Air transport is central to global connectivity, but also a growing source of carbon emissions. Existing evidence shows that international trade and transport affect emissions through composition, efficiency, and scale effects. Indeed, changes in transport costs alter both the organisation and the volume of international flows (Cristea et al. 2013, Hummels et al. 2011). In the aviation sector, one of the drivers of transport costs, beyond the oil price and carbon pricing (Fageda and Teixidó 2022), is regulatory restrictions on market access and connectivity (Piermartini and Rousová 2013).

Among these regulatory restrictions, a key but rather overlooked policy is the signing of air service agreements (ASAs). By liberalising market access — through provisions on entry, capacity, and pricing — air service agreements reshape airline networks and the allocation of traffic across routes and carriers and by the same token carbon emissions embedded in international flights.

In a recent study (Fontagné et al. 2026), we assess the implications of the signing of air service agreements on emissions, drawing on detailed global data on flight bookings and flight networks for the period 2012-2019. We show that air service agreements significantly improve the environmental efficiency of air transport by reducing emissions per passenger: airlines reorganise international routes, serve more passengers, operate more direct connections, and hence reduce the amount of carbon emissions per passenger. However, shorter and more comfortable air routes (i.e. direct rather than connecting flights) generate increases in demand for flights that, at least partially, offset the benefit associated with reductions in emissions per passenger. By considering the complex direct and indirect effects associated with the signing of air service agreements (ASAs), we are ultimately able to highlight the trade-off between scale and the operational efficiency of airlines.

Figure 1 documents the expansion of fully liberalised air service agreements and the corresponding evolution of passenger traffic. Over 2012-2019, the number of country pairs covered by such agreements increased substantially, and passenger flows on these routes grew significantly faster than on other routes. This pattern indicates that the content of agreements — particularly provisions on capacity, pricing, and market access — plays a central role in shaping air transport outcomes. Routes covered by more liberalised agreements experience stronger growth in traffic, making air service agreements a key policy lever affecting both the scale of air transport and the organisation of flight networks. This provides the basis for analysing their impact on route structure and emissions.

Figure 1 Fully liberalised air service agreements and passengers

Source: Fontagné et al. (2026), based on SABRE ticket data and ICAO air service agreements.
Note: The figure reports, for 2012–2019, the evolution of country pairs covered by Air Service Agreements (ASAs), distinguishing between fully liberalised and not fully liberalised agreements, together with the corresponding evolution of international passenger traffic. Fully liberalised air service agreements include agreements with unrestricted traffic rights, multiple designation, free capacity setting, and flexible pricing provisions.

Reorganisation of international aviation routes

What are the main direct effects of the signing of air service agreements on flight traffic, route organisation, and emissions? The results point to two distinct mechanisms operating in opposite directions: Air service agreements help to increase the number of passengers and reduce the number of stopovers required to travel from the country of departure to the country of destination. Based on our estimates of the impact of air service agreements over the 2012–2019 period — obtained using the heterogeneity-robust nonlinear difference-in-differences estimator of Wooldridge (2023) and implemented within a gravity framework following Nagengast and Yotov (2025) — the number of passengers increases by 12.7%, while the average distance decreases by about 1.4% and the number of flight segments by around 3.9%, resulting in a comparable reduction in emissions per passenger. These changes are consistent with a reorganisation of routes towards more direct connections, reducing fuel-intensive take-off and landing phases and improving overall network efficiency. Taken together, these estimates highlight a clear trade-off. Air service liberalisation improves the efficiency of international flight connectivity but expands the overall scale of air transport. In aggregate, and considering only direct effects, the increase in traffic dominates, so that total emissions rise despite measurable gains in per-passenger efficiency.

The impact of air service agreements in terms of reduction in emissions concentrates on large and high-performing carriers. For airlines in the top segment — whether defined by revenue (top 25) or service quality (Skytrax top 50) — air service agreements lead to a decline in total emissions of about 3%, along with a comparable reduction in emissions per passenger. By contrast, there is no statistically significant effect for other airlines.

These findings point to a composition effect. Liberalisation reallocates traffic towards airlines that operate more efficient networks and aircraft and are better able to reorganise routes and reduce stopovers. As a result, part of the overall efficiency gains associated with air service agreements reflects a shift in market shares towards more efficient carriers.

Dynamic effect of air service liberalisation

A dynamic analysis confirms these findings. Figure 2 shows the dynamic effects of air service agreements around their entry into force. Pre-treatment coefficients are close to zero and do not display systematic trends, providing support for the causal interpretation of our results (i.e. absence of differential pre-trends between treated and control routes).

Following the signing of air service agreements, all outcomes adjust in a consistent direction. Distance and the number of flight segments decline, indicating a reorganisation of routes towards more direct connections. These changes translate into a reduction in emissions per passenger, which becomes more pronounced over time as airlines progressively adjust their networks. By contrast, total emissions increase after the agreement, reflecting the expansion in passenger traffic and the gradual scaling-up of operations on liberalised routes.

Overall, the dynamic patterns confirm the coexistence of three effects: a technical effect — improving efficiency at the route level; a composition effect — concentrating gains among more efficient airlines; and a scale effect — increasing total traffic and emissions. These effects emerge shortly after liberalisation and strengthen over time, consistent with gradual adjustments in network structure and demand responses.

Figure 2 Dynamic effects of air service agreements on route structure and emissions

Source: Fontagné et al. (2026).
Note: The figure reports event-study estimates of the effects of Air Service Agreements (ASAs) on emissions per passenger (EP), total emissions (passenger-constrained), average distance, and number of legs. Estimates are obtained with the heterogeneity-robust nonlinear difference-in-differences estimator of Wooldridge (2023), implemented in a gravity setting following Nagengast and Yotov (2025). The control group includes country pairs that are not yet treated and country pairs that remain untreated throughout the sample period. All specifications include origin-year, destination-year, and country-pair fixed effects. Coefficients are reported as percentage changes. Red markers denote pre-treatment coefficients, while blue markers denote post-treatment effects. Shaded areas represent confidence intervals. Standard errors are two-way clustered by origin and destination country. ***, ** denote statistical significance at the 1% and 5% levels.

Quantification of the net effect of air service agreements

To assess the aggregate implications of these effects, we employ a structural gravity model akin to a full endowment general equilibrium framework in the trade literature, calibrated using the elasticities (i.e. estimation results) discussed above. The model incorporates the estimated changes in route structure and transport costs, allowing for endogenous adjustments in demand, network organisation and average prices across all country pairs. The results confirm that the scale effect dominates once system-wide responses are taken into account.

In a counterfactual exercise in which we turn off the existing network of air service agreements, international passenger traffic is found to be 2.8% lower and total emissions 0.9% lower, implying that the current network of agreements already raises emissions despite improving efficiency. Extending air service agreements to all country pairs would increase passenger traffic by 2.3% and total emissions by 1.2%, while reducing emissions per passenger by about 1.1%. Finally, a fully liberalised aviation network in which all country pairs share full-liberalisation air service agreements, would generate the strongest effects: passenger traffic increases by 4.0%, total emissions rise by 1.7%, and emissions per passenger decline by 2.3%.

These findings have direct implications for transport and climate policy. Although air service agreements are designed to enhance connectivity and competition, their environmental impact ultimately depends on the balance between efficiency gains, composition effects, and the expansion of traffic. The bottom line of our exercise is that liberalisation alone, although beneficial for the passengers and potentially in terms of activity generated at origin and destination, is unlikely to reduce total emissions in aviation, even if it improves efficiency at the route level and lowers emissions per passenger.

This implies that policies regulating access to the aviation market must be complemented by first-best instruments that directly target emissions, such as carbon pricing, mandates for sustainable aviation fuels, or incentives for fleet modernisation and the adoption of more fuel-efficient aircraft. More generally, our findings highlight the importance of jointly considering both intensive and extensive margins when assessing the environmental impacts of trade and transport policies, as efficiency gains may be offset by an increase in scale.

References

Cristea, A, D Hummels, L Puzzello and M Avetisyan (2013), “Trade and the greenhouse gas emissions from international freight transport”, Journal of Environmental Economics and Management 65(1): 153–173.

Fageda, X and J Teixidó (2022), “Pricing carbon in aviation: Evidence from the European Emissions Trading System”, Journal of Environmental Economics and Management 111: 102591.

Fontagné, L, C Mitaritonna, G Orefice and G Santoni (2026), “Air Service Agreements, Connectivity and Emissions”, CEPII Working Paper 2026-04.

Hummels, D, L Puzzello and A Cristea (2011), “Trade and greenhouse-gas emissions: How important is international transport?”, VoxEU.org, 17 December.

Nagengast, A J and Y V Yotov (2025), “Staggered difference-in-differences in gravity settings: Revisiting the effects of trade agreements”, American Economic Journal: Applied Economics 17(1): 271–296.

Piermartini, R and L Rousová (2013), “The Sky Is Not Flat: How Discriminatory Is the Access to International Air Services?”, American Economic Journal: Economic Policy 5(3): 287–319.

Wooldridge, J M (2023), “Simple approaches to nonlinear difference-in-differences with panel data”, The Econometrics Journal 26(3): C31–C66.