Toyota #1 globally ยท FANUC 38% of robot production ยท Canon 98% of cameras
Diamond firing on all cylinders โ rivalry, demanding buyers, deep clusters
Global-facing, rapid product iteration, high innovation rate
๐ The Sheltered Economy
Protected & inefficient: agriculture, retail, construction โ cartels, zero global presence
The Retail Bottleneck: Large Store Law blocked efficient retailers for decades
The Danger Signal: protectionism breeds stagnation โ drag on living standards
EVs: 1.35% domestic adoption โ lowest in the developed world
"Porter wrote this warning in 1990. Thirty-five years later โ which Japan won? And what does that tell us about whether the diamond is still the right lens for 2025?"
Porter, M.E. (1990). Ch. 8; Motor Authority 2025; IFR World Robotics 2024.
๐ OPTIONAL โ INTERACTIVE SCENARIO ยท ~3โ4 MIN
What Would MITI Do?
The Scenario โ Japan, 1975
A Japanese firm wants to manufacture memory chips. Engineers: yes. Capital: no. Intel and Texas Instruments already dominate globally.
YOU ARE MITI. What do you do?
OPTION A
Direct subsidy + import tariffs to shield from foreign competition
OPTION B
Cooperative research โ pool basic R&D, preserve product rivalry
OPTION C
Channel low-interest capital through postal savings system
OPTION D
Let the market decide โ no intervention
What MITI Actually Did โ A, B and C. Never D.
โ
VLSI cooperative research (1976โ1979) โ ยฅ72B joint R&D across NEC, Hitachi, Fujitsu, Mitsubishi, Toshiba. Signalled all firms to invest heavily in private R&D alongside it
โ
Postal savings capital deployed โ low-interest loans via Japan Development Bank. Patient capital unavailable on open markets
โ
Domestic rivalry preserved โ MITI did NOT consolidate firms. All five competed ferociously. By 1985: Japan held 50% of global DRAM market
The Lesson
MITI's role was to signal, fund basic research, direct capital โ not pick winners or suppress rivalry. When MITI tried to consolidate firms, companies ignored them. Rivalry survived. That was the real advantage.
The Question for Today
Japan is spending $11.4B on Rapidus to build 2nm chips by 2027. Same MITI logic โ different era. Will it work when TSMC took 40 years to build its process lead?
Porter, M.E. (1990). Ch. 8 โ Role of Government; Brookings (2023).
Cluster depth = durable advantage. Cameras, robotics, machine tools all held 40 years later
โ
Selective disadvantages spur innovation. Oil shocks โ world-leading energy efficiency
โ
No rivalry = no global success. Agriculture, construction, retail โ zero international presence today
What Porter Missed โ
โ
Modularization breaks vertical integration. When industries split into horizontal layers, Japan's keiretsu became a cost burden โ DRAM, TVs, PCs all collapsed
โ
Demanding buyers became a trap. The Galรกpagos Syndrome โ Japan built world-class products for uniquely Japanese needs the world didn't share. Smartphones, EVs, electronics
The analytical point: Porter's diamond predicts which industries win within an existing structure. It does not predict what happens when the industry structure itself changes.
A keiretsu links companies through cross-shareholding. Not a cartel. A structure that moves information, capital, and components faster than any arm's-length market could manage.
Three Layers โ Each a Diamond Determinant
Bank + Trading Co. โ patient capital and global distribution. Factor conditions + supporting industries.