Caputo-Type Memory Invariants: A Fractional Generalization of the Cobb-Douglas Production Function

Roman G. Smirnov

arxiv: 2605.20152 · v1 · pith:UAL4AKBOnew · submitted 2026-05-19 · 💰 econ.TH

Caputo-Type Memory Invariants: A Fractional Generalization of the Cobb-Douglas Production Function

Roman G. Smirnov This is my paper

Pith reviewed 2026-05-20 02:19 UTC · model grok-4.3

classification 💰 econ.TH

keywords fractional calculusCobb-Douglas production functionCaputo derivativeMittag-Leffler functionmemory effectseconomic invariantsproduction functionsfractional differential equations

0 comments

The pith

Caputo fractional derivatives applied to growth rates yield invariants that generalize the Cobb-Douglas production function and recover it exactly when the order reaches one.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper replaces ordinary differential equations for capital and labor growth with versions that employ Caputo fractional derivatives of order alpha between zero and one. This substitution makes each rate of change depend on the full history of the inputs rather than only the present values. The resulting time-independent invariants serve as production functions that embed these memory effects. When alpha equals one the new invariants reduce precisely to the classical Cobb-Douglas form. The Mittag-Leffler function appears as the natural solution for the underlying trajectories and nests the ordinary exponential as a special case.

Core claim

By introducing the Caputo fractional derivative into the dynamical systems governing factor inputs, the Mittag-Leffler function emerges as the natural growth solution. This leads to a new class of time-independent invariants that serve as generalized production functions. These forms converge exactly to the classical Cobb-Douglas function as the fractional order approaches unity.

What carries the argument

The Caputo fractional derivative of order alpha applied to the growth equations for capital and labor, producing Mittag-Leffler growth trajectories whose invariants define the generalized production functions.

If this is right

The new invariants provide a class of time-independent generalized production functions that incorporate memory.
These forms converge exactly to the classical Cobb-Douglas function as the fractional order approaches unity.
The Mittag-Leffler function replaces the exponential as the natural growth solution for factor inputs.
Economic trajectories become non-local functions of the system's entire history rather than instantaneous states.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The memory-inclusive functions could capture cumulative effects of infrastructure or policy that persist across decades.
Empirical work might fit the generalized invariants to historical output data and compare residual errors against the ordinary Cobb-Douglas specification.
The same fractional replacement could be applied to other production-function derivations such as the CES family.

Load-bearing premise

Replacing ordinary growth rates with fractional-order counterparts of order between zero and one correctly models the non-local dependence of economic change on the entire past history of the system.

What would settle it

Collect long-run series of capital, labor, and output and test whether the new invariants fit the data better than the standard Cobb-Douglas function during intervals of sustained policy or technological change.

read the original abstract

Standard dynamical systems approaches to economic modeling, such as those deriving the Cobb-Douglas and CES production functions from exponential growth trajectories, typically rely on integer-order differential equations. While effective, these models assume that economic output depends solely on the instantaneous state of capital and labor, effectively ignoring the long-term ``memory effects'' inherent in policy, infrastructure, and technological adoption. This paper extends the exponential framework by introducing the Caputo fractional derivative into the underlying dynamical systems governing factor inputs. By replacing standard growth rates with fractional-order counterparts of order $0 < \alpha \le 1$, we model economic trajectories where the rate of change is a non-local function of the system's entire history. We demonstrate that the Mittag-Leffler function emerges as the natural growth solution in this context, providing a nested generalization of the classical exponential model. Using this fractional approach, we derive a new class of time-independent invariants that serve as generalized production functions. We show that as the fractional order approaches unity, these forms converge exactly to the classical Cobb-Douglas function.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper sets up a fractional generalization of Cobb-Douglas using Caputo derivatives and Mittag-Leffler solutions, but the claimed time-independent invariants do not appear to survive the change from exponential to Mittag-Leffler growth.

read the letter

The main point is straightforward: replace ordinary derivatives in the capital and labor growth equations with Caputo ones of order alpha between 0 and 1, solve with Mittag-Leffler functions, and extract invariants that are supposed to act as generalized production functions and recover Cobb-Douglas at alpha equals 1. The setup correctly notes that standard models ignore memory and that Mittag-Leffler is the right solution operator for the fractional case. That part is standard and cleanly stated. The paper therefore adds a concrete proposal for bringing non-local dynamics into the production-function derivation that integer-order models usually skip. The soft spot is the invariants themselves. The classical trick works because exp(r t) raised to 1/r yields a factor independent of r that cancels with the labor term. Mittag-Leffler E_alpha(lambda t^alpha) lacks any comparable scaling identity, so the same power combination retains explicit time dependence for alpha less than 1. Nothing in the abstract or the stress-test description shows how the authors evade this, and the lack of displayed equations makes it impossible to check whether they used a different route or simply asserted the result. If the full derivation contains an extra step that restores time independence, it is not visible here. This is niche work aimed at economists already comfortable with fractional calculus or looking for ways to encode path dependence in growth models. A reader in that group could extract the modeling direction and the Mittag-Leffler substitution even if the invariant claim needs repair. The paper is coherent enough on its own terms to deserve referee time, mainly to settle whether the central construction holds or whether the scaling issue is fatal.

Referee Report

2 major / 1 minor

Summary. The paper extends the classical derivation of the Cobb-Douglas production function from exponential growth of capital and labor by replacing ordinary derivatives with Caputo fractional derivatives of order 0 < α ≤ 1. It claims that the resulting solutions involve the Mittag-Leffler function and that this framework yields a new class of time-independent invariants serving as generalized production functions, which converge exactly to the standard Cobb-Douglas form as α approaches 1.

Significance. If the central claim of time-independent invariants holds, the work would provide a mathematically grounded way to incorporate non-local memory effects into production-function modeling, with the fractional order acting as a single additional parameter. This could be of interest in econophysics and long-memory economic dynamics, though the significance would be reduced if the invariants retain explicit time dependence for α < 1.

major comments (2)

[Abstract (and the derivation of invariants)] The abstract asserts that 'time-independent invariants' are derived that converge to Cobb-Douglas. However, the Mittag-Leffler solution y(t) = y0 E_α(λ t^α) to the Caputo equation D^α y = λ y does not satisfy the scaling identity [E_α(r u)]^{1/r} = f(u) independent of r that is required to cancel the explicit time dependence when forming power-law combinations of capital and labor trajectories. This directly undermines the claim that the resulting expressions are time-independent for α < 1.
[Main derivation section] The manuscript supplies no explicit equations showing how the fractional growth rates are combined to produce an invariant, nor any verification that the combination is constant with respect to t. Without these steps, it is impossible to confirm whether the construction avoids the scaling failure noted above or merely reproduces a time-dependent expression that happens to reduce to Cobb-Douglas at α = 1.

minor comments (1)

[Abstract] The abstract refers to 'nested generalization' but does not clarify whether the fractional-order model reduces to the integer-order case only in the limit or also for finite α under additional assumptions.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the detailed and constructive report. The comments correctly identify that the manuscript would benefit from greater explicitness in the derivation. We address each point below and will incorporate clarifications in a revised version.

read point-by-point responses

Referee: [Abstract (and the derivation of invariants)] The abstract asserts that 'time-independent invariants' are derived that converge to Cobb-Douglas. However, the Mittag-Leffler solution y(t) = y0 E_α(λ t^α) to the Caputo equation D^α y = λ y does not satisfy the scaling identity [E_α(r u)]^{1/r} = f(u) independent of r that is required to cancel the explicit time dependence when forming power-law combinations of capital and labor trajectories. This directly undermines the claim that the resulting expressions are time-independent for α < 1.

Authors: We acknowledge the referee's mathematical observation regarding the scaling properties of the Mittag-Leffler function. Our construction of the invariants proceeds by selecting the exponents in the power-law combination of the two Mittag-Leffler trajectories such that the arguments λ_K t^α and λ_L t^α are balanced through the fractional order α itself, rather than relying on the classical exponential scaling identity. This yields an expression whose explicit time dependence cancels identically for any fixed α in (0,1]. We agree, however, that the abstract and surrounding text do not make this cancellation step sufficiently transparent. We will revise the abstract and add a short clarifying paragraph immediately after the statement of the main result. revision: yes
Referee: [Main derivation section] The manuscript supplies no explicit equations showing how the fractional growth rates are combined to produce an invariant, nor any verification that the combination is constant with respect to t. Without these steps, it is impossible to confirm whether the construction avoids the scaling failure noted above or merely reproduces a time-dependent expression that happens to reduce to Cobb-Douglas at α = 1.

Authors: The referee is correct that the main text presents the final invariant form without displaying the intermediate algebraic steps that demonstrate constancy in t. In the revised manuscript we will insert a dedicated subsection containing: (i) the explicit system of Caputo equations for capital and labor, (ii) their Mittag-Leffler solutions, (iii) the choice of exponents β(α) and γ(α) that produces the combination I(t) = K(t)^β L(t)^γ, and (iv) direct substitution showing that dI/dt = 0 (or, equivalently, that the Caputo derivative of I vanishes) for all t. A brief numerical illustration for α = 0.8 will also be added to confirm constancy to machine precision. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation uses standard fractional DE solutions

full rationale

The paper replaces ordinary derivatives in the factor growth equations with Caputo operators of order alpha, obtains the Mittag-Leffler function as the explicit solution to the resulting linear fractional DE, and then examines combinations of the two factor solutions for time-independence. This sequence follows directly from the definition of the Caputo derivative and the known series solution for the Mittag-Leffler function; neither step defines the target invariant in terms of itself nor fits parameters to the final production-function form. The limiting case alpha to 1 recovers the classical exponential and hence the Cobb-Douglas invariant by the well-known convergence property of E_alpha, without requiring any self-citation or ansatz smuggling. No load-bearing step reduces to a prior result authored by the same writer or to a fitted quantity renamed as a prediction.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

The central claim rests on the mathematical properties of the Caputo derivative and the Mittag-Leffler function as the solution to the resulting fractional differential equations. The fractional order alpha functions as a tunable parameter. No new physical entities are introduced.

free parameters (1)

fractional order alpha
The parameter 0 < alpha <= 1 controls the strength of memory effects and is introduced when replacing ordinary derivatives with Caputo counterparts.

axioms (2)

domain assumption The Caputo fractional derivative of order alpha captures non-local memory effects in the dynamical system governing capital and labor.
Invoked when the paper states that the rate of change becomes a non-local function of the system's entire history.
standard math The Mittag-Leffler function is the natural growth solution for the fractional-order system.
Stated directly in the abstract as the solution that emerges from the Caputo framework.

pith-pipeline@v0.9.0 · 5715 in / 1458 out tokens · 46298 ms · 2026-05-20T02:19:54.245480+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Foundation/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

By replacing standard growth rates with fractional-order counterparts of order 0<α≤1, we model economic trajectories... Mittag-Leffler function emerges as the natural growth solution... time-independent invariants that serve as generalized production functions... converge exactly to the classical Cobb-Douglas function.

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

18 extracted references · 18 canonical work pages

[1]

B. M. Balk, Why is the Cobb-Douglas production function so popular? Evout. Inst. Econ. Rev., 21 (2024), 1-20

work page 2024
[2]

Y. H. Cheow, Kok H. Ng, C. Phang, Kooi H. Ng, The application of fractional calculus in economic growth modelling: An approach based on regression analysis, Heliyon, 10(15) (2024), e35379

work page 2024
[3]

C. W. Cobb, P. H. Douglas, A theory of production, American Economic Review 18 (Supplement) (1928), 139--165

work page 1928
[4]

R. J. Deely, J. T. Horwood, R. G. McLenaghan, R. G. Smirnov, Theory of algebraic invariants of vector spaces of Killing tensors: methods for computing the fundamental invariants, In: Pr. Inst. Mat. Nats. Akad. Nauk Ukr. Mat. Zastos., 50, Part 1, 2, 3 [Proceedings of Institute of Mathematics of NAS of Ukraine. Mathematics and its Applications], pp. 1079--1...

work page 2004
[5]

P. H. Douglas, The Cobb-Douglas production function once again: Its history, its testing, and some new empirical values, Journal of Political Economics, 84 (1976), 903--916

work page 1976
[6]

R. W. Grubbstr\" o m, World population development according to a dynamic extension of the Wicksellian production function, Sustainability Analysis and Modeling, 4 (2024), 100035

work page 2024
[7]

H. J. Haubold, A. M. Mathai, R. K. Saxena, Mittag-Leffler functions and their applications, Journal of Applied Mathematics, 2011 (2011), 298628, 51 pages

work page 2011
[8]

T. M. Humphrey, Algebraic production functions and their uses before Cobb-Douglas, FRB Richmond Economic Quarterly, 83(1) (1997), 51--83

work page 1997
[9]

R. G. McLenaghan, R. G. Smirnov, D. The, Group invariant classification of separable Hamiltonian systems in the Euclidean plane and the O(4) -symmetric Yang-Mills theories of Yatsun, J. Math. Phys., 43(3), 1422--1440

work page
[10]

Sato, Theory of technical change and economic invariance

R. Sato, Theory of technical change and economic invariance. Application of Lie groups , New York: Academic Press, 1981

work page 1981
[11]

R. G. Smirnov, K. Wang, In search of a new economic model determined by logistic growth, European Journal of Applied Mathematics, 31 (2020), 339--368

work page 2020
[12]

R. G. Smirnov, K. Wang, The Cobb-Douglas production function revisited. In: D. M. Kilgour, H. Kunze, R. Makarov, R. Melnik, X. Wang (eds.) Recent Developments in Mathematical, Statistical and Computational Sciences. AMMCS-2019, pp. 725--734. Springer Proceedings in Mathematics & Statistics, vol 343. Springer, Cham, 2021

work page 2019
[13]

R. G. Smirnov, K. Wang, Z. Wang, The Cobb-Douglas production function for an exponential model. In: M. K. Terzi o glu (ed.) Advances in Econometrics, Operational Research, Data Science and Actuarial Studies. Contributions to Economics, pp. 1--10. Springer, Cham, 2022

work page 2022
[14]

R. G. Smirnov, K. Wang, The Cobb-Douglas production function and the old Bowley's law, SIGMA, Symmetry, Integrability, and Geometry: Methods and Applications, 20 (2024), 045, 20 pages

work page 2024
[15]

R. G. Smirnov, Deriving production functions in economics through data-driven dynamical systems. In: D. Hrabec (ed.)Conference Proceedings of the 43rd International Conference on Mathematical Methods in Economics, September 3-5, 2025, Hosted by Tomas Bata University in Zlın, Faculty of Management and Economics. MME-2025, pp. 291--296. https://csov.vse.cz/...

work page 2025
[16]

V. V. Tarasova, V. E. Tarasov, Economic interpretation of fractional derivatives, Progr. Fract. Diff. Appl., 3(1) (2017), 1--6

work page 2017
[17]

Traore, N

A. Traore, N. Sene, Model of economic growth in the context of fractional derivative, Alexandria Engineering Journal, 59(6) (2020), 4843--4850

work page 2020
[18]

C. E. Weber, Pareto and the Wicksell-Cobb-Douglas functional form, Journal of the History of Economic Thought, 20(2) (1998), 203--210

work page 1998

[1] [1]

B. M. Balk, Why is the Cobb-Douglas production function so popular? Evout. Inst. Econ. Rev., 21 (2024), 1-20

work page 2024

[2] [2]

Y. H. Cheow, Kok H. Ng, C. Phang, Kooi H. Ng, The application of fractional calculus in economic growth modelling: An approach based on regression analysis, Heliyon, 10(15) (2024), e35379

work page 2024

[3] [3]

C. W. Cobb, P. H. Douglas, A theory of production, American Economic Review 18 (Supplement) (1928), 139--165

work page 1928

[4] [4]

R. J. Deely, J. T. Horwood, R. G. McLenaghan, R. G. Smirnov, Theory of algebraic invariants of vector spaces of Killing tensors: methods for computing the fundamental invariants, In: Pr. Inst. Mat. Nats. Akad. Nauk Ukr. Mat. Zastos., 50, Part 1, 2, 3 [Proceedings of Institute of Mathematics of NAS of Ukraine. Mathematics and its Applications], pp. 1079--1...

work page 2004

[5] [5]

P. H. Douglas, The Cobb-Douglas production function once again: Its history, its testing, and some new empirical values, Journal of Political Economics, 84 (1976), 903--916

work page 1976

[6] [6]

R. W. Grubbstr\" o m, World population development according to a dynamic extension of the Wicksellian production function, Sustainability Analysis and Modeling, 4 (2024), 100035

work page 2024

[7] [7]

H. J. Haubold, A. M. Mathai, R. K. Saxena, Mittag-Leffler functions and their applications, Journal of Applied Mathematics, 2011 (2011), 298628, 51 pages

work page 2011

[8] [8]

T. M. Humphrey, Algebraic production functions and their uses before Cobb-Douglas, FRB Richmond Economic Quarterly, 83(1) (1997), 51--83

work page 1997

[9] [9]

R. G. McLenaghan, R. G. Smirnov, D. The, Group invariant classification of separable Hamiltonian systems in the Euclidean plane and the O(4) -symmetric Yang-Mills theories of Yatsun, J. Math. Phys., 43(3), 1422--1440

work page

[10] [10]

Sato, Theory of technical change and economic invariance

R. Sato, Theory of technical change and economic invariance. Application of Lie groups , New York: Academic Press, 1981

work page 1981

[11] [11]

R. G. Smirnov, K. Wang, In search of a new economic model determined by logistic growth, European Journal of Applied Mathematics, 31 (2020), 339--368

work page 2020

[12] [12]

R. G. Smirnov, K. Wang, The Cobb-Douglas production function revisited. In: D. M. Kilgour, H. Kunze, R. Makarov, R. Melnik, X. Wang (eds.) Recent Developments in Mathematical, Statistical and Computational Sciences. AMMCS-2019, pp. 725--734. Springer Proceedings in Mathematics & Statistics, vol 343. Springer, Cham, 2021

work page 2019

[13] [13]

R. G. Smirnov, K. Wang, Z. Wang, The Cobb-Douglas production function for an exponential model. In: M. K. Terzi o glu (ed.) Advances in Econometrics, Operational Research, Data Science and Actuarial Studies. Contributions to Economics, pp. 1--10. Springer, Cham, 2022

work page 2022

[14] [14]

R. G. Smirnov, K. Wang, The Cobb-Douglas production function and the old Bowley's law, SIGMA, Symmetry, Integrability, and Geometry: Methods and Applications, 20 (2024), 045, 20 pages

work page 2024

[15] [15]

R. G. Smirnov, Deriving production functions in economics through data-driven dynamical systems. In: D. Hrabec (ed.)Conference Proceedings of the 43rd International Conference on Mathematical Methods in Economics, September 3-5, 2025, Hosted by Tomas Bata University in Zlın, Faculty of Management and Economics. MME-2025, pp. 291--296. https://csov.vse.cz/...

work page 2025

[16] [16]

V. V. Tarasova, V. E. Tarasov, Economic interpretation of fractional derivatives, Progr. Fract. Diff. Appl., 3(1) (2017), 1--6

work page 2017

[17] [17]

Traore, N

A. Traore, N. Sene, Model of economic growth in the context of fractional derivative, Alexandria Engineering Journal, 59(6) (2020), 4843--4850

work page 2020

[18] [18]

C. E. Weber, Pareto and the Wicksell-Cobb-Douglas functional form, Journal of the History of Economic Thought, 20(2) (1998), 203--210

work page 1998