AI Disruption Script: Developed Economies Bear Greater Damage, Long-term Blind Spots Remain in Market

Overview

This Bernstein research report aims to correct common market misconceptions about AI's disruptive impact. The mainstream view holds that emerging markets relying on low-cost labor arbitrage (such as India and Indonesia) will suffer the most severe shocks, but the research uses employment structure, salary differentials, and national income data to demonstrate that high-income developed economies are actually the most damaged in the long run. The report points out that large-scale AI adoption will trigger a global demand collapse similar to the prisoner's dilemma, but at the same time, AI could become an 'equalizer' that narrows wealth gaps between countries, albeit at the cost of extreme wealth concentration toward the top 1% within each country, with mass consumption economy facing collapse, ultimately inevitably provoking strong government regulation.

Core views

Structural differences that defy common sense: The market generally believes knowledge-intensive industries will be hit first, so low-cost labor countries face the greatest threat. However, data reveals that information and knowledge industry employment accounts for 6-10% in high-income countries, compared to only 1-4% in middle-income countries; service industry contribution to GDP exceeds 70% in developed countries, while in countries like India it is only about 50-55%. This means developed economies have higher structural exposure to AI. Economic logic of wage substitution: From a salary perspective, the cost savings from replacing an Indian software engineer earning $10,000 annually are minimal, while replacing a U.S. engineer earning $130,000 represents transformative-level gains. If a 'human-machine collaboration' model is introduced, retaining a $40,000 Indian engineer is more economical than retaining a $200,000 U.S. engineer. Therefore, white-collar positions in high-income countries have stronger substitution logic support. Market myopia and paradox: Despite fundamentals showing greater long-term risks for developed economies, emerging markets like India and Indonesia have performed worst in the AI wave, while countries like the U.S. and South Korea have surged. The research explains this stems from market short-sightedness: emerging markets lack pure beneficiaries in the AI supply chain, and broad-based indices have higher IT services weight (about 10%), with short-term damage expectations being overly amplified, while long-term macro demand collapse has not been priced in. Demand collapse in the prisoner's dilemma: Unlike previous industrial revolutions, AI aims to substitute cognitive tasks rather than expand new domains. The report traces how companies pursuing cost reduction compete to adopt AI, with initial productivity and demand rising in tandem; but after crossing a critical point, the number of people displaced exceeds new jobs created, and demand begins to decline. Individual companies cannot stop adoption (or they will be eliminated in competition), ultimately forming a 'Pareto-inefficient Nash equilibrium' — rising productivity with shrinking demand, sounding the death knell for the global economy. Restructuring of the income pyramid and 'equalizer' effect: AI will erode about 31% of national income in middle-income countries (the 9% share excluding the top 1% from the top 10%), but erosion in high-income countries could exceed 27% (due to higher knowledge industry share). Estimates show AI shock could reduce national income by approximately 10% in middle-income countries and about 22% in high-income countries. Ironically, while AI intensifies domestic wealth concentration toward the top 1%, it may actually narrow income gaps between rich and poor countries, becoming a 'great equalizer'. However, the 'upgrading' path for poor countries is completely locked, while rich countries lose substantial existing wealth. Consumption shock and government regulatory endgame: The top 1% will capture most of the national income, while the remaining 99% become 'more equal' in their poverty internally. This will deal devastating blows to mass discretionary consumption brands that depend on volume, with only ultra-luxury brands serving the top 1% likely to survive. The research asserts that AI will ultimately become a heavily regulated industry, with governments inevitably intervening to introduce Universal Basic Income (UBI), minimum employment guarantees, or AI taxes. In an endgame where wealth rather than income determines consumption, high-income countries will suffer far greater consumption damage than middle-income countries due to their higher share of consuming households.

Analysis framework

The research adopts a contrarian framework centered on 'penetrating short-term market pricing to find long-term structural blind spots.' First, using cross-country macro data (employment share, GDP contribution) and micro data (absolute salary comparisons), it deconstructs the common-sense fallacy that 'low-cost countries suffer the most,' establishing the baseline judgment that developed economies have higher exposure. Second, introducing a game theory perspective (Nash equilibrium/prisoner's dilemma), it traces the dynamic macro path of AI transitioning from 'productivity booster' to 'demand destroyer,' noting that technology adoption spontaneously slides into a Pareto-inefficient equilibrium. Finally, based on income distribution pyramid data, it quantifies potential national income shock magnitudes across different income tiers and countries (10% vs 22%), deriving the paradox of AI as an 'inter-country equalizer' while being a 'domestic concentrator,' and据此预判了政府监管必然介入的终局形态。

Methodology notes

• Event Game Theory & Behavioral FinanceGame Theory Analysis

  Prisoner's Dilemma and Pareto-inefficient Nash Equilibrium

  The research uses game theory to explain AI adoption: companies compete to adopt AI for cost reduction, even if total demand contracts they cannot unilaterally exit, ultimately falling into a Nash equilibrium where everyone is worse off but cannot escape, revealing how spontaneous technology expansion can trigger systemic demand collapse.

• Industry/Industrial Analysis FrameworkSupply-demand framework

  Decoupling of Productivity and Demand (AI Paradox)

  Traditional technologies enhance productivity and create new demand, but AI substituting cognitive roles causes the demand curve to begin declining before productivity peaks, breaking the common wisdom of synchronized productivity and demand growth.

• Macroeconomic framework

  Income Distribution Pyramid and Macro Shock Quantification

  The research disaggregates national income by tiers such as top 1% and top 10%, and combines with AI's impact ratios on different industry employment to estimate potential national income decline levels for countries at various development stages, an effective method for evaluating macro-structural shocks.

• Industry/Industrial Analysis FrameworkVolume-price decomposition

  Wage Arbitrage and Marginal Substitution Returns

  The research notes that AI substitution's economic driver lies in absolute salary differentials: substituting low salaries ($10,000) has low marginal returns, while substituting high salaries (over $100,000) has high returns. This determines stronger substitution logic for positions in high-income countries, a key dimension for assessing AI shock geographic distribution.

• Event Game Theory & Behavioral FinanceExpectation Gap/Expectation Management

  Market Myopia and Long-term Expectation Gap

  Market pricing only reflects the next few quarters (lack of AI beneficiaries, high IT weight), while ignoring decades of macro disruption ahead (demand collapse, income restructuring). This time-dimension expectation gap creates potential opportunities or blind spots.

Key data

• High-income Country Information/Knowledge Industry Employment Share6-10%Far higher than middle-income countries' 1-4%, indicating higher exposure to disruption
• High-income Country Service Industry GDP Contribution>70%Middle-income country median is only 55%, developed economies have deeper service dependency
• Indian Software Engineer Average Annual SalaryAbout $10,000Compared to approximately $130,000 in the U.S. and $80,000 in the UK, replacing high-salary positions has far greater marginal returns than low-salary
• Middle-income Country Broad-based Index IT WeightAbout 10%Compared to 2.9% in high-income countries, showing obvious short-term market drag effect
• Estimated National Income Decline from AI in High-income CountriesAbout 22%Compared to approximately 10% decline in middle-income countries, developed economies suffer larger absolute long-term damage
• Middle-income Country Top 10% Income ShareOver 50%Top 3 countries (South Africa, Brazil, India) reach as high as 60%, with higher inequality but lower knowledge position share
• Top 1-10% Population Income Share in Middle-income Countries31%Compared to 27% in high-income countries, this group suffers most from AI erosion
• High-income vs Middle-income Country Top 10% Wealth ShareAbout 64%Similar across both after removing extreme values, establishing comparison basis for long-term consumption shock

Impact & implications

The research believes the current extreme pessimism toward IT services and other industries represents short-term overpricing. In the long run, these industry punishments should be distributed more broadly across economies. For high-income countries, since their wealth distribution is similar to middle-income countries and they have higher share of consuming households, their damage in mass discretionary consumption will be more severe, with only ultra-luxury brands serving the top 1% likely to survive. For middle-income countries, although relative income gaps may narrow, the path to 'upgrading' to high-income status is completely locked, and the labor arbitrage model ends. Ultimately, AI will trigger a global demand collapse, inevitably provoking strong government intervention (such as UBI, AI taxes, minimum employment), which will become a defining characteristic of the full AI penetration phase.

Risks

• Early strong government intervention (such as UBI, AI taxes) interrupts AI's natural expansion path
• AI adoption speed or substitution effects fall below expectations, delaying the demand collapse critical point
• Emerging markets successfully transform and find new growth models, breaking free from labor arbitrage dependence

What to watch

• AI-related regulatory policies and legislative developments in various countries (especially minimum employment and AI taxes)
• Developed economy knowledge industry employment data and wage growth stagnation signals
• Emerging market IT service industry transformation to AI services with real changes in profitability and revenue