AI improvements boost returns to AI capital, shift investment toward AI and generate a front-loaded output boom that slightly raises fertility; by contrast, longer lifespans increase saving, lower interest rates and fertility, and account for most macro volatility.
This paper develops a general equilibrium overlapping-generations model with endogenous fertility, in which firms accumulate both physical and artificial intelligence (AI) capital, and uses it to study the macroeconomic transmission of two structural disturbances: an AI technology shock and a longevity shock. The AI shock acts as a capital-demand disturbance: it raises all rates of return, most sharply the return to AI capital, reallocates investment from physical to AI capital, and produces a front-loaded output expansion that decays monotonically. The longevity shock acts as a saving-supply disturbance: it deepens the aggregate capital stock, compresses returns and the real interest rate, and generates hump-shaped, persistent dynamics. The two shocks move fertility in opposite directions: AI raises it modestly through an income effect, while longevity lowers it by strengthening the life-cycle saving motive and the cost of childrearing. A forecast-error variance decomposition attributes most aggregate volatility to the longevity shock, while the AI shock dominates the variance of the return to AI capital. Fertility is strongly countercyclical and almost perfectly negatively correlated with hours worked, placing household time allocation at the center of the mechanism. Robustness checks across the capital share, the shock persistence, and the utility specification show that only an empirically implausible labor-AI elasticity reverses the wage and fertility signs. A welfare analysis finds the AI shock welfare-improving under complementarity, whereas longevity produces a short-run welfare loss that recedes as capital deepening raises wages, since households initially compress consumption and fertility to finance a longer retirement.
Summary
Main Finding
Automation (an AI-technology shock) and population aging (a longevity shock) operate as opposite-signed disturbances to the same equilibrium price — the return to capital — with distinct macro and demographic consequences. An AI shock behaves like a capital-demand shock that raises all returns (especially the return to AI capital), reallocates investment toward AI, and produces a front-loaded output expansion; it mildly raises fertility via an income effect and can be welfare-improving under labor–AI complementarity. A longevity shock acts like a saving-supply shock that deepens the capital stock, compresses returns and the real interest rate, generates hump-shaped persistent dynamics, and lowers fertility as households reallocate time and saving toward longer retirement. These return responses are structurally invariant to preferences; only the labor–AI elasticity of substitution can flip some wage/fertility signs, and only beyond empirically implausible values.
Key Points
-
Model scope
- Overlapping-generations (OLG) general equilibrium life-cycle model with endogenous fertility and stochastic survival (longevity).
- Households choose consumption, saving, and number of children; child-rearing incurs a parental time cost that reduces market labor supply.
- Firms produce with physical capital, labor, and AI capital. AI capital is modeled as a distinct capital stock and enters a nested CES technology: AI substitutes for labor in automatable tasks but can complement labor in aggregate depending on the labor–AI elasticity.
- Annuitization parameter (µ) allows partial annuity markets and accidental bequests.
-
Central mechanism
- Two distinct structural shocks: an AI productivity (technology) shock and a longevity (survival probability) shock.
- AI shock: capital-demand disturbance → raises marginal products and returns (RTS especially high for AI capital), shifts investment toward AI, output jumps on impact then decays monotonically.
- Longevity shock: saving-supply disturbance → strengthens life-cycle saving motive, increases aggregate capital (capital deepening), compresses returns and interest rate, produces hump-shaped persistent dynamics.
- Because they affect the capital market from opposite sides (demand vs. supply), the two shocks leave opposite imprints on returns and the real interest rate.
-
Fertility and time allocation
- Fertility responds in opposite directions: AI mildly raises fertility (income effect); greater longevity lowers fertility (higher life-cycle saving motive and higher opportunity cost of parental time).
- Fertility is strongly countercyclical and nearly perfectly negatively correlated with hours worked, highlighting household time allocation (market work vs. child-rearing) as central to transmission.
-
Quantitative/second-moment results
- Forecast-error variance decomposition: longevity shock accounts for most aggregate volatility (output, consumption, etc.); AI shock dominates variance of the return to AI capital.
- Investment reallocation: physical → AI capital when AI productivity rises.
- Welfare: AI shock can be welfare-improving under complementarity between labor and AI; longevity generates a short-run welfare loss that fades as capital deepening raises wages.
-
Analytical robustness
- The opposite-sign responses of returns to the two shocks are robust and provable (preference-invariant).
- Only the labor–AI elasticity of substitution can reverse wage and fertility responses — but reversal requires an elasticity value well outside standard empirical estimates.
- Results robust to variations in capital share, shock persistence, and utility specification.
Data & Methods
-
Model and theoretical methods
- A deterministic/stochastic overlapping-generations general equilibrium model with two-period agents (young/old), endogenous fertility, stochastic survival, partial annuitization, and monopolistic competition in varieties.
- Production: final-good CES aggregator across varieties; intermediate firms use nested CES/Cobb–Douglas combining physical capital, labor, and AI capital (AI as a separable factor).
- Closed-economy setup with population-dependent mass of firms (extensive margin linked to demographic structure).
- Household first-order conditions analyzed with Implicit Function Theorem; comparative statics derived analytically (including proofs that return-sign asymmetry is preference-invariant and derivation of threshold elasticity separating complementarity/substitution regimes — Corollary 2.1).
-
Quantitative methods
- Calibration to stylized features of advanced economies (empirical motivation drawn from U.S. demographic and AI-adoption facts); numerical experiments compute impulse-response functions (IRFs) to AI and longevity shocks.
- Second-moment analysis: forecast-error variance decomposition to apportion aggregate volatility across the two shocks.
- Welfare analysis: exact consumption-equivalent decomposition of welfare effects for each shock; links shocks to dynamic efficiency (Diamond, 1965) considerations.
- Robustness checks: vary capital share, shock persistence, utility specification, and the labor–AI elasticity of substitution.
-
Data note
- The paper is primarily theoretical/quantitative with calibration and numerical simulation rather than empirical estimation. Empirical motivation cites demographic and AI-adoption statistics (e.g., rising old-age shares, falling fertility, rapid AI investment/adoption).
Implications for AI Economics
-
Conceptual implications
- Treat AI as a distinct capital stock in macro models. Rapid, quality-improving AI investment can best be modeled as accumulation of an AI-specific capital factor with its own pricing and substitutability properties vis-à-vis labor and physical capital.
- Viewing AI shocks as capital-demand disturbances explains rising returns to capital, sectoral reallocation toward AI, and front-loaded output effects that differ qualitatively from shocks that expand saving.
-
Empirical priorities
- Estimate the labor–AI elasticity of substitution precisely. This parameter determines whether AI complements or substitutes labor in aggregate and therefore whether wages (and fertility responses) rise or fall after AI progress.
- Improve measurement of AI capital accumulation (spending on compute, models, data, software) and separate it from traditional capital in national accounts, since AI capital has distinct macro effects.
- Quantify the contribution of demographic shocks (longevity) versus technology shocks to aggregate volatility using structural VARs or DSGE estimation that include AI-capital measures.
-
Policy implications
- Distributional and stabilization policy should account for opposite effects of aging and automation on returns and interest rates: aging tends to compress returns (capital deepening) while AI raises returns (capital demand).
- Fertility and household time allocation matter: childcare time costs and labor-market policies interact with automation and longevity. Policies supporting parental time (childcare services, family-friendly labor policies) can alter saving and fertility responses and hence long-run capital accumulation.
- Social insurance and annuity markets matter: degree of annuitization influences retirement consumption and saving behavior; reforms in pensions/annuity markets will interact with both longevity and automation dynamics.
-
Directions for further research
- Extend to open-economy settings to examine cross-border capital flows and how aging/automation interact with international spillovers.
- Introduce heterogeneity (skill, sectoral exposure to AI) to study inequality and distributional effects of AI vs. aging.
- Incorporate endogenous technology adoption driven by demographic signals (older populations inducing more automation) and model policy levers (taxation/subsidies, training programs, childcare subsidies) to evaluate welfare-optimal responses.
- Empirical validation: use micro- and macro-evidence (firm-level AI adoption, local labor markets, household time-use surveys) to calibrate and test the nested-CES specification and the model’s predicted co-movements (returns, wages, fertility, hours).
Summary takeaway: Integrating endogenous fertility, longevity, and an explicit AI capital factor in a life-cycle general equilibrium framework clarifies that AI and aging are mirror-image shocks to capital pricing with distinct macro, demographic, and welfare consequences; measurement of AI capital and the labor–AI elasticity are key empirical priorities, and policy design must account for these opposing forces.
Assessment
Claims (11)
| Claim | Direction | Outcome | Confidence & Evidence | Details |
|---|---|---|---|---|
| An AI technology shock acts as a capital-demand disturbance: it raises all rates of return, most sharply the return to AI capital. Firm Productivity | positive | rates of return (aggregate and to AI capital) |
Reading fidelity
high
Study strength
medium
|
not reported
|
| The AI shock reallocates investment from physical to AI capital. Task Allocation | mixed | investment allocation between physical and AI capital |
Reading fidelity
high
Study strength
medium
|
not reported
|
| The AI shock produces a front-loaded output expansion that decays monotonically. Firm Productivity | positive | aggregate output (time path) |
Reading fidelity
high
Study strength
medium
|
not reported
|
| A longevity shock acts as a saving-supply disturbance: it deepens the aggregate capital stock. Fiscal And Macroeconomic | positive | aggregate capital stock |
Reading fidelity
high
Study strength
medium
|
not reported
|
| The longevity shock compresses asset returns and lowers the real interest rate, and generates hump-shaped, persistent dynamics. Fiscal And Macroeconomic | negative | asset returns and real interest rate (time path/persistence) |
Reading fidelity
high
Study strength
medium
|
not reported
|
| The two shocks move fertility in opposite directions: the AI shock raises fertility modestly through an income effect, while the longevity shock lowers fertility by strengthening life-cycle saving motives and increasing the cost of childrearing. Other | mixed | fertility (birth rate/children per household) |
Reading fidelity
high
Study strength
medium
|
modest (qualitative)
|
| A forecast-error variance decomposition attributes most aggregate volatility to the longevity shock, while the AI shock dominates the variance of the return to AI capital. Fiscal And Macroeconomic | mixed | variance decomposition of aggregate volatility and variance of return to AI capital |
Reading fidelity
high
Study strength
medium
|
not reported
|
| Fertility is strongly countercyclical and almost perfectly negatively correlated with hours worked in the model, placing household time allocation at the center of the mechanism. Other | negative | correlation between fertility and hours worked (countercyclicality) |
Reading fidelity
high
Study strength
medium
|
almost perfectly negatively correlated (qualitative)
|
| Robustness checks across the capital share, shock persistence, and the utility specification show that only an empirically implausible labor–AI elasticity reverses the wage and fertility signs. Wages | mixed | signs of wage and fertility responses to shocks under parameter variations |
Reading fidelity
high
Study strength
medium
|
reversal only under empirically implausible labor–AI elasticity (qualitative)
|
| Welfare analysis finds the AI shock welfare-improving under complementarity between labor and AI capital. Other | positive | household welfare (utility) |
Reading fidelity
high
Study strength
medium
|
not reported
|
| Longevity produces a short-run welfare loss that recedes as capital deepening raises wages, since households initially compress consumption and fertility to finance a longer retirement. Other | mixed | household welfare over time (short-run loss, subsequent recovery) |
Reading fidelity
high
Study strength
medium
|
short-run welfare loss (qualitative)
|