Automation and AI are tilting production toward technological capital, shrinking labour’s share and eroding payroll‑based revenues; policymakers must rethink tax bases and pension designs or face mounting fiscal and social strain.

Main Finding

The paper argues that the rise of automation, robots, and AI is shifting value creation from human labor to technological capital (K_T), undermining wage‑based PAYG pension financing. Demographic ageing (in the EU) and demographic decline plus emigration and informality (in Albania) together create a structural challenge: fewer contributors and a declining labor share of GDP. To preserve equity, intergenerational justice, and fiscal sustainability, pension systems must broaden their funding base to capture a portion of capital/automation‑generated value (e.g., levies on automation profits, capital income contributions, state stakes in technology), alongside institutional reforms (formalization, migrant integration) — rather than relying on parametric fixes alone.

Key Points

• Conceptual shift: production is moving from Y = f(L, K, T, E) toward a technological‑capital dominated K_T regime; this reduces labor’s share and the wage base for PAYG pensions.
• Empirical indicators highlighted:
  • EU labor share of GDP fell from 56.05% (2010) to 55.10% (2024), with a sharper drop to 53.62% in 2022; capital share rose from 43.95% to 44.90% (peaking 46.38% in 2022).
  • Albania: population ~10% lower (2000–2023); ~39% of working‑age Albanians live abroad; informal employment ≈34%; youth employment ~30%.
• Methodological stance: a conceptual‑normative analysis rather than a numerical forecasting exercise. Uses principle testing (fairness, efficiency, solidarity), institutional assessment, and normative evaluation of alternative financing.
• Normative framework: three guiding principles for reform — Equity (capture value regardless of source), Solidarity (extend protection to informal workers and migrants), and Sustainability (financial viability across generations).
• Policy proposals:
  • For EU: levies/“automation dividends” on AI/robotics profits, earmarking portions of corporate/capital taxes for pensions, modest parametric adjustments, stakeholder engagement for legitimacy.
  • For Albania: accelerate formalization and enforcement, integrate emigrants and informal workers into contribution systems, broaden funding sources (including taxation of capital/automation gains), and strengthen institutional capacity.
• Limitation: the study is normative/conceptual; it does not present calibrated macro or microsimulation models or precise sustainability thresholds.

Data & Methods

• Data cited (secondary sources and aggregated indicators):
  • Labor vs. capital share of GDP (2010–2024), old‑age dependency ratio projections to 2050 (EU), contributor/retiree series, Albanian demographic and labor‑market statistics (INSTAT, World Bank).
  • Literature grounding: Acemoglu & Restrepo; Korinek & Stiglitz; Piketty; Atkinson; Esping‑Andersen; Barr; OECD and European Commission reports.
• Methods:
  • Comparative, conceptual–normative approach: contrasts EU (ageing + automation) with Albania (emigration + informality).
  • Analytical tools: principle testing (fairness, efficiency, solidarity), institutional assessment (formalization, enforcement, fiscal capacity), and normative evaluation of financing alternatives (capital taxation, automation levies, public equity).
  • Visual/empirical support: time series illustrations (labor/capital shares, old‑age dependency, contributors per retiree) used narratively; no new econometric estimation or calibrated projections provided.

Implications for AI Economics

• Model specification: macro and labor models should explicitly incorporate technological capital (K_T) and endogenous substitution effects between AI/automation and labor when assessing distributional and fiscal outcomes.
• Public finance and tax design:
  • AI/automation generates rents that conventional payroll taxes miss; economists should evaluate efficiency and incidence of proposals like automation levies, targeted capital taxes, or dividends linked to AI use.
  • Measurement challenge: assignability of “AI‑generated value” to firms/sectors and distinguishing returns to embodied vs. disembodied technological capital—critical for tax base design.
• Social insurance financing: declining labor shares imply PAYG instability; research should quantify how much of AI/capital returns would need to be captured to stabilize pensions and assess macroeconomic side effects (innovation incentives, investment).
• Distributional dynamics: automation‑driven capital gains may exacerbate inequality; AI economics must link technology adoption decisions to general equilibrium effects on wages, employment, profits, and public revenues.
• Institutional considerations: ownership structures (concentrated tech incumbents vs. broad ownership) shape how automation rents are distributed; policy efficacy depends on corporate structure, international tax coordination, and enforcement capacity—especially in emerging economies.
• Policy evaluation agenda: require integrated frameworks combining growth, distribution, and public‑finance modules to simulate tradeoffs (e.g., dynamic stochastic general equilibrium models with heterogeneous agents and explicit K_T).
• Research gaps signaled by the paper: need for empirical quantification of substitution elasticities between labor and K_T, calibrated estimates of pension funding shortfalls under different automation scenarios, and country‑specific feasibility studies for automation taxation and state equity options.

If you want, I can: (a) extract specific policy options into a short policy brief, (b) draft a simple stylized macro model that embeds K_T and PAYG financing to explore sensitivities, or (c) produce an annotated bibliography of the cited works relevant to AI economics and pension finance. Which would be most useful?