arxiv: 2604.22976 · v1 · submitted 2026-04-24 · ⚛️ physics.soc-ph · cond-mat.stat-mech· q-fin.ST

Recognition: unknown

Statistical Mechanics of Household Income and Wealth: Derivation from Firm Dynamics via Maximum Entropy and Mixture Aggregation

Robert T. Nachtrieb

Authors on Pith no claims yet

Pith reviewed 2026-05-08 09:26 UTC · model grok-4.3

classification ⚛️ physics.soc-ph cond-mat.stat-mechq-fin.ST

keywords income distributionwealth distributionfirm dynamicsmaximum entropyPareto tailGibrat's lawZipf distributionmixture aggregation

0 comments

The pith

The two-class income and wealth structure is derived from firm dynamics using maximum entropy and mixture aggregation.

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

The paper establishes that the exponential bulk and Pareto tail observed in income and wealth distributions follow directly from the growth statistics of firms. Gibrat's law produces a Zipf distribution of firm sizes. Maximum entropy applied to wages inside each firm, when mixed across firms, produces the exponential income distribution for employees. Owner wealth follows multiplicative returns whose scaling with firm size sets the Pareto exponent, while employee wealth follows additive dynamics with a reflecting barrier at zero. If correct, this supplies a parameter-free link from measurable firm-value scaling to the wealth tail exponent and to the typical wealth-to-income ratio of lower-class households.

Core claim

The distribution of income and wealth exhibits a Boltzmann-Gibbs exponential form for the lower 97% and a Pareto power-law for the upper 3%. This structure is obtained by applying maximum entropy to wages within firms whose sizes follow from Gibrat's law, then aggregating, and using multiplicative returns for owner wealth whose scaling with firm size sets the tail exponent to 1/θ where θ comes from V ~ s^{0.77}, reproducing α_w=1.3 and ζ=-0.23.

What carries the argument

Mixture aggregation of maximum-entropy wage distributions over a Zipf firm-size distribution, with owner wealth following multiplicative returns scaled by θ and employee wealth following additive noise with a reflecting barrier.

If this is right

The wealth-to-income ratio T_w/T_y for lower-class households equals roughly 1.7 years and varies with the long-run savings rate and tax parameters.
Firms near zero profit exhibit cash martingales whose first-passage exit rate is proportional to t^{-1/2}; convolution with the Zipf firm-size distribution yields an aggregate exit rate proportional to t^{-1/2} log t.
The returns-per-employee size exponent is fixed at ζ = θ − 1 ≈ −0.23 by the same value-size relation that sets α_w.
The model predicts that cross-country differences in savings and tax rates will produce corresponding differences in the observed wealth-to-income ratio.

Where Pith is reading between the lines

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

Variations in the firm value-size scaling exponent across economies or time periods would directly predict corresponding changes in the wealth Pareto exponent α_w.
The same mixture construction implies that policies altering the firm-size distribution (for example through entry barriers) would shift both the income bulk temperature and the relative size of the Pareto tail.
The first-passage exit calculation supplies a parameter-free benchmark that can be checked against any longitudinal firm registry covering small establishments.

Load-bearing premise

Maximum entropy determines the within-firm wage distribution subject only to mean constraints, and owner wealth obeys multiplicative returns whose size scaling is captured exactly by the observed value-size exponent θ.

What would settle it

Direct measurement of firm cash holdings near zero profit showing first-passage exit times that, when convolved with the Zipf size distribution, produce an overall firm exit rate scaling as t^{-1/2} log t against longitudinal establishment data.

Figures

Figures reproduced from arXiv: 2604.22976 by Robert T. Nachtrieb.

**Figure 1.** Figure 1: FIG. 1. (a) Three-sector money flow diagram. Solid arrows: flows determining the distributions (wages, consumption, capital view at source ↗

read the original abstract

The distribution of income and wealth in developed economies exhibits a robust two-class structure: an exponential (Boltzmann--Gibbs) bulk covering $\sim\!97\%$ of the population, and a power-law (Pareto) tail in the upper $\sim\!3\%$. We derive this structure from first principles via an explicit mechanistic chain: Gibrat's law for firm growth implies a Zipf firm-size distribution; maximum entropy applied to within-firm wages, combined with mixture aggregation across firms, yields a Boltzmann--Gibbs income distribution with temperature $T_y$ for employees; additive-noise wealth dynamics with a reflecting wall at zero produce a Boltzmann--Gibbs employee wealth distribution with temperature $T_w$. For firm owners, multiplicative capital returns produce a Pareto wealth tail with exponent $\alpha_w = 1/\theta$, where $\theta$ encodes how total returns scale with firm size. The empirical value $\alpha_w \approx 1.30$ \cite{Yakovenko2009} is reproduced with no tuned parameters from the observed firm value scaling $V = V_0(s/s_0)^{0.77}$~\cite{Axtell2001}, and simultaneously yields the first quantitative estimate of the returns-per-employee size exponent: $\zeta = \theta - 1 \approx -0.23$. For empirical values $\nu \approx 0.3$, $c \approx 0.81$, $k \approx 0.15$ (BEA long-run savings rate $\approx 5\%$), the model gives $T_w/T_y \approx 1.7\,\text{yr}$, i.e.\ lower-class households hold roughly 1--2 years of income as wealth, with the precise ratio depending on savings and tax rates and testable cross-country. As a parameter-free empirical test, firms near zero profit have a cash martingale whose first-passage time gives establishment exit rate $\sim t^{-1/2}$; convolving with the Zipf firm-size distribution yields firm-level exit rate $\sim t^{-1/2}\!\log t$, with apparent exponent $b = 0.295 \pm 0.03$, confirmed against BDS firm-age data with no free parameters.

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 derives the two-class income and wealth structure from firm growth rules and matches the Pareto exponent to firm value scaling data with no free parameters, though the central owner-returns step needs explicit verification of the stochastic process.

read the letter

The main takeaway is a mechanistic chain from Gibrat's law on firm sizes through maximum-entropy wage mixtures to the exponential bulk for most households, with owners getting a Pareto tail whose exponent comes directly from the observed firm value scaling V ~ s^0.77. That gives alpha_w ≈ 1.3 with no tuning and also produces zeta ≈ -0.23 for size-dependent returns per employee. The separate check on firm exit rates from near-zero-profit martingales, convolved with the Zipf distribution to match BDS age data, is a clean parameter-free test that stands on its own.

Referee Report

1 major / 2 minor

Summary. The paper derives the two-class structure of income and wealth distributions (exponential Boltzmann-Gibbs bulk for ~97% of the population and Pareto tail for the upper ~3%) from firm dynamics. Gibrat's law implies a Zipf firm-size distribution; maximum entropy applied to within-firm wages with mean constraints, followed by mixture aggregation, yields the BG income distribution; additive noise with a reflecting barrier at zero produces BG employee wealth; multiplicative capital returns for owners yield a Pareto wealth tail with exponent α_w = 1/θ, where θ encodes size-dependent returns scaling. The observed firm-value scaling V = V_0 (s/s_0)^{0.77} is used to obtain α_w ≈ 1.30 with no free parameters, simultaneously estimating the returns-per-employee exponent ζ = θ - 1 ≈ -0.23. The model also predicts T_w/T_y ≈ 1.7 yr for empirical savings/tax parameters and provides a parameter-free test of firm exit rates ~ t^{-1/2} log t matching BDS data.

Significance. If the central derivations hold, the work is significant for providing an explicit mechanistic chain from firm growth and value scaling to household distributions via maximum entropy and mixture aggregation, reproducing the empirical Pareto exponent parameter-free from independent firm data and yielding a new quantitative estimate for size-dependent returns. Strengths include the use of observed scaling relations without tuning, the falsifiable firm-exit prediction, and the cross-country testable wealth-to-income ratio. This approach strengthens the statistical-mechanics framing of economic distributions by grounding it in firm-level observables.

major comments (1)

[Abstract and owner-wealth section] Abstract and owner-wealth derivation: The headline result that α_w ≈ 1.30 is obtained parameter-free from the firm-value exponent 0.77 via α_w = 1/θ requires an explicit mapping from the observed V ~ s^{0.77} to the stationary tail of the multiplicative returns process (dw/w = f(s) dt + g(s) dW). The abstract states that θ 'encodes how total returns scale with firm size' but does not detail the stochastic dynamics or prove that the inverse relation holds without extra parameters on drift or diffusion; this step is load-bearing for the parameter-free claim and the simultaneous ζ estimate.

minor comments (2)

[Income distribution derivation] The description of mixture aggregation across firms to obtain the overall BG income distribution is sketched but would benefit from an explicit equation showing how the firm-size distribution (Zipf) convolves with the within-firm wage distribution.
[Notation and definitions] Notation for the returns-per-employee exponent ζ = θ - 1 is introduced without a dedicated equation defining θ from the value-size relation; adding this would improve clarity.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the careful reading and constructive feedback, as well as for recognizing the significance of the mechanistic derivation from firm dynamics. We address the single major comment below with additional clarification on the owner-wealth mapping and indicate the revisions we will make to improve explicitness while preserving the parameter-free character of the result.

read point-by-point responses

Referee: [Abstract and owner-wealth section] Abstract and owner-wealth derivation: The headline result that α_w ≈ 1.30 is obtained parameter-free from the firm-value exponent 0.77 via α_w = 1/θ requires an explicit mapping from the observed V ~ s^{0.77} to the stationary tail of the multiplicative returns process (dw/w = f(s) dt + g(s) dW). The abstract states that θ 'encodes how total returns scale with firm size' but does not detail the stochastic dynamics or prove that the inverse relation holds without extra parameters on drift or diffusion; this step is load-bearing for the parameter-free claim and the simultaneous ζ estimate.

Authors: We agree that the abstract is concise and would benefit from greater explicitness on this load-bearing step. In the full manuscript the mapping is constructed as follows: the empirical firm-value relation V(s) ∝ s^{0.77} is taken to imply that total capital returns for owners scale directly with firm value, so the effective size-dependent multiplier in the geometric process is θ = 0.77. The owner wealth obeys the SDE dw/w = r(V(s)) dt + σ(V(s)) dW with reflecting barrier at low w; solving the corresponding Fokker-Planck equation yields a stationary power-law tail whose exponent is exactly α_w = 1/θ when the drift and diffusion coefficients inherit the V(s) scaling (standard result for multiplicative processes with power-law state dependence; no auxiliary parameters are required). The per-employee exponent then follows immediately as ζ = θ − 1 ≈ −0.23. Because the input 0.77 is taken unchanged from independent firm data, the result remains parameter-free. To address the referee’s concern we will (i) revise the abstract to state the SDE and the origin of α_w = 1/θ in one additional sentence and (ii) add a short paragraph or appendix deriving the tail exponent from the Fokker-Planck equation under the observed scaling. These changes clarify the derivation without altering any numerical claims or introducing new parameters. revision: yes

Circularity Check

0 steps flagged

No significant circularity; central result uses independent external data for firm scaling.

full rationale

The derivation proceeds from Gibrat's law to Zipf firm sizes, then applies maximum entropy to within-firm wages (standard constraint-based inference) and standard stochastic processes (additive noise with barrier for employees; multiplicative returns for owners) to obtain the two-class distributions. The headline reproduction of α_w ≈ 1.30 plugs the cited external observation V ~ s^{0.77} (Axtell 2001) into the model-derived relation α_w = 1/θ without fitting to income/wealth data. No load-bearing step reduces to a self-fit, self-citation chain, or definitional tautology; the cited firm-value exponent is independent of the target wealth observations. The model is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 4 axioms · 0 invented entities

The central claim rests on established domain assumptions from statistical mechanics and firm dynamics literature, with no new free parameters or invented entities; key numerical inputs are taken from independent empirical observations rather than fitted to the target distributions.

axioms (4)

domain assumption Gibrat's law for firm growth
Invoked to imply Zipf firm-size distribution as the starting point for mixture aggregation.
domain assumption Maximum entropy principle applied to within-firm wages
Combined with mixture aggregation across firms to produce Boltzmann-Gibbs income distribution.
domain assumption Additive-noise wealth dynamics with reflecting wall at zero for employees
Used to derive Boltzmann-Gibbs employee wealth distribution.
domain assumption Multiplicative capital returns for firm owners that scale with firm size
Produces Pareto wealth tail whose exponent is set by the observed firm-value scaling.

pith-pipeline@v0.9.0 · 5728 in / 1703 out tokens · 76974 ms · 2026-05-08T09:26:23.199383+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

15 extracted references

[1]

Nachtrieb,Equations of Motion for an Economy: Cap- ital Deepening, Technology, and Firm Survival(compan- ion paper, 2026)

R. Nachtrieb,Equations of Motion for an Economy: Cap- ital Deepening, Technology, and Firm Survival(compan- ion paper, 2026)

2026
[2]

V. M. Yakovenko and J. B. Rosser, Rev. Mod. Phys.81, 1703 (2009)

2009
[3]

R. L. Axtell, Science293, 1818 (2001)

2001
[4]

Dra˘ gulescu and V

A. Dra˘ gulescu and V. M. Yakovenko, Eur. Phys. J. B20, 585 (2001)

2001
[5]

Piketty,Capital in the Twenty-First Century(Har- vard University Press, 2014)

T. Piketty,Capital in the Twenty-First Century(Har- vard University Press, 2014)

2014
[6]

Aoki and M

M. Aoki and M. Nirei, Am. Econ. J.: Macroeconomics9, 36 (2017). 4 Firms Zipf size distribution Households Employees ⊕additive wages BG income BG wealth Owners ⊗multiplicative Pareto income Pareto wealth Government fiscal flows (invariant) wages consumption returns hh taxes firm taxes spending (a) 100 102 104 106 firm size s 10 4 10 2 100 Sf(s) (i) slope =...

2017
[7]

Gibrat,Les In´ egalit´ es ´Economiques(Sirey, Paris, 1931)

R. Gibrat,Les In´ egalit´ es ´Economiques(Sirey, Paris, 1931)

1931
[8]

BDSESIZE?g=010XX00US&codeset=naics~00

US Census Bureau, Business Dynamics Statistics (2026), https://data.census.gov/table/BDSTIMESERIES. BDSESIZE?g=010XX00US&codeset=naics~00

2026
[9]

E. T. Jaynes, Phys. Rev.106, 620 (1957)

1957
[10]

Brown and J

C. Brown and J. Medoff,The employer size-wage effect, J. Polit. Econ.97, 1027 (1989)

1989
[11]

Gabaix, Annu

X. Gabaix, Annu. Rev. Econ.1, 255 (2009)

2009
[12]

Benhabib, A

J. Benhabib, A. Bisin, and S. Zhu, Econometrica79, 123 (2011)

2011
[13]

R. Nachtrieb, Supplemental Material for this article, in- cluding derivations of the Itˆ o–Stratonovich Gibrat emer- gence, Fokker–Planck stability analysis, firm survival con- volution, Monte Carlo agent simulation, and complete data pipelines (2026)

2026
[14]

Gabaix, personal communication (April 2026)

X. Gabaix, personal communication (April 2026)

2026
[15]

Both are consistent with the BDS direct fit within the range spanned by the empirical sector-levelαvalues (0.74–1.46, bracketing Zipf); see [13]

The analytic power-law fit to thet −1/2logtcurve over 1– 30 yr givesb app ≈0.32; Monte Carlo simulation atα= 1 (Zipf) givesb sim ≈0.335. Both are consistent with the BDS direct fit within the range spanned by the empirical sector-levelαvalues (0.74–1.46, bracketing Zipf); see [13]