroll. Their commercial treaties with Mamluk Egypt and later with the Ilkhanate were not goodwill gestures; they were attempts to purchase predictability on routes where the tollkeeper was also the policeman.
The arithmetic still surprises. Overland freight from China to the Levant cost roughly three to four times as much per unit weight as sea freight would have, if sea freight had been available end-to-end. But the markup on silk and spices in Venetian or Genoese markets was so large, sometimes tenfold on raw silk between Cathay and the Rialto, that the extra transport cost shrank to a rounding error against the margin. What destroyed profit was not cost per mile. It was theft, spoilage, and the catastrophic loss of an entire cargo to a raid or a storm with no legal recourse. That tenfold markup implies the overland premium was, at most, a minor line item next to the downside of a single uninsured loss.
What the coast was actually missing
There is a second mechanism that gets less attention: the infrastructure of credit. The landlocked routes of the medieval period were threaded with cities, caravanserais, and relay stations spaced roughly a day's march apart. That spacing was not accidental. It created a network of known intermediaries, people a Genoese factor in Caffa could correspond with, extend credit to, and eventually litigate against through the emerging body of merchant law. The commenda contract, the instrument that let a Venetian investor fund a caravan without traveling himself, only worked if there were identifiable counterparties at each node.
Coastal alternatives often lacked this entirely. The western African coast, the Arabian Sea shore, the outer Atlantic margin: these were not empty, but their commercial infrastructure was not legible to European merchant houses in the same way. A Venetian firm could not extend a letter of credit to a port it had never factored, could not insure a cargo through routes where its legal instruments had no standing, and could not collect on a debt in a jurisdiction its consuls had never penetrated. Ask yourself: what is a cheaper route actually worth if you cannot enforce delivery at the far end?
Consider two Genoese merchants, call them Marco and Benedetto, both shipping Armenian silk westward in the 1320s. Marco takes a caravan route through Tabriz, paying tolls at four checkpoints but holding a commenda contract underwritten by a Genoese notary, with a correspondent agent in each city. Benedetto experiments with a coastal dhow route, cheaper per bale, but outside the network of Genoese consulates and with no instrument to enforce delivery. Marco's costs are higher. His variance is lower. Over ten voyages, Marco retires. Benedetto does not. The spread between their outcomes is not explained by geography; it is explained by the institutional layer one of them had access to and the other did not.
This is the lesson the merchant republics were teaching, though they never articulated it this way. A route's value is inseparable from the legal and financial architecture built on top of it. Geography sets the menu. Law, credit, and enforcement capacity determine what you can order. The landlocked Silk Road was, by the fourteenth century, a more legible commercial environment for a sophisticated Italian house than any coastal shortcut, and legibility, the ability to price risk, extend credit, and collect on a contract, is worth more than cheap freight. Capital followed systems then. It still does.