Hubbry Logo
search
logo
Overlapping generations model avatar
Overlapping generations model
Overlapping generations model
current hub
O

Overlapping generations model

logo
Community Hub0 Subscribers

Wikipedia

Overlapping generations model
View on Wikipedia
from Wikipedia
For the topic in population genetics, see Overlapping generations.
Part of a series on
Macroeconomics
Federal Reserve
Basic concepts
Policies
Models
Related fields
Schools
Mainstream

Heterodox

People
See also

The overlapping generations (OLG) model is one of the dominating frameworks of analysis in the study of macroeconomic dynamics and economic growth. In contrast to the Ramsey–Cass–Koopmans neoclassical growth model in which individuals are infinitely-lived, in the OLG model individuals live a finite length of time, long enough to overlap with at least one period of another agent's life.

The OLG model is the natural framework for the study of: (a) the life-cycle behavior (investment in human capital, work and saving for retirement), (b) the implications of the allocation of resources across the generations, such as Social Security, on the income per capita in the long-run,[1] (c) the determinants of economic growth in the course of human history, and (d) the factors that triggered the fertility transition.

History

[edit]

The construction of the OLG model was inspired by Irving Fisher's monograph The Theory of Interest.[2] It was first formulated in 1947, in the context of a pure-exchange economy, by Maurice Allais, and more rigorously by Paul Samuelson in 1958.[3] In 1965, Peter Diamond[4] incorporated an aggregate neoclassical production into the model. This OLG model with production was further augmented with the development of the two-sector OLG model by Oded Galor,[5] and the introduction of OLG models with endogenous fertility.[6][7]

Books devoted to the use of the OLG model include Azariadis' Intertemporal Macroeconomics[8] and de la Croix and Michel's Theory of Economic Growth.[9]

Pure-exchange OLG model

[edit]
Generational Shifts in OLG Models

The most basic OLG model has the following characteristics:[10]

  • Individuals live for two periods; in the first period of life, they are referred to as the Young. In the second period of life, they are referred to as the Old.
  • A number of individuals are born in every period. denotes the number of individuals born in period t.
  • denotes the number of old people in period t. Since the economy begins in period 1, in period 1 there is a group of people who are already old. They are referred to as the initial old. The number of them can be denoted as .
  • The size of the initial old generation is normalized to 1: .
  • People do not die early, so .
  • Population grows at a constant rate n:
  • In the "pure exchange economy" version of the model, there is only one physical good and it cannot endure for more than one period. Each individual receives a fixed endowment of this good at birth. This endowment is denoted as y.
  • In the "production economy" version of the model (see Diamond OLG model below), the physical good can be either consumed or invested to build physical capital. Output is produced from labor and physical capital. Each household is endowed with one unit of time which is inelastically supply on the labor market.
  • Preferences over consumption streams are given by
where is the rate of time preference.

OLG model with production

[edit]

Basic one-sector OLG model

[edit]

The pure-exchange OLG model was augmented with the introduction of an aggregate neoclassical production by Peter Diamond.[4]  In contrast, to Ramsey–Cass–Koopmans neoclassical growth model in which individuals are infinitely-lived and the economy is characterized by a unique steady-state equilibrium, as was established by Oded Galor and Harl Ryder,[11] the OLG economy may be characterized by multiple steady-state equilibria, and initial conditions may therefore affect the long-run evolution of the long-run level of income per capita.

Since initial conditions in the OLG model may affect economic growth in long-run, the model was useful for the exploration of the convergence hypothesis.[12]

Convergence of OLG Economy to Steady State

The economy has the following characteristics:[13]

  • Two generations are alive at any point in time, the young (age 1) and old (age 2).
  • The size of the young generation in period t is given by Nt = N0 Et.
  • Households work only in the first period of their life and earn Y1,t income. They earn no income in the second period of their life (Y2,t+1 = 0).
  • They consume part of their first period income and save the rest to finance their consumption when old.
  • At the end of period t, the assets of the young are the source of the capital used for aggregate production in period t+1.So Kt+1 = Nt,a1,t where a1,t is the assets per young household after their consumption in period 1. In addition to this there is no depreciation.
  • The old in period t own the entire capital stock and consume it entirely, so dissaving by the old in period t is given by Nt-1,a1,t-1 = Kt.
  • Labor and capital markets are perfectly competitive and the aggregate production technology is CRS, Y = F(K,L).

Two-sector OLG model

[edit]

The one-sector OLG model was further augmented with the introduction of a two-sector OLG model by Oded Galor.[5] The two-sector model provides a framework of analysis for the study of the sectoral adjustments to aggregate shocks and implications of international trade for the dynamics of comparative advantage. In contrast to the Uzawa two-sector neoclassical growth model,[14] the two-sector OLG model may be characterized by multiple steady-state equilibria, and initial conditions may therefore affect the long-run position of an economy.

OLG model with endogenous fertility

[edit]

Oded Galor and his co-authors develop OLG models where population growth is endogenously determined to explore: (a) the importance the narrowing of the gender wage gap for the fertility decline,[6] (b) the contribution of the rise in the return to human capital and the decline in fertility to the transition from stagnation to growth,[7][15] and (c) the importance of population adjustment to technological progress for the emergence of the Malthusian trap.[16]

Dynamic inefficiency

[edit]

One important aspect of the OLG model is that the steady state equilibrium need not be efficient, in contrast to general equilibrium models where the first welfare theorem guarantees Pareto efficiency. Because there are an infinite number of agents in the economy (summing over future time), the total value of resources is infinite, so Pareto improvements can be made by transferring resources from each young generation to the current old generation,[17] similar to the logic described in the Hilbert Hotel. Not every equilibrium is inefficient; the efficiency of an equilibrium is strongly linked to the interest rate and the Cass Criterion gives necessary and sufficient conditions for when an OLG competitive equilibrium allocation is inefficient.[18]

Another attribute of OLG type models is that it is possible that 'over saving' can occur when capital accumulation is added to the model—a situation which could be improved upon by a social planner by forcing households to draw down their capital stocks.[4] However, certain restrictions on the underlying technology of production and consumer tastes can ensure that the steady state level of saving corresponds to the Golden Rule savings rate of the Solow growth model and thus guarantee intertemporal efficiency. Along the same lines, most empirical research on the subject has noted that oversaving does not seem to be a major problem in the real world.[citation needed]

In Diamond's version of the model, individuals tend to save more than is socially optimal, leading to dynamic inefficiency. Subsequent work has investigated whether dynamic inefficiency is a characteristic in some economies[19] and whether government programs to transfer wealth from young to poor do reduce dynamic inefficiency[citation needed].

Another fundamental contribution of OLG models is that they justify existence of money as a medium of exchange. A system of expectations exists as an equilibrium in which each new young generation accepts money from the previous old generation in exchange for consumption. They do this because they expect to be able to use that money to purchase consumption when they are the old generation.[10]

See also

[edit]

References

[edit]

Further reading

[edit]

Grokipedia

Overlapping generations model

View on Grokipedia
from Grokipedia
The overlapping generations (OLG) model is a discrete-time framework in macroeconomics that examines intertemporal decision-making and resource allocation in economies where agents live for a finite number of periods, with successive cohorts overlapping in each period, contrasting with representative-agent models that assume infinite-lived agents.[1] Introduced by Paul Samuelson in 1958, the pure-exchange version of the model provided a foundational explanation for the positive value of fiat money in equilibrium, as young agents accept intrinsically worthless currency in exchange for goods, anticipating its future usability by the next generation.[2] Peter Diamond extended the model in 1965 by integrating production functions and capital accumulation, allowing analysis of long-run growth equilibria, savings behavior, and the neutrality of public debt under neoclassical assumptions.[3] Key features of the OLG model include sequential markets without bequest motives, leading to potential Pareto inefficiencies in competitive equilibria relative to a social planner's optimum, as demonstrated by Samuelson's "golden rule" savings criterion where over-accumulation of capital can reduce welfare.[4] Applications span monetary theory, where it rationalizes the demand for money absent lump-sum transfers; public finance, evaluating pay-as-you-go social security systems that may crowd out private savings; and demographic shifts, modeling aging populations' impacts on capital intensity and fiscal sustainability.[5] Despite its simplicity—often assuming two-period lives and logarithmic utility—the model's tractability has made it influential in policy simulations, though extensions to heterogeneous agents and uncertainty address limitations in capturing realistic heterogeneity.[6]

Historical Development

Origins in Pure Exchange Models

The overlapping generations (OLG) model in its pure exchange form emerged as a framework to analyze intertemporal resource allocation among agents with finite lifetimes in an economy without production. Maurice Allais first formulated the basic structure in 1947 within an appendix to his book Économie et Intérêt, modeling a pure-exchange economy where successive generations overlap and trade claims on goods across periods.[7] This approach addressed intergenerational transfers but received limited attention initially due to its publication in French and lack of widespread dissemination.[8] Paul Samuelson independently and more rigorously introduced the OLG model in 1958 through his seminal paper "An Exact Consumption-Loan Model of Interest with or without the Social Contrivance of Money," published in the Journal of Political Economy.[9] Samuelson's setup posits an infinite-horizon economy populated by agents who live for two periods: the young, endowed with a perishable consumption good, and the old, who seek to consume but hold no endowment.[2] Absent production, young agents must transfer resources to the old via consumption loans or fiat money to enable old-age consumption, with equilibrium interest rates determined by population growth and time preferences.[9] This pure exchange OLG framework highlighted the role of money as a store of value in stationary economies and demonstrated that decentralized equilibria could feature positive interest rates even without productive capital, contrasting with neoclassical growth models.[2] Samuelson showed that such equilibria might not achieve Pareto optimality, introducing the possibility of dynamic inefficiency where the economy over-accumulates stores of value relative to the golden rule level.[9] These insights laid the groundwork for later extensions, emphasizing the model's utility in capturing realistic demographic structures absent in representative-agent infinite-horizon frameworks.[10]

Introduction of Production and Key Extensions

Peter Diamond introduced production into the overlapping generations model in his 1965 paper "National Debt in a Neoclassical Growth Model," published in the American Economic Review.[3] This built upon Paul Samuelson's 1958 pure exchange framework by adding a production sector where output is generated using capital and labor inputs via a neoclassical aggregate production function exhibiting constant returns to scale and Inada conditions.[11] In the model, young agents supply labor inelastically, receive wages, and allocate income between current consumption and savings, which form the capital stock for the subsequent period after accounting for population growth at rate nn.[12] Elderly agents consume from their accumulated savings plus rental returns, with no further labor supply or bequests.[13] Firms operate competitively, hiring labor at wage wt=f(kt)−ktf′(kt)w_t = f(k_t) - k_t f'(k_t) and capital at rental rate rt+1=f′(kt+1)r_{t+1} = f'(k_{t+1}), where f(kt)f(k_t) denotes per capita output from capital per worker ktk_t and the production function F(Kt,Lt)F(K_t, L_t).[12] Household optimization yields savings that determine capital accumulation via kt+1=s(wt,rt+1)/(1+n)k_{t+1} = s(w_t, r_{t+1}) / (1+n), leading to a dynamic system with a unique stable steady state under standard parameter values, such as discount factor β<1\beta < 1 and capital share ε∈(0,1)\varepsilon \in (0,1).[13] Diamond's formulation enabled analysis of long-run equilibria and transitions, highlighting how savings-driven capital accumulation differs from exogenous growth assumptions in earlier Solow-style models.[3] Key extensions incorporated fiscal elements within Diamond's structure, including government debt and pay-as-you-go social security, where unfunded liabilities crowd out private capital and may induce dynamic inefficiency if steady-state capital exceeds the Golden Rule level kGk_G satisfying f′(kG)=n+δf'(k_G) = n + \delta.[13] Diamond demonstrated that internal debt reduces capital accumulation by transferring resources intertemporally, while external debt's effects depend on interest rates relative to growth.[3] Later historical developments, such as Auerbach and Kotlikoff's 1987 multi-period OLG models with heterogeneous agents and life-cycle heterogeneity, expanded the framework for computational general equilibrium analysis of demographic shifts and policy reforms, popularizing its use in public finance simulations.[14]

Evolution in Macroeconomic Theory

The overlapping generations (OLG) model emerged as a foundational tool in macroeconomic theory through Paul Samuelson's 1958 analysis of a pure exchange economy, where finite-lived agents overlap across periods, enabling the demonstration of fiat money's positive value as a store of worth despite its intrinsic worthlessness and highlighting potential inefficiencies in decentralized equilibria compared to optimal social allocations.[9] This framework contrasted with infinite-horizon representative agent models by incorporating realistic demographic structure, where agents save for retirement without intergenerational altruism, leading to equilibria that may fail Pareto optimality due to missing intertemporal markets.[12] Peter Diamond's 1965 extension integrated production into the OLG setup, adapting neoclassical growth dynamics to finite lives and showing that competitive steady-state capital accumulation could exceed the golden rule level, resulting in dynamic inefficiency where the interest rate falls below the growth rate, rendering public debt or resource transfers Pareto improving.[3] This revealed a key theoretical insight: market-driven savings, motivated by life-cycle needs rather than dynastic preferences, could overaccumulate capital relative to socially optimal paths, influencing debates on fiscal policy's role in correcting such distortions.[15] Post-Diamond developments expanded OLG applications to monetary and fiscal analysis, with extensions incorporating endogenous money supply and banking to explain inflation dynamics and seigniorage, while emphasizing the model's capacity to generate valued intrinsically useless assets through sequential trading.[4] In the 1980s, Auerbach and Kotlikoff pioneered large-scale, multi-period OLG simulations for quantitative policy evaluation, modeling heterogeneous cohorts with age-specific earnings and consumption to assess tax reforms, social security sustainability, and demographic transitions like aging populations, which revealed transition generations bearing disproportionate burdens in pay-as-you-go systems.[16][14] By the late 20th century, OLG frameworks evolved into computational workhorses for addressing heterogeneity absent in representative agent models, facilitating analysis of inequality, lifecycle risk-sharing, and aggregate fluctuations driven by cohort effects.[17] In modern macroeconomics, OLG structures inform heterogeneous agent New Keynesian models, incorporating nominal rigidities and demographic variation to better capture policy transmission, such as how monetary shocks disproportionately affect younger borrowers versus older savers, thus providing microfoundations for empirical regularities in consumption and savings behavior.[6] This evolution underscores OLG's enduring role in privileging agent-specific horizons over infinite-lived approximations, yielding predictions aligned with observed intergenerational trade-offs.

Core Model Frameworks

Pure Exchange OLG Model

The pure exchange overlapping generations (OLG) model posits an economy spanning infinite discrete time periods $ t = 0, 1, 2, \dots $, where each generation lives for two periods—youth and old age—overlapping with the subsequent generation. [18] Agents born in period $ t $, denoted generation $ t $, are endowed with $ w > 0 $ units of a perishable consumption good solely in their youth (period $ t $) and nothing in old age (period $ t+1 $); the good cannot be stored or produced.[19] [18] Population size of newborns grows exogenously at constant rate $ n \geq 0 $, so the number of young in period $ t $ is $ N_t^t = (1+n)^t $.[19] [20] Preferences are time-separable and concave, with utility $ u(c_t^t, c_{t+1}^t) = U(c_t^t) + \beta U(c_{t+1}^t) $, where $ c_t^t $ is young-age consumption, $ c_{t+1}^t $ old-age consumption, $ U(\cdot) $ strictly increasing and concave, and $ 0 < \beta < 1 $ the discount factor.[19] In autarky—absent any mechanism for intertemporal transfer—each generation consumes its full endowment when young ($ c_t^t = w )andstarveswhenold() and starves when old ( c_{t+1}^t = 0 $), yielding zero old-age utility and precluding consumption smoothing despite agents' willingness to trade current for future goods. [18] This outcome is Pareto dominated, as reallocations could increase welfare without harming prior generations, highlighting the model's demonstration of market failure in achieving optimality without coordination.[20] To facilitate exchange, the model introduces fiat money as a non-productive store of value with fixed aggregate supply $ M \geq 0 $, intrinsically worthless but accepted due to anticipated future exchange. [21] Let $ p_t $ denote the price level (goods per unit money) in period $ t $; young agents maximize utility subject to budget $ c_t^t + m_t / p_t = w $ (where $ m_t $ is real money balances acquired) and anticipated old-age $ c_{t+1}^t = m_t / p_{t+1} $.[19] [21] The first-order condition yields $ U'(c_t^t) = \beta U'(c_{t+1}^t) (p_t / p_{t+1}) $, equating marginal utilities adjusted for the money return $ p_t / p_{t+1} = 1 + r_{m,t} $.[19] Money market clearing requires aggregate demand by young equals supply: $ N_t^t m_t = M $, so per-young balances $ m_t = M / N_t^t $.[21] In stationary monetary equilibrium (constant allocations, prices adjusting to growth), young consumption solves $ c^y + (M / p) / N_0 = w $ and $ c^o = (M / p) / N_0 \cdot (1+n) $ (old outnumber young by factor $ 1+n $), with $ U'(c^y) = \beta U'(c^o) (1+n) $.[19] [21] The implied money return is $ r_m = n $, or gross rate $ 1 + r_m = 1 + n $; for $ n > 0 $, this falls below the autarkic marginal product of capital (effectively infinite, as goods perish), rendering the equilibrium dynamically inefficient per the Golden Rule, where capital (here, claims on future goods) yields returns below population growth.[20] [19] If $ n = 0 ,moneycansustainzeronominal[interest](/page/Interest)(, money can sustain zero nominal [interest](/page/Interest) ( i = 0 $), mimicking optimal loan rates; otherwise, positive growth dilutes money's purchasing power, bounding transfers. Multiple equilibria exist: the autarkic (zero money value, $ p_t = 0 $) or positive-value monetary steady state, with sunspot-driven fluctuations possible under certain utilities.[21]

One-Sector Production OLG Model

The one-sector production overlapping generations (OLG) model integrates capital accumulation and neoclassical production into the basic OLG framework, allowing analysis of intergenerational resource allocation through savings and investment. Developed by Peter Diamond in 1965, it features competitive firms producing output using capital and labor inputs under constant returns to scale, with output serving dual roles as consumption goods and capital for future production.[12][13] Individuals live for two periods: the young supply inelastic labor, receive wages, consume part of earnings, and save the rest by acquiring capital claims; the old consume returns on prior savings without working.[22] Population grows exogenously at constant rate n>0n > 0, ensuring overlapping cohorts where each period's young generation equals the prior young's size times (1+n)(1+n).[12] Households derive utility from consumption in both periods via time-separable preferences, typically U(ctt)+βU(ct+1t)U(c_t^t) + \beta U(c_{t+1}^{t}) where β∈(0,1)\beta \in (0,1) discounts future utility and UU exhibits positive marginal utility with diminishing returns, such as CRRA form.[12] The young in period tt, facing wage wtw_t, solve max⁡ctt,stU(ctt)+βU((1+rt+1)st)\max_{c_t^t, s_t} U(c_t^t) + \beta U((1 + r_{t+1}) s_t) subject to ctt+st=wtc_t^t + s_t = w_t, yielding savings st=wt−ctts_t = w_t - c_t^t where marginal conditions equate U′(ctt)=β(1+rt+1)U′(ct+1t)U'(c_t^t) = \beta (1 + r_{t+1}) U'(c_{t+1}^t).[13] Aggregate savings by the young generation, scaled by cohort size Ntt=(1+n)tN_t^t = (1+n)^t, finance next-period capital Kt+1=stNttK_{t+1} = s_t N_t^t. Firms operate in competitive markets with production function Yt=F(Kt,Lt)Y_t = F(K_t, L_t) exhibiting Inada conditions: positive but diminishing marginal products, constant returns, and output malleability for consumption or investment.[22] Labor supply LtL_t equals young population NttN_t^t, as old retire; capital KtK_t derives from prior savings. Profit maximization yields factor prices: rental rate rt=FK(Kt,Lt)r_t = F_K(K_t, L_t) and wage wt=FL(Kt,Lt)w_t = F_L(K_t, L_t), with per-effective-worker terms kt=Kt/Ltk_t = K_t / L_t, f(kt)=F(kt,1)f(k_t) = F(k_t, 1), so rt=f′(kt)r_t = f'(k_t) and wt=f(kt)−ktf′(kt)w_t = f(k_t) - k_t f'(k_t).[13] Goods market clears via CttNtt+Ctt−1Ntt−1+Kt+1+(1−δ)Kt=F(Kt,Lt)C_t^t N_t^t + C_t^{t-1} N_t^{t-1} + K_{t+1} + (1 - \delta) K_t = F(K_t, L_t), where δ\delta is depreciation (often 1 for full in simple versions).[12] Competitive equilibrium sequences satisfy household optimization, firm pricing, capital accumulation Kt+1=st(wt,rt+1)LtK_{t+1} = s_t(w_t, r_{t+1}) L_t, and market clearing for all tt, with transversality ensuring finite asset holdings asymptotically. Steady-state analysis sets growth rates to zero post-per-capita: kt+1=kt=k∗k_{t+1} = k_t = k^* implies s(w(k∗),r(k∗))=k∗(1+n+g+δ)s(w(k^*), r(k^*)) = k^* (1 + n + g + \delta) if including tech growth gg, but base Diamond omits gg.[13] Existence and uniqueness hinge on parameters; for Cobb-Douglas f(k)=kαf(k) = k^\alpha, α<1\alpha < 1, steady k∗k^* solves implicit savings equaling investment needs, with capital deeper than Ramsey due to operative bequest motive absence, potentially yielding dynamic inefficiency if r∗<nr^* < n.[22][13]

Multi-Sector and Endogenous Fertility Extensions

Multi-sector extensions of the overlapping generations (OLG) model introduce heterogeneity in production technologies across goods, departing from the single aggregate output assumption of the Diamond (1965) framework to capture structural features like sector-specific capital and relative price dynamics. In a canonical two-sector OLG model, one sector produces a consumption good using labor and capital, while the other produces an investment good, enabling analysis of intertemporal substitution and potential business cycles driven by sector reallocations. [23] [24] These models characterize global dynamics, including saddle-path stability under gross substitutability conditions, and demonstrate how sector-specific shocks can generate oscillatory equilibria absent in one-sector setups. [25] Further refinements incorporate heterogeneous capital across sectors, allowing for non-homothetic preferences or vintage-specific technologies, which reveal inefficiencies in capital allocation and transitions to steady states influenced by demographic rates. [25] Empirical applications, such as multi-sector OLG simulations for open economies, quantify welfare effects of trade liberalization by tracing intergenerational resource shifts across industries like manufacturing and services. [26] Endogenous fertility extensions embed household decisions on family size into the OLG structure, treating population growth as an outcome of utility maximization rather than an exogenous parameter, thus linking demographics to savings and capital accumulation. Agents balance child-rearing costs—often modeled as time or resource inputs—against benefits like old-age support or altruism, yielding fertility rates ntn_t that respond to wages, interest rates, and public policies. [27] [28] In such models, higher wage rates can reduce fertility via substitution toward child quality (e.g., education) or income effects, reconciling opposing theoretical predictions in static frameworks. [29] These extensions highlight dynamic efficiency challenges: endogenous fertility may lead to over-accumulation if parental altruism distorts transfers, violating Pareto optimality unless property rights enforce minimum bequests. [30] Applications to policy, such as pension reforms, show that pay-as-you-go systems can lower fertility by weakening intergenerational links, slowing long-run growth unless offset by human capital investments. [31] Steady-state analysis often reveals a unique balanced growth path where fertility aligns with discount factors, supporting positive population growth under neoclassical production. [27]

Fundamental Properties

Equilibrium Dynamics and Capital Accumulation

In the Diamond (1965) overlapping generations model with production, competitive equilibrium dynamics are governed by the intertemporal choices of households and profit-maximizing behavior of firms, leading to a recursive law of motion for the capital-labor ratio. Households, living for two periods, allocate earnings from labor supplied inelastically when young between current consumption and savings, which fund consumption when old; savings earn returns from capital rented to firms. Firms produce output using capital and labor via a neoclassical production function $ y_t = f(k_t) $ with constant returns, Inada conditions, and full depreciation for simplicity in basic formulations, yielding factor prices as marginal products: wage $ w_t = f(k_t) - k_t f'(k_t) $ and rental rate $ r_{t+1} = f'(k_{t+1}) $.[12][13] Aggregate savings per young worker $ s_t $ thus depend on current wages and anticipated future returns, with the savings function $ s(k_t) $ derived from the first-order condition of household utility maximization, typically $ s_t = w_t \cdot \frac{1}{1 + \left( \frac{1 + r_{t+1}}{\beta^{-1}} \right)^{1/\sigma - 1}} $ under CRRA preferences with elasticity $ \sigma $, where $ \beta $ is the discount factor. The next-period capital-labor ratio evolves as $ k_{t+1} = \frac{s(k_t)}{1 + n} $, with $ n $ the exogenous population growth rate, ensuring market clearing in goods and capital markets.[12][13][15] The steady-state capital stock $ k^* $ satisfies $ s(k^) = (1 + n) k^ $, uniquely determined under standard assumptions where $ s(k) $ is increasing in wages (hence decreasing in $ k $) and concave, intersecting the $ (1 + n) k $ line once with slope less than $ 1 + n $. Transitional dynamics exhibit monotonic convergence to $ k^* $: if initial $ k_0 < k^* $, capital accumulates as savings exceed depreciation and dilution by population growth, raising $ k_t $ over time; conversely, if $ k_0 > k^* $, capital decumulates. This stability mirrors Solow-Swan dynamics but arises endogenously from decentralized saving decisions rather than an exogenous savings rate.[12][15][22] Capital accumulation in equilibrium thus reflects the balance between productive capacity (via marginal returns) and intergenerational altruism (via $ \beta $), with paths potentially featuring dynamic inefficiency if $ k^* > k_g $, where $ k_g $ maximizes steady-state consumption $ c^y + c^o / (1 + n) $ by equating $ f'(k_g) = n + \delta ;here,over−accumulationoccurswhenpatienthouseholds(; here, over-accumulation occurs when patient households ( \beta (1 + n) > 1 $) save excessively relative to the social optimum, lowering returns below growth rates. Empirical calibrations, such as those matching U.S. data on savings rates and returns, often indicate such inefficiency, supporting policies like public debt to redistribute resources intertemporally.[13][22][4]

Dynamic Efficiency and the Golden Rule

In overlapping generations (OLG) models with production, dynamic efficiency refers to the property that the economy's steady-state allocation cannot be Pareto improved by discarding some capital stock, thereby freeing resources for greater aggregate consumption without harming any generation.[4] This contrasts with infinite-horizon representative agent models, where competitive equilibria are always dynamically efficient due to transversality conditions ensuring optimal capital accumulation.[13] In the canonical Diamond OLG framework, each generation consists of agents living for two periods: the young supply labor inelastically and save part of their wage income for old-age consumption, while capital is produced via aggregate output Yt=F(Kt,Lt)Y_t = F(K_t, L_t) assuming constant returns and Inada conditions.[4] The steady-state capital per worker k∗k^* solves the condition where young agents' savings equal (1+n)k∗(1+n)k^*, with population growth rate n>0n > 0, wage w(k)=F(k,1)−kFk(k,1)w(k) = F(k,1) - k F_k(k,1), and rental rate r(k)=Fk(k,1)r(k) = F_k(k,1), often under log utility and Cobb-Douglas production yielding k∗=(αβ(1+n)1−β(1−α(1+n)))1/(1−α)k^* = \left( \frac{\alpha \beta (1+n)}{1 - \beta (1 - \alpha (1+n))} \right)^{1/(1-\alpha)} where α\alpha is capital's share.[13] Dynamic inefficiency arises when the steady-state net marginal product of capital r∗<nr^* < n, implying over-accumulation of capital relative to the level that maximizes intergenerational welfare.[4] In this case, reducing k∗k^* by Δ>0\Delta > 0 each period allows the economy to permanently increase consumption: the immediate gain from lower investment exceeds the future loss from reduced capital income, as the low return on capital fails to justify its opportunity cost in terms of foregone consumption.[13] This inefficiency stems from decentralized saving decisions ignoring the externalities on future generations' capital dilution via population growth; high time preference (low β\beta) or low capital share α\alpha exacerbates it, potentially making fiat money or government debt Pareto improving by crowding out capital.[4] Empirical assessments, such as those calibrating to U.S. data, suggest rare but plausible inefficiency in post-WWII economies if safe asset returns persistently fall below nn, though stochastic extensions complicate tests by requiring checks on expected returns.[32] The golden rule capital stock kGRk_{GR} maximizes steady-state per capita consumption c=F(k,1)−nkc = F(k,1) - n k (ignoring depreciation for simplicity), satisfying the first-order condition Fk(kGR,1)=nF_k(k_{GR},1) = n.[4] Unlike the modified golden rule in Ramsey models, where Fk=n+δ+ρF_k = n + \delta + \rho with subjective discount ρ\rho, the OLG golden rule equates the marginal product to growth nn alone, reflecting the social planner's effective discount rate tied to demographic expansion rather than individual impatience.[13] The competitive steady state achieves the golden rule only if agents' savings propensity aligns precisely with kGRk_{GR}, typically when β(1+r(kGR))=1\beta (1 + r(k_{GR})) = 1, but deviations occur due to incomplete altruism across generations.[4] If k∗>kGRk^* > k_{GR}, the economy is dynamically inefficient (r∗<nr^* < n); if k∗<kGRk^* < k_{GR}, it is efficient but potentially under-accumulating, precluding gains from money issuance.[13] Policy interventions like capital income taxes or pay-as-you-go social security can shift toward kGRk_{GR}, though their welfare effects depend on initial conditions and whether inefficiency prevails.[32]

Pareto Optimality Conditions

In the overlapping generations (OLG) framework, a feasible allocation is Pareto optimal if no alternative feasible allocation exists that increases the utility of at least one generation without reducing the utility of any other generation. This definition extends across the infinite sequence of generations, accounting for the sequential entry and exit of cohorts. Unlike in infinitely lived representative-agent models where competitive equilibria are typically Pareto optimal under standard assumptions, OLG equilibria often fail this criterion due to missing markets for intergenerational trade and pecuniary externalities from aggregate capital decisions, which impose uninternalized costs or benefits on future cohorts.[33][12] In pure exchange OLG models without production, competitive equilibria under autarky or with non-monetary assets are generally Pareto optimal, as they clear markets period-by-period given fixed endowments and resemble Edgeworth-box allocations within overlapping cohorts. However, equilibria involving fiat money introduce potential inefficiency: the stationary monetary equilibrium may be Pareto dominated by the autarkic one if money facilitates suboptimal intertemporal transfers, though Pareto improvements are constrained by the zero initial endowment of money for the first generation.[4][34] In production-based OLG models, such as the one-sector Diamond (1965) framework with constant population growth rate n > 0, Cobb-Douglas production f(k) = k^α, depreciation rate δ, and two-period lived agents with utility u(c_t^t) + β u(c_{t+1}^{t}), the steady-state competitive equilibrium is Pareto optimal if and only if it satisfies dynamic efficiency, defined by the condition that the steady-state real interest rate r ≥ n. This ensures the capital-labor ratio k^ * ≤ k_g, the golden rule level maximizing steady-state per capita consumption, where f'(k_g) = n + δ. Equivalently, excess capital accumulation (k^ > k_g*, implying r < n) allows Pareto-improving policies, such as lump-sum taxes on capital income to reduce savings and reallocate resources toward current consumption, benefiting all generations from the initial reform onward without harming prior cohorts, as the marginal return on capital falls below the economy's growth rate.[12][35][13] To verify dynamic efficiency formally, compute the golden rule from the steady-state resource constraint c_y + c_o = f(k) - (n + δ)k, where young consumption c_y and old c_o are aggregated; maximization yields the first-order condition f'(k_g) = n + δ. In equilibrium, r = f'(k^) - δ*, derived from agents' optimality u'(c_y) = β(1 + r) u'(c_o) and market clearing k_{t+1} = s(k_t) / (1 + n), with savings s depending on wages w = f(k) - k f'(k). Parameter values yielding β(1 + n) > 1 (low discount, high growth) can produce k^ > k_g* and thus inefficiency; empirical calibrations with α ≈ 0.3, n ≈ 0.01, δ ≈ 0.05, β ≈ 0.96 often yield r > n, implying efficiency.[13][35] For transitional dynamics or non-steady-state allocations, Pareto optimality requires solving a planning problem maximizing ∑{t=0}^∞ λ_t [U_t + β U{t+1}] subject to aggregate constraints k_{t+1}(1 + n) = f(k_t) - c_t^y - c_t^o / (1 + n), where λ_t are Pareto weights ensuring feasibility and no-arbitrage across generations. First-order conditions imply intergenerational Euler equations linking marginal utilities via β(1 + n) u'(c_{t+1}^{t+1}) / u'(c_t^t) = f'(k_{t+1}) / (1 + δ), adjusted for weights; violations signal inefficiency amenable to transfers. In stochastic extensions, interim Pareto optimality equates to exchange efficiency in contingent claims equilibria, with dominant root conditions on price matrices ensuring no arbitrage opportunities across states.[36][37][38]
ConditionImplication for Pareto OptimalitySupporting Model Feature
r ≥ n in steady stateEfficient (no overaccumulation)Diamond production OLG; holds if β low or α high[35][13]
k^ > k_g* (f'(k_g) = n + δ)Inefficient; Pareto improvable via reduced savingsArises from high patience (β close to 1/(1 + n))
Monetary steady state vs. autarkyMay be dominated if money crowds out goods transfersPure exchange OLG; initial old generation constraint[4]

Policy Implications and Applications

Intergenerational Transfers and Social Security

In overlapping generations models, intergenerational transfers through pay-as-you-go (PAYGO) social security systems involve levying payroll taxes on the young to finance lump-sum benefits for the elderly, effectively redistributing resources across generations without accumulating funded assets.[4] This mechanism, analyzed in Samuelson's 1958 pure exchange framework, enables efficient intertemporal allocation by mimicking the role of money in providing a store of value when private intergenerational lending is infeasible due to the absence of double coincidence of wants.[39] Extending to production economies, Diamond's 1965 model incorporates capital accumulation, where PAYGO social security reduces the disposable income of young agents, thereby decreasing their savings and lowering the steady-state capital stock per worker.[40] The implicit rate of return on PAYGO contributions equals the population growth rate nn, as benefits depend on the ratio of workers to retirees.[13] The welfare implications hinge on dynamic efficiency: an economy is dynamically inefficient if the gross marginal product of capital f′(k)+1−δ<1+nf'(k) + 1 - \delta < 1 + n, implying overaccumulation of capital relative to the golden rule level.[4] In such cases, introducing or expanding PAYGO social security reduces excessive capital, raising steady-state welfare by shifting resources toward consumption via a higher-return transfer mechanism.[13] Conversely, in dynamically efficient equilibria where f′(k)+1−δ>1+nf'(k) + 1 - \delta > 1 + n, PAYGO distorts savings downward, potentially lowering welfare and burdening future generations with lower capital returns.[40] Empirical assessments, such as those examining postwar data, generally find real economies dynamically efficient, with marginal products of capital exceeding population growth rates, suggesting PAYGO systems may reduce long-run growth by crowding out productive investment.[32] However, low real interest rates in recent decades, around 1-2% net of depreciation in advanced economies as of 2020, have fueled debates on borderline inefficiency, though structural factors like risk premiums complicate direct comparisons to model predictions.[32] Policy analyses in OLG frameworks thus emphasize transitioning to funded systems to mitigate intergenerational inequities when efficiency holds.[41]

Monetary Neutrality and Fiat Money

In the overlapping generations (OLG) framework, fiat money—currency with no intrinsic value or backing—acquires positive value in equilibrium solely through its role as a store of value facilitating intertemporal exchange across non-overlapping generations. Agents live two periods: the young are endowed with goods for consumption and production (in extended models) but must provide for retirement, while the old hold no endowments and rely on accumulated assets. Without money, autarkic equilibria prevail, with each generation consuming only its endowment in youth and zero in old age, yielding dynamically inefficient outcomes if population grows. Fiat money resolves this by enabling the young to transfer resources to the old in exchange for claims redeemable against future young generations, as formalized in Samuelson's 1958 model where money circulates sequentially without medium-of-exchange frictions in pure exchange settings.[42][4] Monetary neutrality in OLG models refers to the proposition that changes in the nominal money supply level do not alter real allocations in stationary equilibria. A one-time proportional increase in the money stock, such as a doubling distributed as a lump-sum transfer to the old, raises the price level proportionally via the quantity theory (where velocity is constant due to fixed trading patterns), imposing an inflation tax on subsequent holders but leaving steady-state per capita consumption and real money balances unchanged once adjusted. This holds in both pure exchange and production variants, as the real return on money aligns with demographic or productivity growth rates (e.g., gross population growth 1+n1+n), independent of the absolute money level. However, neutrality is fragile: if alternative assets like capital yield higher returns, agents may abandon fiat money, rendering it valueless in non-monetary equilibria.[43][44] Superneutrality—the invariance of real variables to the money growth rate—fails in OLG models, distinguishing them from representative-agent frameworks. Steady-state money growth μ>1\mu > 1 finances seigniorage (e.g., government spending) through inflation, reducing the real return on money from 1+n1+n to approximately (1+n)/μ(1+n)/\mu, which lowers young agents' incentives to save in money as the marginal rate of substitution shifts. In pure exchange economies, this decreases real money demand, raising young consumption and lowering old consumption, effectively reducing intergenerational transfers. In Diamond-style production OLG models with capital, inflation substitutes away from capital toward money (or vice versa), altering steady-state capital accumulation via a Tobin-like effect: higher μ\mu depresses the money-capital ratio, potentially increasing capital if money crowds it out, though empirical calibration suggests contractionary impacts dominate under standard parameters. Optimal policy implies zero or negative money growth to maximize welfare, aligning the money return with the golden rule rate, as positive growth introduces deadweight losses from distorted savings.[43][45][46] These properties underscore fiat money's role as an implicit social contrivance for credit in OLG settings, but also highlight policy pitfalls: unanticipated expansions enable Pareto improvements for current generations at future expense, while anticipated growth erodes efficiency. Extensions with cash-in-advance constraints or transaction costs reinforce non-superneutrality, as inflation raises effective costs of holding money, further distorting allocations unless constraints bind uniformly across equilibria.[47][46]

Public Debt Sustainability and Fiscal Policy

In overlapping generations (OLG) models, public debt sustainability depends on the relationship between the real interest rate $ r $ and the economy's growth rate $ g $ (typically population growth $ n $ plus productivity growth in production versions). If $ r < g $, the government can maintain a stable debt-to-GDP ratio indefinitely by rolling over debt, as the expanding tax base covers interest payments without requiring proportionally rising primary surpluses. This allows a constant primary deficit to finance debt service, enabling perpetual issuance without default, unlike scenarios where $ r > g $ necessitate surpluses to prevent explosive debt dynamics.[4][48] Fiscal policy through debt issuance affects resource allocation across generations, crowding out private capital accumulation as young agents divert savings from productive assets to government bonds. In Diamond's production OLG framework, this reduces the steady-state capital stock $ k^* $, raising $ r $ endogenously and altering intergenerational transfers: current generations benefit from bond holdings and lower taxes, while future ones bear repayment via distortionary taxes or reduced wages. Sustainability limits exist; the maximum debt-to-output ratio $ b^* $ is finite, determined by factors like capital share $ \alpha $, discount factor $ \beta $, and tax distribution $ \gamma $ (e.g., $ b^* \approx 0.044 $ for $ \alpha = 0.3 $, $ \beta = 0.5 $, $ \gamma = 0.5 $), beyond which equilibria collapse toward zero capital and consumption.[49][4] Debt can enhance welfare in dynamically inefficient economies where $ r < n + g_y $ (with $ g_y $ productivity growth), as crowding out mitigates capital overaccumulation relative to the Golden Rule level $ f'(k_{GR}) = n ,potentiallyyieldingParetoimprovementsviareducedexcesssaving.However,calibratedmodelstypicallyindicatedynamicefficiency(, potentially yielding Pareto improvements via reduced excess saving. However, calibrated models typically indicate dynamic efficiency ( r > n + g_y $), implying debt reduces long-run output and welfare through persistent crowding out. Empirical observations, such as U.S. $ r \approx 1%-2% $ and $ g \approx 2.5% $ post-1945, suggest $ r < g $ holds in low-rate environments, supporting sustainability but raising risks of sudden collapses if parameters shift (e.g., via aging populations lowering $ n $). Policy must weigh these against intergenerational inequities, as debt effectively taxes unborn generations without their consent.[4][48][49]

Criticisms, Debates, and Empirical Considerations

Theoretical Limitations and Assumptions

The overlapping generations (OLG) model assumes agents live for a finite number of periods, typically two, during which the young supply inelastic labor, earn wages, save for retirement, and the old dissave entirely from accumulated assets, with generations overlapping such that both cohorts are active in goods and capital markets each period.[50] Population size grows exogenously at a constant rate nn, ensuring a stable demographic structure without endogenous fertility or mortality risks in baseline formulations.[36] Individuals maximize lifetime utility, often under separable logarithmic or constant relative risk aversion forms, subject to intertemporal budget constraints that link current savings to future consumption via market interest rates, assuming perfect capital markets and no borrowing constraints.[50] Production occurs via neoclassical aggregate technology exhibiting constant returns, diminishing marginal products, and Inada conditions for interior solutions, frequently specified as Cobb-Douglas for analytical convenience.[36] A core assumption is the absence of operative bequests or altruism toward non-overlapping descendants, which prevents infinite-horizon dynastic behavior and enables distinct predictions like non-neutrality of public debt or positive value for fiat money; however, this is theoretically fragile, as even modest intergenerational altruism—modeled via utility linkages to offspring—collapses the framework toward representative-agent infinite-horizon equivalence, undermining OLG-specific policy irrelevance results.[51] The two-period lifespan abstracts from multi-stage human development, including childhood dependency, skill accumulation, and variable retirement durations, limiting the model's capacity to incorporate realistic life-cycle dynamics such as precautionary motives or human capital depreciation.[14] Representative agent homogeneity within cohorts ignores intra-generational variation in endowments, abilities, and preferences, which can amplify inequalities or alter aggregate savings paths in ways unaccounted for in symmetric setups, reducing applicability to heterogeneous economies.[52] Deterministic environments with rational expectations and perfect foresight preclude stochastic shocks to productivity, longevity, or preferences, potentially overstating equilibrium uniqueness while understating multiplicity from sunspots or coordination failures observed in extensions.[53] Closed-economy formulations exclude international capital flows or trade, constraining analysis of open-economy spillovers despite overlapping generations' emphasis on temporal linkages.[50] These assumptions yield tractable equilibria but at the cost of generality; specific functional forms ensure closed-form solutions yet restrict robustness, as departures—like non-convexities or elastic labor—can induce corner solutions or indeterminacy absent in baseline cases.[36] The infinite-horizon aggregate with finite individual lives inherently permits dynamic inefficiency, where capital exceeds the golden rule level due to excess savings motivated by old-age needs rather than productive optima, a pathology tied to the no-altruism premise but empirically rare under plausible parameters.[50]

Debates on Dynamic Inefficiency in Real Economies

In the overlapping generations (OLG) framework, dynamic inefficiency occurs when the steady-state capital stock exceeds the golden rule level, such that the interest rate falls below the economy's growth rate, allowing Pareto-improving resource reallocation across generations via reduced capital accumulation or increased public debt.[32] Empirical debates center on whether this condition holds in real economies, with tests comparing historical asset returns to per capita output growth rates; if average returns consistently exceed growth, efficiency prevails, as excess returns would otherwise attract overinvestment.[10] Key evidence draws from long-run data series, though interpretations vary due to distinctions between safe and risky returns, stochastic shocks, and measurement of growth. A seminal study by Abel, Mankiw, Summers, and Zeckhauser (1989) analyzed data from the United States, United Kingdom, Canada, Japan, Germany, France, Italy, and Australia spanning 1870–1986, finding no evidence of dynamic inefficiency.[32] They employed Euler equation tests and compared safe rates (e.g., short-term government bonds) to growth, noting that while safe rates occasionally fell below growth (e.g., in the US averaging 1% real vs. 2% growth), equity returns averaged 6–7% real, exceeding growth by a wide margin.[10] Additionally, net investment consistently fell short of capital income shares (e.g., US profits exceeded investment every year since 1929), implying no overaccumulation of capital that could sustain inefficiency.[32] This supports dynamic efficiency, as competitive markets would erode any persistent return-growth gap through capital inflows. Recent contributions challenge these findings amid secular declines in safe rates, particularly post-2008 in advanced economies like the US, where real safe rates have averaged near zero or negative relative to 1.5–2% GDP growth, prompting arguments for inefficiency under OLG logic.[54] Geerolf (2023) reassesses AMMZ tests using updated data and alternative metrics, such as investment exceeding capital income in aggregate, concluding that the US economy exhibits dynamic inefficiency, potentially rationalizing policies like expansive public debt or pay-as-you-go pensions.[55] However, Farhi and Gourio (2023) counter that low safe rates alone are inconclusive in stochastic OLG models with production and growth uncertainty, as risk premia and incomplete markets preserve efficiency despite observed r < g for safe assets; their simulations show that empirical return-growth comparisons must account for volatility to avoid false inefficiency signals.[54] These disputes highlight methodological tensions, with efficiency tests sensitive to asset class weighting and growth definitions, though broad historical patterns favor efficiency over widespread overaccumulation.[56]

Empirical Tests and Evidence

Empirical assessments of the overlapping generations (OLG) model have centered on testing predictions of dynamic efficiency and the crowding-out effects of intergenerational transfers, such as pay-as-you-go social security systems. Direct identification remains challenging due to confounding factors like unobserved heterogeneity and general equilibrium effects, leading to reliance on reduced-form regressions, historical natural experiments, and calibrations informed by aggregate data.[57][58] A key test for dynamic inefficiency—where capital accumulation exceeds the golden rule level, implying potential gains from reduced saving—involves comparing net investment rates to after-tax capital income or safe real interest rates to economic growth rates. Abel, Mankiw, Summers, and Zeckhauser (1989) analyzed U.S. data from 1929 to 1985, finding gross investment consistently below gross capital income by over 8% of GNP annually, with dividend yields exceeding 23%, concluding dynamic efficiency and rejecting overaccumulation in the nonfinancial corporate sector and overall economy.[32] Similar patterns held in other major OECD economies from 1960 to 1984.[32] However, Geerolf (2018) reassessed these findings using updated OECD data incorporating land rents (up to 17% of GDP in Japan versus 5% in earlier estimates) and mixed incomes, determining that sufficient conditions for efficiency fail in advanced economies, with unambiguous inefficiency in Japan and South Korea due to over-accumulated capital and returns on investment below costs of capital.[55] Evidence on social security's impact supports OLG predictions of partial crowding out of private savings, as mandatory transfers reduce incentives for lifecycle accumulation absent strong bequest motives. A survey of over 100 studies finds nearly 70% report statistically significant negative effects on private wealth, with the median estimating a small but nonzero displacement.[57] Feldstein's (1974) time-series analysis of U.S. data indicated social security expansions reduced national saving by 30-50%, a result echoed in microeconometric work showing household wealth accumulation falls with expected benefits.[59][60] Historical evidence from Prussia's 19th-century pension introduction, using a difference-in-differences approach, reveals substantial crowding out of workers' savings, consistent with OLG mechanisms in pre-existing saving environments.[61] Critiques, such as computational errors noted by Leimer and Lesnoy, have tempered some estimates, but the literature consensus affirms partial rather than full offset, implying lower steady-state capital in OLG frameworks.[61][58] Tests of public debt non-neutrality in OLG settings draw indirect support from rejections of Ricardian equivalence, where households fail to fully anticipate future taxes, leading to higher consumption and lower saving in response to deficits. Calibrated OLG models incorporating distortionary taxes show debt expansions reduce welfare unless offset by efficiency gains, aligning with empirical patterns of debt-GDP ratios influencing growth without full private offsets.[62] However, aggregate evidence remains debated, with low safe rates post-2008 prompting OLG-based arguments for mild inefficiency enabling sustainable debt, though rate-growth comparisons alone prove inconclusive.[54] Overall, while OLG predictions hold qualitative traction, quantitative magnitudes vary, underscoring the model's utility for framing but not fully resolving empirical ambiguities in intergenerational resource allocation.

Modern Extensions

Heterogeneous Agents and Stochastic Shocks

Extensions of the overlapping generations (OLG) model to heterogeneous agents incorporate differences in endowments, productivity, preferences, or risk aversion across individuals within or across cohorts, departing from the assumption of identical agents in standard formulations.[63] This heterogeneity arises from factors such as varying labor skills or innate abilities, leading to dispersed income and wealth distributions that influence aggregate savings and capital accumulation.[64] For instance, models with ordinal skill groups assign agents to productivity classes where wages depend on age and skill level, generating realistic inequality patterns observable in empirical data.[65] Incorporating idiosyncratic shocks—such as random changes in individual productivity or health—further enriches these frameworks by introducing uncertainty that agents mitigate through precautionary savings and incomplete markets.[66] In stochastic OLG models, agents transition between productivity states probabilistically, affecting lifetime consumption and bequest decisions, which in turn impact intergenerational resource allocation.[66] Aggregate shocks, like economy-wide productivity fluctuations, are modeled alongside these to capture business cycle dynamics, where heterogeneous responses amplify or dampen macroeconomic volatility.[67] Equilibrium prices in such settings emerge from competitive interactions under incomplete insurance, often resulting in higher equity premia compared to representative-agent benchmarks due to uninsurable risks.[68] These extensions enable quantitative analysis of policy effects, such as progressive taxation or social security, on welfare distribution across heterogeneous groups.[64] For example, stochastic OLG models with heterogeneous agents have been employed by the U.S. Congressional Budget Office to evaluate tax reforms, revealing intra- and intergenerational trade-offs not apparent in homogeneous setups.[69] Computationally, solving these models requires numerical methods like decomposition algorithms or Markov equilibrium characterizations to handle the high dimensionality from agent diversity and stochastic processes.[70][71] Empirical calibration draws on panel data for shock persistence and heterogeneity, though challenges persist in matching long-run asset pricing anomalies without ad hoc adjustments.[63]

Computational and Applied OLG Models

Computational overlapping generations (OLG) models extend analytical frameworks by incorporating agent heterogeneity, stochastic shocks, and finite horizons, necessitating numerical solution methods to compute equilibria. These models typically discretize time into periods spanning lifetimes (e.g., 55-80 periods to represent working and retirement phases), with agents solving lifecycle problems under perfect foresight or rational expectations. Projection methods, such as those using Chebyshev polynomials or Smolyak grids, approximate policy functions and aggregate dynamics, enabling the handling of high-dimensional state spaces. Iterative algorithms, including stacked-time simulations and decomposition techniques, converge to steady states or transition paths by solving sequences of Euler equations and market-clearing conditions across cohorts.[72][73][74] The Auerbach-Kotlikoff (AK) framework, developed in the early 1980s, pioneered large-scale computational OLG analysis by modeling heterogeneous households with age-specific productivity, bequests, and government policies. Initial implementations, such as Auerbach, Kotlikoff, and Skinner (1983), used backward-forward induction to trace intergenerational transfers under tax reforms, revealing how policies like consumption taxes shift burdens across cohorts. Extensions in the 1990s incorporated elastic labor supply and progressive taxation, solved via nonlinear solvers that iterate over capital-labor ratios and factor prices. By the 2010s, AK-style models simulated 80-period economies with aggregate shocks, quantifying generational risk exposure—e.g., finding that uninsurable productivity shocks amplify welfare variance by 20-30% across birth cohorts without financial diversification.[16][75][76] In applied settings, these models inform policy evaluation by tracing long-run equilibria and transitional dynamics. For instance, simulations of U.S. Social Security privatization in AK models (e.g., Altig, Auerbach, and Kotlikoff, 1997) project GDP gains of 4-8% over decades but initial output dips from capital deepening delays, with net present value welfare varying by -2% for early retirees to +5% for future workers. Demographic applications calibrate fertility and longevity shocks, as in Auerbach and Kotlikoff (1987), forecasting fiscal gaps from aging populations—e.g., a 1% population growth decline raising debt-to-GDP ratios by 50-100% absent reforms. Gender-focused extensions, like those modeling childcare access, demonstrate how public investments boost female labor participation, yielding 1-2% annual growth in computable OLG setups with 50-year horizons. Stochastic variants, solved via parameterized expectations (e.g., Christiano-Fisher algorithm), assess climate policy transitions, revealing carbon taxes' intergenerational costs mitigated by revenue recycling into lump-sum rebates. Such analyses underscore OLG models' utility in causal policy inference, though convergence relies on grid fineness and initial guesses, with computational demands scaling exponentially in cohort size.[75][14][77]

Applications to Demographics, Climate, and Growth

OLG models incorporate demographic variables such as fertility rates, longevity, and cohort sizes to examine their effects on macroeconomic aggregates. Declining fertility and increasing life expectancy lead to population aging, which raises the old-age dependency ratio and shifts resources from investment to retiree consumption, potentially lowering savings rates and long-run capital accumulation. Simulations in OLG frameworks for OECD economies project that such aging reduces GDP growth through diminished labor supply and strained public pension systems, with early retirement options exacerbating fiscal pressures.[14] Endogenous fertility decisions, integrated into agents' utility maximization, respond to economic incentives like child-rearing costs and pension benefits, influencing aggregate human capital formation and intergenerational transfers.[14] In climate policy applications, OLG models capture the absence of complete markets between non-overlapping generations, leading to underinvestment in mitigation as finitely lived agents discount future damages without altruism spanning unborn cohorts. This structure reveals distributional conflicts hidden in representative agent models, where private impatience diverges from social optimality, and asset prices—such as rents from scarce resources like fossil fuels—transmit intergenerational effects via life-cycle savings. Fiscal policies, including carbon taxes recycled through public debt or transfers, serve as redistribution mechanisms to align efficiency with equity, breaking Ricardian equivalence and enabling higher welfare gains than lump-sum alternatives.[78] Peer-reviewed extensions emphasize that climate damages acting on production factors necessitate dual policy instruments to address both externalities and intra-generational disparities.[78] For economic growth, OLG models extended with production functions analyze steady-state capital intensity and transition dynamics under decentralized decisions. In Diamond's 1965 formulation, young agents inelastically supply labor, allocate wages between consumption and savings invested as capital, while firms produce under constant returns to scale, determining factor prices via marginal products. The resulting equilibrium equates aggregate savings to net investment, yielding a unique steady-state capital-labor ratio below the golden rule level that maximizes consumption per capita, implying potential Pareto improvements from forced savings policies.[3] These applications highlight how fiscal distortions or technological shocks affect growth paths, with recent variants exploring multiplicity of steady states and endogenous transitions influenced by demographic factors.[79]

References

  1. 27. The Overlapping Generations Model
  2. Overlapping Generations: The First Jubilee
  3. [PDF] National Debt in a Neoclassical Growth Model
  4. [PDF] The Overlapping Generations Model
  5. Overlapping Generations - an overview | ScienceDirect Topics
  6. Overlapping Generations (OLG) Model - INOMICS
  7. The Overlapping Generations Model in 1947 - jstor
  8. Overlapping Generations - Oxford Academic - Oxford University Press
  9. An Exact Consumption-Loan Model of Interest with or without the ...
  10. Overlapping Generations: The First Jubilee
  11. http://www.jstor.org/stable/1809231
  12. [PDF] The Diamond (1965) OLG Model - JHU Economics
  13. [PDF] Diamond Overlapping Generations Model 1 Setting 2 Equilibrium
  14. [PDF] Use of Overlapping Generations Model in Modeling Demographic ...
  15. [PDF] 14.452 Economic Growth: Lecture 8, Overlapping Generations
  16. [PDF] Using Dynamic General Equilibrium Models for Policy Analysis
  17. [PDF] Part 2: Overlapping Generations Models - LSE Press
  18. [PDF] The Overlapping-Generations Model, I: The Case of Pure - Karl Shell
  19. [PDF] Chapter 9 Overlapping Generations Models - Jose-Elias Gallegos
  20. [PDF] An Exact Consumption-Loan Model of Interest with or without the ...
  21. [PDF] The Overlapping-Generations Model, II: The Case of Pure - Karl Shell
  22. [PDF] The Diamond Model: Overlapping Generations
  23. A Two-Sector Overlapping-Generations Model - jstor
  24. Cyclical equilibria in multi-sector productive economies: The role of ...
  25. [PDF] Research Article Two-sector overlapping generations model with ...
  26. an Analysis Using a Multi-Sector OLG Model in an Open Economy
  27. [PDF] Endogenous Fertility in an Overlapping Generations Growth Model
  28. [PDF] Endogenous Fertility in an Overlapping Generations Growth Model
  29. Fertility and Wage Rates in an Overlapping-Generations Model - jstor
  30. [PDF] Property Rights and Efficiency in OLG Models with Endogenous ...
  31. [PDF] An Overlapping-Generations Model with Endogenous Fertility
  32. [PDF] Assessing Dynamic Efficiency: Theory and Evidence
  33. [PDF] 14.452 Economic Growth: Lecture 8, Overlapping Generations
  34. [PDF] Chapter 4 The overlapping generations (OG) model
  35. [PDF] Overlapping Generations Model - fep.up.pt
  36. [PDF] 14.452 Economic Growth: Lecture 7, Overlapping Generations
  37. Dominant root characterization of Pareto optimality and the ...
  38. [PDF] Dynamic Efficiency and Pareto Optimality in a Stochastic OLG Model ...
  39. [PDF] Pay-As-You-Go Social Security and the Aging of America
  40. [PDF] 14.452 Economic Growth: Lecture 8, Overlapping Generations
  41. Social security, growth, and welfare in overlapping generations ...
  42. [PDF] 2. The Overlapping Generations Model of Fiat Money - We will now ...
  43. [PDF] The Role of Overlapping—Generations Models in ionetary Economics
  44. The Overlapping Generations Model of Money
  45. Monetary policy rules in an OLG model with non-superneutral money
  46. overlapping generations model: neutrality and optimality of ... - jstor
  47. Chapter 7 Overlapping generations models with money and ...
  48. 10 Public Debt | Intermediate Macroeconomics
  49. [PDF] i Maximum Sustainable Government Debt in the Overlapping ...
  50. [PDF] Infinite-Horizon and Overlapping-Generations Models
  51. NBER WORKING PAPER SERIES OVERLAPPING GENERATIONS ...
  52. [PDF] Overlapping Generations Models and Graded Age-Set Societies
  53. The economics of indeterminacy in overlapping generations models
  54. Low safe interest rates: A case for dynamic inefficiency?
  55. [PDF] Reassessing Dynamic Efficiency - François Geerolf
  56. [PDF] A case for dynamic inefficiency? - IRIS
  57. Does Social Security Crowd out Private Wealth? A Survey of the ...
  58. [PDF] Does Social Security Crowd Out Private Savings?
  59. [PDF] An Examination of Empirical Tests of Social Security and Savings¹
  60. The ideas and influence of Martin Feldstein, 1939-2019 | CEPR
  61. Does Social Security Crowd Out Private Savings? The Case of ... - jstor
  62. [PDF] Public Debt in Calibrated OLG Models: Fiscal Arithmetic versus ...
  63. [PDF] The Implications of Heterogeneity and Finite Lives for Asset Pricing
  64. Analyzing Fiscal Policies in a Heterogeneous-Agent Overlapping ...
  65. [XLS] Summary of OLG Models - SOA
  66. Stochastic Overlapping Generations Models - SpringerLink
  67. Stochastic Analysis of Overlapping Generations Models Under ...
  68. A Long-Lived, Heterogeneous Agent, Overlapping Generations Model
  69. Analyzing Tax Policy Changes Using a Stochastic OLG Model with ...
  70. [PDF] Computation of Equilibria in OLG Models with Many Heterogeneous ...
  71. A characterization of Markov equilibrium in stochastic overlapping ...
  72. [PDF] Solving Stochastic OLG Models Using Chebyshev Parameterized ...
  73. Computing equilibrium in OLG models with stochastic production
  74. [PDF] Computation of Equilibria in OLG Models with Many Heterogeneous ...
  75. [PDF] David Altig - Alan J. Auerbach Laurence J. Kotlikoff
  76. [PDF] Generational Risk–Is It a Big Deal? Simulating an 80-Period OLG ...
  77. A COMPUTABLE OVERLAPPING GENERATIONS MODEL FOR ...
  78. [PDF] Why OLG models are needed in climate economics - IAMC
  79. Overlapping Generations Models, Multiplicity of Steady States and ...
User Avatar
No comments yet.