Two Japans in one economic snapshot
The brightest lights in Japan’s economy illuminate semiconductor supply chains, precision-equipment plants and the industrial companies taking orders from worldwide investment in computing. A different picture emerges on shopping streets where customers examine prices before spending. The October 6 Reuters Tankan captured the divide: manufacturers’ sentiment rose to +22 from +21 in September, while the nonmanufacturing index fell to +23 from +29. These are diffusion-style balances of company responses, not percentage changes in sales or production. They describe confidence—and expose two very different sources of demand.
The industrial economy rides a global technology cycle
Precision machinery led the surveyed manufacturing categories, gaining nine points to +38, Reuters reported. Semiconductor-related orders helped, while metal-products and steel respondents also showed improvement. The mechanism is broader than chips alone. Investment in data centers creates orders for factory tools, measurement systems, components, materials and industrial services. Japanese suppliers can benefit even when the customer building the new computing facility is overseas. Yet this is a cyclical demand story, not a guarantee that the entire factory sector is insulated from changes in trade, energy costs or a future slowdown in AI capital spending.
The less visible price squeeze
A business can sell more goods without feeling more profitable if ingredients, imported materials, electricity and wages rise even faster. Reuters reported a 15-point decline to -40 in the food-manufacturing category—a manufacturing subgroup, not part of its nonmanufacturing index. Weak household purchasing power made passing on costs difficult. The survey also recorded deteriorating sentiment in retail, information and communications, real estate and construction, while wholesalers benefited from reconstruction demand. That unevenness matters: “services slump” is too sweeping a description for sectors with different customer bases and cost structures.
An October 9 release makes the story more concrete
Japan’s Statistics Bureau provided unusually direct evidence of household restraint on October 9. In August, average consumption expenditure for households of two or more people was ¥310,975, 1.0% lower than a year earlier in nominal terms and 3.1% lower after inflation. Workers’ households received average monthly income of ¥616,704, up 1.3% nominally but down 0.9% in real purchasing-power terms. The monthly seasonally adjusted real spending measure actually rose 0.1%. The monthly rise and annual decline are not contradictory: they compare different periods and answer different questions.
Do not confuse two surveys called Tankan
The Reuters Tankan is a private poll: Reuters surveyed 508 businesses between September 18 and October 2 and received 215 responses. The Bank of Japan’s September Tankan, released October 1, surveyed 9,104 firms of different sizes and achieved a 99.4% response rate, with replies collected from August 26 to September 30. Among large companies, the BOJ manufacturing business-conditions DI rose from +22 in June to +24 in September, while nonmanufacturing slipped from +37 to +35. The broad contrast resembles the Reuters reading. The numbers cannot be stitched together into one time series or described as if they were the same sample.
Small firms complicate the easy narrative
The BOJ survey also found improvement among medium-sized manufacturers, whose DI climbed from +17 to +23, and small manufacturers, rising from +9 to +14. For nonmanufacturers, the medium-company reading declined from +26 to +24; smaller firms stayed at +15. The shortage of workers remained a powerful constraint: the all-industry employment-conditions DI stood at -38, meaning businesses reporting too few workers outnumbered those reporting too many. A country can have weak consumer demand in some sectors and acute labor scarcity in others. The two conditions coexist because of demographics, location, skills and the way demand is distributed.
How Japan arrived at this point
Following the collapse of its asset bubble, Japan spent decades wrestling with stagnant demand, persistent price weakness and unusually low interest rates. The BOJ expanded monetary easing dramatically from 2013 and ended negative interest rates in 2024. By 2026, many businesses instead faced a combination of more expensive imports, shortages of workers, higher financing costs and pressure to raise pay. A shift from deflation to inflation does not automatically translate into greater prosperity. The crucial question is whether productivity and nominal compensation grow fast enough to improve real household purchasing power.
The missing link between industrial orders and household demand
An equipment maker serving overseas semiconductor clients and a restaurant serving a neighborhood earn their revenue from very different engines. The first depends on global investment plans and may gain or lose from currency swings. The second depends on local disposable income and how much customers are willing to pay for meals after housing, transport and utility bills. Workers can contribute to export successes and still feel poorer if the gains are concentrated in particular firms, occupations or regions. That gap helps explain why a favorable business-conditions survey need not produce a stronger consumer mood.
What executives and investors should actually monitor
Rather than treat Japan as one simple recovery story, analysts should separate exporters from domestically oriented firms and trace exposure to AI capital spending, dollar-denominated imports, borrowing costs and the ability to pass on higher prices. For manufacturers, order backlogs, utilization, customer concentration and overseas sales matter. For domestic services, real wages, household budgets, foot traffic and sales volumes offer a stronger reality check. Business-sentiment data are useful leading indicators, but optimism does not prove profits have already been earned.
What would close the divide?
A sustainable expansion would require international industrial demand to translate more broadly into domestic investment, wages and spending without losing gains to inflation. The October Reuters poll anticipated modest further improvement among manufacturers and weaker sentiment for nonmanufacturers, but those forecasts remain opinions about the future. Upcoming family-budget, employment, retail and corporate-earnings releases will test them. Japan’s factories and its consumers need not move in opposite directions forever. For now, the most revealing fact is that they are not moving together.