How Far Can the Evidence Take Us?

The Japanese parent–affiliate evidence supports a clear mechanism, but it does not turn every weak-currency episode into the same story. This chapter tests how far the explanation can travel: whether competing financial and input-cost channels fit the data, which results survive changes in modeling choices, how foreign demand changes the response, and what the cross-country evidence adds. The reader’s practical task is to recognize a plausible organizational interpretation without confusing it with a universal or fully causal law.

Competing explanations meet the Japanese data

A useful explanation must account for the pattern as a whole. More-exposed foreign affiliates gained income and employment. Payments to Japanese parents increased. At home, exposed parents raised average wages and non-production support employment, while other employment fell, total employment stayed broadly flat, and domestic fixed assets declined. A theory that explains only higher foreign profits is incomplete; a theory that predicts broad domestic hiring is inconsistent with the parent-side results.

One alternative is a financial-windfall channel. A weaker yen can increase the yen value of profits earned abroad. If parents repatriated that income and used it to overcome financing constraints, the domestic response should be stronger among firms that were initially more financially constrained. The paper measures initial leverage as total liabilities relative to assets and tests whether the revenue-side exposure effect changes with that position. The interaction is insignificant across the parent outcomes. More leveraged parents do not show a systematically stronger response in wages, employment composition, or fixed assets.

This does not prove that financial flows are irrelevant. Affiliate-to-parent payments do rise, and the income connection is part of the story. The result says something narrower: the observed parent-side reorganization is not especially concentrated among the firms for which a financial relaxation should have mattered most. The evidence fits a change in the value of foreign operations and the support needed to run them better than a simple balance-sheet windfall spent on domestic expansion.

A second alternative is a vertical input channel. If an affiliate buys a large share of its inputs from Japan, yen depreciation may reduce the foreign-currency cost of those Japanese inputs. The affiliate could then expand because it has become a cheaper downstream producer, not because the parent redirected scarce support toward it. The paper controls for purchase-side exchange-rate exposure and tests whether the revenue-side response is stronger when an affiliate’s purchases from Japan are a larger share of total purchases. The interaction is not positive across the tested outcomes; for profit, it is negative.

Again, the conclusion is conditional. Japanese inputs may matter for particular firms or products, but the data do not support that channel as the general explanation for why the revenue-side exposure is associated with higher affiliate activity and parent support. The revenue and purchase measures also move differently: greater purchase-side exposure generally reduces affiliate value added, employment, and dispatched workers. That contrast is useful because it shows that the results are not simply an undifferentiated response to any foreign-currency movement.

A third possibility is that support functions are readily scalable. If coordination and commercial capabilities could expand without opportunity cost, an affiliate might grow while domestic operations remained unaffected.