ECONOMICS · MARKETS · Capital
Infrastructure Economics
When Is Extra Harbour-Pilot Capacity Worth Paying For in Rotterdam?
Rotterdam's 29 October pilotage warnings reported a demand surge, not measured vessel delays. Rokin's illustrative 48-hour model asks when reserve pilotage capacity is worth paying for and tests it against rescheduling arrivals. The result is a useful distinction: added service can shrink a long recovery queue, while shifting scheduled arrivals can simply replace unexpected waiting with planned delay.
Rokin Research
Key takeaways
The transition’s investable bottlenecks are increasingly physical: connection queues, grid equipment, permitting capacity, and skilled delivery.
1. In a constructed 48-hour Rotterdam pilotage scenario, three additional effective positions avoid 168.75 queued assignment-hours, not measured incident delays. 2. The value of each further reserve position diminishes as the queue is reduced; timing and deployability matter more than a regional workforce total. 3. Rescheduling arrivals can replace unexpected waiting with planned deferral rather than eliminate vessel-time cost; other port constraints may still bind.
On 29 October 2026, Rotterdam's port information system announced two separate periods of special pilotage-capacity arrangements. The first began at 01:53 and was deactivated at 07:59. A second began at 22:58. In each activation, the port warned that a surge in pilot demand could leave vessels waiting three hours or longer for boarding. It did not report how many vessels actually waited or for how long. [1][2][3]
That distinction changes the economic question. Rather than assign a speculative bill to the episode, it is more useful to ask when an extra unit of pilotage capacity creates enough value to justify its cost. The answer depends less on annual staffing totals than on how closely demand approaches service capacity during a particular shift.
A short surge can have a long tail
Pilotage is a service that cannot simply be stockpiled. When more assignments become ready than qualified pilots can start, a queue forms. Once arrivals ease, the backlog disappears only as quickly as the difference between incoming work and available service capacity allows.
Consider a deliberately constructed 48-hour example. Assume 25 simultaneously effective pilotage positions, each tied up for an average of 3.5 hours per assignment, including the activities that determine when it can take another job. That implies capacity of about 7.14 assignments per hour. Against six ordinary hourly requests, there is just over one assignment per hour of spare capacity.
Now suppose demand reaches 11 assignments per hour for six hours. The queue grows by roughly 3.86 assignments each hour. When requests return to six per hour, it shrinks at only about 1.14 per hour. The surge ends well before its waiting-time consequences do. These numbers are assumptions used to expose the mechanism, not the observed Rotterdam order book.
The relevant quantity is the area under the queue over time: one assignment waiting for an hour contributes one queued assignment-hour. The model equates this with a vessel-hour only under its assumption of one pilotage assignment per vessel. It is not a measurement of time at anchor, berth delay or cargo loss.
Figure 1. The same constructed workload under four operating regimes. The shaded band is an illustrative six-hour pulse, not the 29 October event. Source: Rokin Research, 48-hour pilotage model v0.2.
What the next position changes
In the scenario, three additional effective positions are staffed for 18 hours. Service capacity rises from about 7.14 to eight assignments per hour during that window. The initial surge still exceeds available service, but the queue is smaller and the subsequent recovery faster.
Across the 48-hour horizon, the baseline accumulates 303.75 queued assignment-hours. The three-position reserve produces 135, avoiding 168.75 hours, or approximately 56%. The modelled queue clears by hour 23 rather than hour 35. These are conditional differences between two hypothetical operating regimes, not an estimate of what could have been saved on 29 October.
What would that capacity be worth? At an assumed incremental staffing cost of €180 per effective position-hour, 3 positions multiplied by 18 hours cost €9,720. Dividing that by 168.75 avoided queued hours gives a break-even value of €57.60 for each avoided waiting hour. If the actual economic value of avoidable waiting were below that threshold, the particular deployment would not pay for itself in this one-episode comparison. If it were higher, it would, provided the additional positions could genuinely deliver the modelled service.
The €57.60 figure is the average break-even value for the three-position package, not the marginal threshold for the third position, a pilot's salary or a measured vessel cost. A continuously available position also does not necessarily correspond to one person over an 18-hour period. Duty limits, qualifications, transfer time and rest requirements still apply.
Why the first reserve position matters most
The relationship is nonlinear. In this example, the first additional effective position removes about 77 queued hours. The second removes a further 54; the third, another 38. Later positions bring progressively smaller reductions because earlier additions have already removed much of the backlog.
Figure 2. Marginal queue-hours avoided in the constructed six-hour demand pulse. Bars are scenario outcomes, not an optimal staffing recommendation. Source: Rokin Research, pilotage model v0.2.
This does not establish that three additional pilots are optimal. In fact, at a separate illustrative waiting value of €1,500 per hour, the model's net benefit continues rising through all eight reserve positions tested. The cost of permanent readiness, the probability of each surge and the practical ability to staff and deploy those positions are outside that one-event calculation.
Moving arrivals may move the cost rather than remove it
There is another response to peak congestion: persuade some vessels to take later slots. In the model, shifting 35% of the 30 above-background surge requests moves 10.5 fluid-equivalent assignments into later hours. That reduces unexpected queueing by 94.5 assignment-hours. But it also moves their intended service time later by 94.5 aggregate hours.
This is not 94.5 hours of free time recovered. If a planned hour of deferral and an unexpected hour of waiting had the same economic cost, the modelled schedule shift would produce no net vessel-time saving before other effects. Its attractiveness depends on whether an earlier notice allows a ship to slow steam, adjust operations or avoid more costly unscheduled idle time. Those advantages are plausible mechanisms, not measured benefits in this study. [7]
Combining rescheduling and reserve staffing produces just 47.5 queued hours in the illustration, but the effects are not additive: both measures are working on the same backlog. The distinction matters to operators comparing staffing, arrival coordination and other ways of protecting schedule reliability.
The real bottleneck still has to be identified
A pilot is only one input into a safe port movement. Qualified availability at the correct boarding area, transfer boats, tugs, visibility, weather and berth access can each impose a separate constraint. Rotterdam activated a visibility arrangement at 07:21 on 29 October, overlapping the earlier pilotage-capacity warning. More available pilots would not necessarily have made weather-restricted movements feasible. [4]
The regional operator describes approximately 220 professional pilots serving around 53,000 vessels annually across Rotterdam-Rijnmond. That provides scale, not a shift roster. The port separately reports 27,384 seagoing visits in 2025, a different geographical and operational denominator. Neither figure can calibrate the number of pilots or queued vessels during these alerts. [5][6]
The commercial decision, therefore, is not simply whether to hire three people. It is whether to fund effective capacity at the hours and locations where it relieves the binding constraint, versus spending on coordination, flexibility or another service bottleneck. Establishing that case requires actual pilot orders, boarding times, deployable capacity, operating restrictions, incident frequency and incremental costs. Until those observations are available, the defensible conclusion is conditional: reserve capacity can have a disproportionate effect near a service threshold, but its realised value depends on the operational bottleneck it can actually remove.
Research note. This article examines an illustrative, deterministic 48-hour queue model prompted by Rotterdam's 29 October 2026 pilotage notices. The 303.75, 135 and 168.75-hour figures, the €57.60 break-even threshold and all staffing and valuation inputs are scenario quantities, not observed event results or a social cost-benefit estimate. The technical report documents the calculation, sensitivity tests and empirical data requirements.
Sources
Port of Rotterdam, Capacity Arrangement Pilots - Activate, 29 October 2026, 01:53. https://pin.portofrotterdam.com/pin-16782
Port of Rotterdam, Capacity Arrangement Pilots - De-activate, 29 October 2026, 07:59. https://pin.portofrotterdam.com/pin-16785
Port of Rotterdam, Capacity Arrangement Pilots - Activate, 29 October 2026, 22:58. https://pin.portofrotterdam.com/pin-16791
Port of Rotterdam, Visibility Arrangement - Activate, 29 October 2026, 07:21–09:19. https://pin.portofrotterdam.com/pin-16784
Nederlands Loodswezen, Regio Rotterdam-Rijnmond, regional profile; accessed 30 October 2026. https://loodswezen.nl/regios/rotterdam-rijnmond/
Port of Rotterdam Authority, Nautical annual figures 2025, 3 February 2026. https://www.portofrotterdam.com/en/news-and-press-releases/nautical-annual-figures-2025-harbour-master-focused-resilience-and-speed
Port of Rotterdam Authority, Port Call Optimisation, general operational context. https://www.portofrotterdam.com/en/port-future/digitisation/improving-logistics-processes/port-call-optimisation
Research calculations: Rokin Research, When Does One More Harbour Pilot Matter?, technical report dated 30 October 2026, §§3–8 and Appendix A; accompanying pilotage workbook v0.2 and independent deterministic replication. The article's modelled quantities are not empirical observations from the cited notices.