Why the Roman Road Network Outlasted Rome

In 1745, a French engineer named Pierre-Marie-Jérôme Trésaguet was assigned to survey the road between Limoges and Paris and found, to his professional embarrassment, that the most structurally sound sections of the route were the ones built by Roman engineers roughly fifteen hundred years earlier. The Roman sections had survived medieval neglect, seasonal flooding, and centuries of heavy cart traffic in better condition than the sections built by his own predecessors over the previous two hundred years. Trésaguet went home and invented modern road engineering — essentially rediscovering Roman techniques he had just seen in the field. His 1775 treatise became the foundation of French road construction for the next century.

This vignette is amusing, but it obscures a more interesting question than “were Roman engineers good?” (They were.) The real question is why the roads survived at all. Most physical infrastructure from antiquity did not. Roman aqueducts fell into disuse within decades of the Western Empire’s collapse; the sophisticated drainage systems of Roman cities silted up and were not maintained; the central heating systems of Roman villas were simply abandoned. Why did the roads persist when everything else decayed? The answer has almost nothing to do with engineering and almost everything to do with economics and geography.

What Roads Actually Are

Roads are commonly understood as physical objects — gravel and stone laid over a substrate in a particular way. But this physical description misses the essential characteristic: a road is a coordination technology. Its value is not intrinsic to the object itself but emerges from the fact that multiple parties have oriented their economic activity around using it. A road that connects two population centers is valuable because merchants, armies, postal services, pilgrims, and migrants all use the same route. Each user benefits from the fact that others are using it, and this mutually reinforcing pattern is what makes roads — once established — extraordinarily resistant to abandonment.

The Roman road network at its peak covered approximately 400,000 kilometers, with about 80,000 kilometers of primary roads surfaced with stone. This was not built primarily as a military asset, although military logistics obviously benefited. It was built as a communications and commerce system, and the commercial activity it enabled rapidly became dependent on its continued operation. Merchant networks organized their supply chains around the road network. Market towns grew at road junctions. Agricultural land near roads commanded premium prices because access to markets was easier. All of this economic activity created constituencies with direct financial interests in maintaining the infrastructure.

When the Western Roman Empire collapsed in the late 5th century, the political authority that had originally built and maintained the roads disappeared. The tax base that had funded maintenance crews evaporated. The centralized administrative apparatus that had coordinated repairs across thousands of kilometers ceased to function. Under any straightforward model of infrastructure maintenance, the roads should have deteriorated rapidly and been abandoned within a generation or two.

But the economic geography the roads had created did not disappear with the empire. The towns were still there. The trade routes were still there. The merchants who needed to move goods from one city to another still needed to move goods. The pilgrim routes to Rome, Compostela, and Jerusalem still needed to exist. The roads continued to be used because the economic activity they supported continued to exist, and continued use — even heavy use — is itself a form of maintenance. Loose stone gets pushed back into place by cart wheels. Ruts get filled by travelers throwing in broken branches and gravel. The road does not maintain itself, but constant use keeps it marginally functional.

Path Dependence and the Lock-In of Routes

The more profound mechanism is path dependence. Once a road establishes itself as the dominant route between two points, it becomes progressively more difficult for any alternative route to compete, regardless of whether the original route is optimal. Economic activity — market towns, inns, smithies, warehouses — clusters along the established route. This clustering makes the established route more valuable for subsequent users, which attracts more activity, which makes it more valuable still.

The Roman road from London to York — the old Roman road known as Ermine Street — is still the basis for the modern A1 motorway. The road from London to Canterbury is still broadly the route of the A2. This is not because Roman engineers made perfect choices about where to put roads. It is because the economic ecosystems that grew up along those roads made deviation increasingly expensive as centuries passed. Every town, every bridge, every market, every institutional network that oriented itself along the Roman alignment added another layer of lock-in.

The medieval English economy had no particular reason to organize itself around routes chosen by Roman military engineers two thousand years earlier. Yet it did, because the inheritance of Roman road alignment came bundled with an inheritance of settlement patterns, market rights, and trade relationships that made departing from it prohibitively costly. A new road from London to York that cut a more efficient modern path would have bypassed all the existing market towns — which meant it would not have had customers, which meant it would not have had inns and stables, which meant it would not have been useful for traveling merchants, which meant it would have been worthless.

This is the economic logic that underlies virtually all infrastructural path dependence. Modern container shipping routes follow paths established by the clipper ships that preceded them, which followed paths established by the sailing routes before them, which were constrained by the geography of winds and currents that sailors learned about incrementally over centuries. The technology changes; the route persists. The coordination point that emerges around a route is more durable than the physical infrastructure itself, and far more durable than the political authority that created it.

The Medieval Maintenance System

It would be wrong to suggest that the Roman roads survived purely through use and without any deliberate maintenance. The medieval period saw a patchwork of institutional arrangements emerge to keep key routes functional, and these arrangements are interesting precisely because they operated without any central authority and without any single entity that “owned” the roads.

The most important mechanism was the corvée — compulsory labor by local residents on roads that ran through their territory. This was a pre-Roman institution that the Roman Empire had formalized, and it persisted in various forms across medieval Europe as the primary mechanism of road maintenance. The logic was straightforward: the people who lived nearest to a road section also benefited most directly from its maintenance, and the compulsory labor obligation distributed the cost of maintenance across the community that benefited from it.

Tolls were the second mechanism. Medieval toll roads — pontage (levied on bridges) and pavage (levied on road surfaces) — were granted by royal charter to individuals or institutions in exchange for commitments to maintain the infrastructure. Monasteries were particularly important in this system. A monastery situated on a major pilgrimage or trade route had obvious interest in keeping the route accessible, received donations from grateful travelers, and had the organizational capacity to manage both collections and labor deployment. Many of the best-maintained road sections in medieval Europe ran through monastic lands precisely for this reason.

The result was not a coherent road network maintained to a uniform standard. It was a patchwork of varying quality, maintained by different institutions under different arrangements, with the best sections typically corresponding to the highest commercial or pilgrimage traffic. But the key routes — the ones that had been Roman roads and retained their economic centrality — generally stayed functional enough to support the traffic that kept them economically relevant. The system was self-reinforcing: importance generated maintenance resources, and maintenance resources preserved importance.

Why Modern Infrastructure Does Not Work This Way

The Roman road story has a direct lesson for how modern infrastructure investment is evaluated and managed, and the lesson is not flattering to current practice. Modern infrastructure investment is typically analyzed on a project-by-project basis, with costs and benefits assessed over a standard horizon of fifteen to thirty years. This framework systematically underestimates the value of investments that create durable coordination points — the kind of path-dependent lock-in that made Roman roads still useful fifteen centuries after their construction.

The economic benefit of the A1 motorway in Britain is not just the traffic it carries today and the time savings it generates for current users. It is the entire economic ecosystem — the cities, the industrial geography, the labor markets — that has oriented itself around that corridor over two thousand years of continuous use. Trying to capture this in a net present value calculation with a thirty-year horizon is not just methodologically difficult; it is categorically wrong. It treats a coordination investment as if it were a consumption investment, and the two work on fundamentally different timescales.

High-speed rail networks illustrate the same point. The case for high-speed rail investment is routinely made and evaluated on the basis of projected ridership and time savings for a population distributed in its current spatial pattern. But large-scale transport infrastructure does not serve a fixed population distribution — it reshapes the population distribution over time. The Shinkansen in Japan did not just connect existing cities faster; it created new cities and industrial clusters in the corridors it ran through, restructured the labor market geography of the Japanese archipelago, and effectively expanded the functional size of the Tokyo and Osaka metropolitan areas by making commuting from two hundred kilometers away viable. The network effects of this spatial restructuring dwarf the direct ridership benefits that justified the initial investment.

Roman road engineers were not doing cost-benefit analysis. They were building military and administrative infrastructure for an empire that needed to move legions and information quickly across vast distances. The economic ecosystems that grew up around their roads were, from the perspective of the original builders, externalities — unintended side effects. But those externalities turned out to be the most durable product of Roman engineering, outlasting not just the roads themselves (many sections are still technically intact) but the civilization that produced them.

The Enduring Logic of Geographic Commitment

The deepest insight of the Roman road story is that infrastructure investment is not primarily a technical or economic decision — it is a geographic commitment. When you build a road, you are not just creating a transport link; you are asserting something about where economic activity will happen. If you are right — if the route you have chosen connects points that will remain economically relevant — the investment will compound across generations in ways that are impossible to predict at the time of construction. If you are wrong, you get an abandoned Roman road in the middle of nowhere, of which there are plenty.

The Romans were right more often than they were wrong, partly because they were good at identifying strategically important places (river crossings, mountain passes, coastal ports) and partly because their military presence itself generated the economic activity that justified the routes they chose. The road to a garrison town was important because the garrison was there; the garrison was there partly because the road made supply logistics manageable; the town that grew up to serve the garrison remained after the garrison left, because the town had become a node in commercial networks that now valued its location.

This circular causation — infrastructure creating economic activity that justifies infrastructure — is why Trésaguet found Roman road sections in better condition than French ones in 1745. The Roman sections were on the routes that had carried the most traffic for the longest time. The French sections were on diversions and additions that had not yet accumulated the economic gravitational pull of the core Roman routes. The engineers who built Ermine Street could not have known that fifteen centuries of traffic would maintain what they built. But the geographic logic they identified — that certain routes between certain places would carry disproportionate economic activity — turned out to be stable across a time horizon that makes almost every other human institution look ephemeral.

Build the right roads. The economics will maintain them for you.