The Weak Yen’s Missing Factory Boom
This series studies Japan’s 2021–22 yen depreciation through the relationship between Japanese multinational parents and their foreign affiliates. Its central question is why a weaker yen can improve Japanese firms’ foreign earnings without bringing a broad wave of production back to Japan. To answer it, the reader needs more than the usual exchange-rate story: they need to distinguish where goods are produced, where income accrues, and which part of a firm’s capability is being expanded or redirected.
The puzzle: a weaker yen, but not a trade-led revival
The standard intuition is straightforward. When a country’s currency depreciates, its goods become cheaper in foreign-currency terms, while imported goods become more expensive at home. Foreign buyers should demand more of the country’s exports, and domestic buyers should switch toward goods made at home. Production in the home economy should rise, improving the trade balance. That is the familiar expenditure-switching channel.
Japan’s experience in 2021–22 does not fit that story neatly. The yen depreciated sharply, yet Japan’s trade balance deteriorated rather than improving. At the same time, the country’s direct-investment-income surplus rose from about 2 percent of GDP in 2020 to about 4 percent in 2022. That income increase partly limited the decline in the current account. In other words, an important part of Japan’s external position was supported by income connected to production abroad, not by a stronger balance of goods trade.
This is not a minor accounting curiosity. Direct-investment income has become a large external flow for Japan. In the longer series examined by the Federal Reserve study, direct-investment-income credits approach 5 percent of GDP. They include distributed earnings, reinvested earnings, and interest on intercompany debt, so they are broader than cash remittances from foreign subsidiaries. They nevertheless provide the balance-of-payments measure most closely connected to the income generated by Japanese-owned foreign affiliates. Recent changes in Japan’s direct-investment-income balance were also large relative to changes in its overall current account, reaching about 0.4 of the latter’s volatility in the paper’s rolling measure.
It is useful to separate three statements that are often collapsed into one. A direct-investment-income credit says that income from an outward investment accrues to residents. A direct-investment-income surplus compares such credits with related debits paid to foreign owners. The current account adds that income balance to the trade balance and other flows. Thus, an improvement in the current account can coexist with a worsening trade balance if the income component improves enough. None of these measures, by itself, says that factories in Japan produced more goods. They answer different questions about cross-border production, ownership, and income.
The observation does not prove that the depreciation caused every movement in these accounts. The 2021–22 period also included pandemic disruptions, supply-chain bottlenecks, changing demand, and other shocks. But it does expose a gap in the simple mental model. If a weaker currency mainly worked by making goods produced in Japan more competitive, why did the trade balance worsen while income from foreign operations became more important?
The missing possibility is that Japanese firms do not serve foreign customers in only one way. A Japanese company may export from a factory in Japan, or it may produce through an affiliate located in the market it wants to serve. The exchange rate changes the value of both activities to the Japanese parent, but the two activities reach different parts of the Japanese economy.
One firm can span two production locations
Consider a Japanese parent with a factory at home and an affiliate overseas. If the parent sells goods made in Japan to foreign customers, the production employs resources in Japan, adds to Japanese GDP, and produces an export receipt. The foreign customer’s payment enters the trade balance. A depreciation can strengthen this route if foreign demand responds to lower foreign-currency prices for Japanese goods.
Now consider the same parent serving foreign customers through its affiliate abroad. The affiliate’s workers, premises, and production are located in the foreign economy. The output therefore does not enter Japan’s GDP, even if the affiliate is fully owned by a Japanese company and even if the parent helped organize or finance it. The affiliate may generate operating income that accrues to the Japanese parent. That income enters Japan’s direct-investment-income balance and contributes to gross national income, but it is not a Japanese export and does not become production located in Japan.
GDP and gross national income answer different questions. GDP asks how much production takes place within a country’s borders. GNI asks how much income accrues to the country’s residents, including income from assets and business relationships abroad. A Japanese-owned factory in Thailand contributes its physical production to Thailand’s GDP, while part of the income that reaches its Japanese parent can contribute to Japan’s GNI. Confusing these measures makes foreign expansion look like domestic expansion when it is not.
A simple thought experiment makes the distinction concrete. Suppose a Japanese parent can supply a foreign customer either with 100 units made in Japan or with 100 units made by its affiliate abroad. In the first case, the relevant production is inside Japan and the foreign sale is an export. In the second, the sale is served by foreign production; Japan may receive parent income, but the 100 units do not become Japanese GDP. The parent can therefore become more successful in serving the foreign customer without the location of the customer’s production changing in Japan’s favor. Ownership links the two operations financially and organizationally; it does not move the foreign affiliate’s workers or output across a border.
The same distinction appears in the external accounts. Japan’s current account can be written as the trade balance plus the direct-investment-income balance and other components. A country can therefore see its current account supported by income from foreign production even when its trade balance is weak. That outcome may be valuable to Japanese residents: parents can receive profits, raise wages, or expand commercial functions at home. But those benefits should not be described as a return of the foreign affiliate’s production to Japan.
There is an important converse case. A foreign multinational’s affiliate that produces inside Japan contributes to Japanese GDP because the production occurs in Japan, even though income paid to the foreign parent is recorded as a direct-investment-income debit. Ownership and production location are related, but they are not the same variable. The question in Japan’s yen episode is specifically about Japanese parents and their affiliates abroad: where their foreign-market activity occurs, and how the resulting income flows back.
This distinction also explains why a larger direct-investment-income surplus is not automatically equivalent to a stronger current-account improvement of the same size. If foreign demand is served through exports, receipts enter the trade balance directly. If it is served through an affiliate, only the portion of foreign operating income that accrues to the Japanese parent enters direct-investment income. The two channels can both benefit a Japanese firm, but they do not have the same effect on domestic production or on the external accounts.
The question is where the yen sends the firm’s next unit of effort
The central issue is therefore not whether depreciation makes Japanese firms richer in yen terms. It can. The more revealing question is what firms do with the higher home-currency value of their foreign activity. If a Japanese company already has a network of affiliates abroad, a weaker yen may make serving those markets through the network more attractive. The firm may expand the overseas operation, send more workers to it, or increase the parent-side coordination needed to support it. None of those responses requires a factory abroad to move back to Japan.
An overseas network also means that the same yen movement does not affect every Japanese parent in the same way. One firm may have affiliates whose sales are concentrated in markets whose currencies rise strongly against the yen. Another may have a similar-looking affiliate in a market whose currency moves less. Their exposure depends on the network’s geography and on how each affiliate earns and spends money. That makes the location of an existing operation economically important before any new investment or relocation decision occurs.
The first distinction is between revenue and cost. A foreign affiliate that earns revenue in a currency appreciating against the yen becomes more valuable when its results are translated into yen. An affiliate that buys many foreign inputs can also face a cost effect. These are not identical forces, and the Japanese evidence treats them separately. The broader point is that an exchange rate reaches a multinational through the structure it already has: its customers, suppliers, affiliates, and parent-side capabilities. It does not begin with a blank map on which every factory can be redrawn.
That possibility changes the meaning of “bringing production home.” Reshoring requires a change in the location of production, not merely a change in the ownership income associated with production. It may also require new domestic machinery, domestic production labor, and enough organizational capacity to operate the plant. A currency movement can raise the return to a foreign-market opportunity without making those domestic conditions the firm’s most valuable use of scarce resources.
The distinction matters for policy and for interpretation. A later IMF assessment describes Japan as facing robust domestic demand and low unemployment alongside high public debt, an aging population, persistent cost-of-living concerns, and labor-market pressure. Those conditions do not determine how the exchange-rate mechanism works, but they remind us that a weak yen arrives in an economy with existing constraints and competing demands. The same currency movement can raise foreign-affiliate income while failing to solve a shortage of domestic labor, productive capacity, or investment.
The strongest claim supported by the Japanese evidence is conditional, not absolute. A weaker yen does not necessarily bring production home when Japanese firms already serve foreign demand through affiliates and when parent-side capabilities are limited or costly to expand. Nor does the evidence say that domestic output must fall after every depreciation. Expenditure switching, domestic demand, and the forces that caused the currency movement can push in the opposite direction.
The practical test is to follow both the money and the work. If only yen-denominated affiliate profits rise, translation may explain much of the change. If affiliate employment and workers dispatched from Japan also rise, the foreign operation is expanding in a more concrete sense. If Japanese parents raise support employment and wages while total employment remains flat, the home response may be reorganization rather than a general production boom. That is the empirical pattern the rest of this series examines.