Author: Denis Avetisyan
A new macroeconomic model suggests that widespread AI adoption, while boosting productivity, could trigger demand shortfalls if labor displacement outpaces the creation of new opportunities.
This paper models a potential macroeconomic crisis stemming from rapid AI adoption, finding that concentrated consumption and financial intermediation risks require proactive fiscal policies and task reinstatement to mitigate demand deficiencies.
Despite unprecedented technological potential, rapid advances in artificial intelligence may paradoxically generate macroeconomic instability. This paper, ‘Abundant Intelligence and Deficient Demand: A Macro-Financial Stress Test of Rapid AI Adoption’, formalizes a macro-financial stress test revealing a core tension: AI-driven abundance can coexist with demand deficiency due to institutional anchors in human cognitive scarcity. We demonstrate that displacement spirals, declining monetary velocity-what we term ‘Ghost GDP’-and collapsing intermediation margins, particularly impacting private credit and mortgage markets, create conditions for explosive crisis unless offset by proactive fiscal policies and task reinstatement. Can preemptive adjustments to economic institutions mitigate these risks and ensure broadly shared gains from the ongoing AI revolution?
The Shifting Sands of Economic Reality
The economic pulse appears to be weakening, as evidenced by a significant decline in the velocity of money – a measure of how quickly funds change hands within the economy. Since 1997, this velocity has fallen by a concerning 36%, indicating that each dollar is circulating less frequently than before. This isn’t merely a temporary fluctuation tied to economic cycles; the sustained decrease suggests deeper structural problems may be at play. A lower velocity can mask underlying economic weakness, as overall demand appears stable even if individual transactions are slowing down. Consequently, standard economic indicators may fail to accurately reflect the true health of the economy, potentially leading to delayed or inadequate policy responses to emerging stagnation.
Economic deceleration isn’t merely a temporary fluctuation within the business cycle, but reflects a profound alteration in how wealth is distributed. Since 1960, the portion of national income flowing to labor has diminished by 19%, signaling a consistent shift of gains towards capital – investments, profits, and asset ownership. This phenomenon creates what is termed ‘Ghost GDP’, where economic figures appear robust due to increased corporate earnings and investment, yet fail to translate into widespread purchasing power. Consequently, even with positive GDP growth, consumer demand can stagnate, as the majority of new wealth concentrates within a smaller segment of the population, ultimately hindering broad-based economic expansion and potentially masking underlying structural vulnerabilities.
A concerning trend indicates that the increasing prevalence of AI-driven automation is not simply reshaping the job market, but potentially instigating a ‘Displacement Spiral’ with significant economic ramifications. Analysis reveals that occupations most susceptible to AI exposure have already experienced a demonstrable decline in wage growth – a decrease of 6.53% between 2019 and 2023. This isn’t an isolated incident; as automation displaces workers, overall demand weakens, leading to further job losses and a contraction of economic activity. The cycle reinforces itself, creating a self-perpetuating downturn where technological advancement, rather than driving prosperity, contributes to prolonged economic stagnation. This suggests that mitigating the negative consequences of AI-driven displacement requires proactive strategies to bolster demand and ensure a more equitable distribution of economic gains.
Pathways to Resilience: Breaking the Feedback Loop
Direct fiscal transfer policies involve the distribution of financial resources directly to individuals or households, typically through mechanisms like universal basic income, expanded unemployment benefits, or stimulus checks. These policies aim to mitigate the economic impact of displacement – whether due to automation, economic shocks, or other factors – by maintaining aggregate demand. When a significant portion of the population experiences income loss, overall consumption declines, potentially leading to a recessionary spiral. Fiscal transfers counteract this by providing disposable income to those most affected, enabling continued spending on essential goods and services. The effectiveness of such policies is predicated on reaching individuals with a high marginal propensity to consume, meaning they are likely to spend the transferred funds rather than save them, thereby maximizing the stimulative effect on the economy.
Task reinstatement strategies are crucial for mitigating the negative economic consequences of AI-driven automation by focusing on the creation of new employment opportunities that leverage human-AI collaboration. Rather than attempting to compete directly with AI in performing existing tasks, these strategies prioritize identifying and developing roles where human skills complement AI capabilities. This includes jobs focused on AI training, data labeling, AI system maintenance, and roles requiring uniquely human skills such as complex problem-solving, critical thinking, and interpersonal communication. Successful task reinstatement requires investment in workforce retraining and education programs aligned with evolving labor market demands, ensuring displaced workers acquire the necessary skills to transition into these new, AI-complementary positions and maintain overall productive capacity.
Effective fiscal and employment interventions require consideration of consumption concentration, as a disproportionately large share of total demand is typically generated by a relatively small percentage of the population. Policies designed to bolster demand must, therefore, prioritize reaching these high-consumption households to maximize impact; broad-based programs, while equitable, may be less effective in preventing demand collapse than targeted support. Data indicates that the top 20% of income earners consistently account for approximately 40% of total consumption expenditure, a figure that can be significantly higher in certain economies; failing to account for this distribution risks diluting the efficacy of interventions and hindering overall economic recovery following displacement events.
Early Warning Signs: The System Speaks
The effectiveness of policy interventions is critically impacted by the time elapsed between problem identification and implementation, known as the Policy Response Lag. Research indicates that a lag exceeding two years, when combined with an inadequate magnitude of corrective transfers, can precipitate a substantial decline in the labor share of national income – potentially exceeding 40%. This signifies a significant shift in income distribution away from labor and towards capital, which can contribute to decreased aggregate demand and increased economic instability. The correlation between prolonged response times, transfer insufficiency, and labor share decline underscores the necessity for rapid and robust policy action in addressing emerging economic challenges.
A decreasing labor share, defined as the proportion of national income compensated to labor, serves as a critical indicator of shifts in economic power and potential macroeconomic instability. Historically, a declining labor share correlates with increasing capital concentration, where a larger percentage of national income accrues to owners of capital rather than workers. This shift can lead to reduced aggregate demand because labor income generally has a higher marginal propensity to consume than capital income. Consequently, a sustained decline in the labor share may signal an impending demand shortfall, potentially hindering economic growth and increasing the risk of recessionary pressures. Analysis demonstrates that prolonged reductions in labor share are often associated with increased income inequality and decreased economic resilience.
Private credit markets, encompassing lending activities outside of traditional banks, demonstrate significant sensitivity to disturbances originating in the mortgage market. This interconnectedness creates a pathway for localized shocks – such as declines in mortgage-backed security values or increased mortgage delinquency rates – to propagate throughout the broader financial system. The amplifying effect stems from the common investor base shared by both sectors, as well as the use of similar collateral and risk models. Consequently, a downturn in the mortgage market can trigger margin calls, forced asset sales, and reduced lending activity within private credit, exacerbating economic slowdowns and increasing the potential for widespread displacement effects, including job losses and business failures.
The implementation of robust early warning indicators is crucial for preemptive risk management within complex economic systems. These indicators, when consistently monitored, allow for the identification of developing imbalances – such as declining labor share or increasing vulnerability in private credit markets – before they contribute to systemic crises. Effective monitoring requires quantifiable metrics and predefined thresholds triggering investigative action or policy adjustments. The timeliness of response is paramount; a policy response lag exceeding two years, combined with insufficient financial transfer, can exacerbate negative outcomes, potentially resulting in a labor share decline of 40% or more. Consequently, continuous data collection, analytical capacity, and established protocols for intervention are essential components of a proactive risk mitigation strategy.
Governing the Inevitable: Shaping the Future
Effective AI governance centers not on hindering technological advancement, but on proactively shaping its course to maximize societal benefit. This requires a shift in perspective – viewing AI not merely as a tool for efficiency gains, but as a powerful force demanding careful stewardship. Rather than imposing restrictions that could stifle creativity, governance frameworks should prioritize the alignment of AI development with broader societal values, such as fairness, transparency, and accountability. This proactive approach involves fostering collaboration between researchers, policymakers, and the public to anticipate and mitigate potential harms, while simultaneously creating incentives for responsible innovation and ensuring that the advantages of artificial intelligence are distributed equitably across all segments of society. Ultimately, successful AI governance recognizes that guiding the technology’s trajectory is not an impediment to progress, but rather a crucial component of realizing its full potential for positive change.
The unchecked advancement of artificial intelligence risks creating a detrimental cycle of displacement and inequality. Without careful consideration, automation could disproportionately impact vulnerable workers, leading to job losses and wage stagnation. This, in turn, could concentrate wealth and power, further exacerbating existing societal divides. However, proactive interventions – such as investment in retraining programs, portable benefits, and policies that encourage profit-sharing – can disrupt this potential negative spiral. By ensuring a broader distribution of the economic gains generated by AI, societies can mitigate the risks of widespread hardship and foster a future where technological progress benefits everyone, rather than just a select few. Addressing these challenges now is crucial to prevent a self-reinforcing cycle that could undermine social cohesion and economic stability.
Economic disruptions stemming from increasing automation necessitate a strengthened social safety net, and automatic fiscal stabilizers represent a proactive approach to mitigating potential harms. These systems, such as unemployment insurance and earned income tax credits, are designed to automatically increase benefits during economic downturns – functioning as a buffer against job displacement and reduced incomes. Unlike discretionary policies requiring legislative action, automatic stabilizers respond immediately to changing economic conditions, providing crucial and timely support to individuals and families. Investment in these programs isn’t merely about damage control; it also facilitates a smoother transition for workers adapting to new roles and industries, potentially fostering innovation by reducing anxieties surrounding technological unemployment and allowing individuals to reskill without facing immediate financial hardship. By proactively bolstering these systems, economies can better absorb the shocks of automation and ensure that the benefits of artificial intelligence are more broadly shared.
The study posits a potential crisis stemming from rapid AI adoption, a scenario echoing the inherent fragility of complex systems. It observes that concentrating consumption among a diminishing number of actors invites systemic stress – a predictable outcome, given the tendency for optimization to create new vulnerabilities. As Tim Berners-Lee once stated, “The Web is more a social creation than a technical one.” This holds true for any system attempting to mediate value; its success relies not on flawless engineering, but on the resilience built through distributed participation. A system that fails to account for the human element, for the messy reality of task reinstatement and fiscal redistribution, is not merely imperfect – it is fundamentally unsustainable. The inevitability of failure, then, isn’t a bug, but a crucial signal for adaptation.
What Lies Ahead?
The exercise reveals, perhaps predictably, that forecasting the impact of abundant intelligence is less about predicting a specific crisis and more about acknowledging the inherent fragility of complex systems. The model, a simplification of an unknowable reality, identifies potential pressure points – a concentration of consumption, a collapse of intermediation – but these are merely symptoms. The true problem isn’t what fails, but that everything optimized will someday lose flexibility. Scalability is just the word used to justify complexity, and complexity, invariably, breeds unforeseen vulnerabilities.
Future work shouldn’t focus on refining the prediction, but on understanding the dynamics of task reinstatement. The model treats labor as a homogeneous block, susceptible to displacement. But the human capacity for adaptation, for finding new niches, remains largely unmodeled. Is it possible to design systems that anticipate, even encourage, this continuous re-skilling? Or is such a notion merely a comforting fiction?
Ultimately, this investigation is a reminder that the perfect architecture is a myth to keep people sane. The real challenge lies not in building resilient systems, but in cultivating ecosystems capable of absorbing inevitable shocks. It’s about accepting that failure isn’t an exception, but the fundamental state of things, and preparing not to prevent it, but to learn from it.
Original article: https://arxiv.org/pdf/2603.09209.pdf
Contact the author: https://www.linkedin.com/in/avetisyan/
See also:
- Building 3D Worlds from Words: Is Reinforcement Learning the Key?
- Securing the Agent Ecosystem: Detecting Malicious Workflow Patterns
- The Best Directors of 2025
- Gold Rate Forecast
- 2025 Crypto Wallets: Secure, Smart, and Surprisingly Simple!
- Mel Gibson, 69, and Rosalind Ross, 35, Call It Quits After Nearly a Decade: “It’s Sad To End This Chapter in our Lives”
- 20 Best TV Shows Featuring All-White Casts You Should See
- TV Shows Where Asian Representation Felt Like Stereotype Checklists
- Umamusume: Gold Ship build guide
- Top 20 Educational Video Games
2026-03-11 06:57