What a 10% Opening Would Change

The incomplete European Banking Union can now be tested in the model rather than described only through observed gaps. This chapter examines the paper’s benchmark experiment: reduce cross-border relationship frictions by 10% while leaving the existing branch and subsidiary network in place. The useful judgment is not whether the resulting GDP number is a forecast, but what the experiment reveals about the channel of integration—more capital and labor made possible by broader access, rather than a large wave of efficient capital reallocation.

The experiment changes one barrier

The counterfactual lowers the cross-border relationship wedge across all euro-area country pairs by 10%. In the model, that is equivalent to increasing the bank-firm relationship shifter for foreign matches by 10%. The change makes foreign relationships more viable, but it does not create new bank branches, subsidiaries, or banking groups. The network of operating institutions is held fixed.

That boundary is essential. The experiment asks what happens when firms can connect more easily to foreign banks that are already represented in the model’s network. It does not ask what happens when regulators also change the incentives for groups to enter, choose a different legal form, or build new local capacity. Nor does a 10% reduction in a model wedge mean that cross-border lending rises by 10% or that GDP rises by 10%. The size of the wedge, the response of lender choice, and the general-equilibrium response are different objects.

The experiment is designed to isolate the relationship-formation channel identified in the data. If the main loss occurs because firms and banks fail to connect, reducing that loss should change the firm’s set of effective financing options. The model then recalculates bank lending, firm inputs, wages, deposits, wholesale funding, and output until the markets clear again.

stateDiagram-v2 state "Baseline euro-area banking network" as Baseline state "10% lower cross-border relationship wedges" as LowerWedges state "Broader realized lender access" as Access state "Lower effective user cost of capital" as UserCost state "Higher capital and labor demand" as Inputs state "Higher aggregate output" as Output state "Uneven gains and weak allocative convergence" as Distribution [*] --> Baseline Baseline --> LowerWedges : Apply the counterfactual LowerWedges --> Access : Existing network stays fixed Access --> UserCost UserCost --> Inputs Inputs --> Output Output --> Distribution

The states emphasize two facts at once. The model opens relationships without rebuilding the bank network, and the response passes through the effective user cost of capital before reaching output. The final state is not “all countries converge.” Aggregate growth and convergence are separate outcomes, and the paper finds that they do not move together in a simple way.

The growth comes from more inputs, not mostly better sorting

The benchmark result is a 1.6% increase in euro-area GDP. Capital rises by 2.4%, labor by 1.6%, and total factor productivity by only 0.06%. The foreign share of bank credit increases by about 1.2 percentage points. These figures are changes relative to the model’s baseline, not annual returns, forecasts, or estimates of an enacted reform.

Read together, the numbers describe an expansion of the production base rather than a one-for-one transfer of existing output.