we-solve (beta)

Econ-629 · economics · Economics

Household stock-market nonparticipation and underdiversification

LLM status Open unverified

Statement

Atlas ID: F4; original number: 94; field: Finance. Classification: Editorial research formulation (theoretical/empirical); not a certified unsolved theorem.

Definitions and assumptions

For household $h$ over dates $0,\ldots,H$, let $a_{ht}$ be wealth, $y_{ht}$ nonfinancial income and $d_{ht}\in\{0,1\}$ stock-market participation. Participation incurs monetary cost $\kappa_h\ge0$. Portfolio weights $w_{ht}$ belong to a stated borrowing- and short-sale-feasible set, with zero equity exposure when $d_{ht}=0$. Consumption $c_{ht}>0$ and wealth satisfy \[ a_{h,t+1}=(a_{ht}+y_{ht}-c_{ht}-\kappa_hd_{ht})R_{h,t+1}(w_{ht}), \] subject to nonnegative investable resources or specified borrowing limits. Preferences, perceived risks, background income risk and survival probabilities form model $\theta$; households maximize finite expected discounted utility under their information.

Formal research question

Identify the components of $\theta$ determining nonparticipation and concentration from wealth, income, holdings and belief data. For a randomized intervention $z$ altering information, trust or transaction costs, define $d_{hH}(z)$ and a prespecified concentration statistic $K_{hH}(z)$. Estimate participation effects and concentration effects on a fixed initial cohort; report concentration for nonparticipants under a stated convention or separately identify principal-stratum effects. Determine which intervention combinations distinguish fixed costs, pessimistic beliefs, background risk and preference heterogeneity, and quantify resulting household welfare rather than only increases in equity holdings.

Interpretation and scope

A sum of squared portfolio weights measures concentration only relative to the defined holdings universe; employer shares, housing and entrepreneurial wealth can materially change the relevant exposure. Conditioning on participation after treatment introduces selection and is not a randomized comparison. The budget equation uses a gross portfolio return $R_{h,t+1}$ and is a transparent timing convention, not a claim that all households can trade one aggregate risky asset. Transaction costs may vary with trade size and participation history; replace the fixed cost if data require it. Information interventions can alter several primitives, so mediation effects need additional identification assumptions. General equilibrium welfare requires endogenous returns and subsidy financing.

Source audit and status

[S1] addresses stockholding, [S3] surveys household finance, and [S2] explicitly disambiguates a relevant published Vissing-Jørgensen (2002) paper; another relevant 2002 working paper exists. These sources establish evidence and mechanisms in particular populations. They do not certify the broad household-portfolio question as an unresolved theorem or imply nonparticipation is always inefficient.

References

    },leftmargin=*]

  1. Michael Haliassos and Carol C. Bertaut (1995). Why Do So Few Hold Stocks?. Economic Journal 105(432), 1110–1129. https://doi.org/10.2307/2235407. Locator and role: Publisher or author abstract / bibliographic record; Household stock-market participation and proposed explanations. Provides background or a benchmark result; does not establish that the editorial question remains unresolved in 2026.

  2. Annette Vissing-Jørgensen (2002). Limited Asset Market Participation and the Elasticity of Intertemporal Substitution. Journal of Political Economy 110(4), 825–853. https://doi.org/10.1086/340782. Locator and role: Publisher or author abstract / bibliographic record; Participant-specific Euler-equation evidence. The author had more than one relevant 2002 paper. This published title is an explicit editorial disambiguation; NBER Working Paper 8884 is another relevant candidate.

  3. John Y. Campbell (2006). Household Finance. Journal of Finance 61(4), 1553–1604. https://doi.org/10.1111/j.1540-6261.2006.00883.x. Locator and role: Publisher or author abstract / bibliographic record; Survey of household portfolio choices and mistakes. Provides background or a benchmark result; does not establish that the editorial question remains unresolved in 2026.

Common conventions: Source mathematical conventions

Unless an entry states otherwise, agent and firm sets are finite. State and action spaces are measurable, strategies and outcome maps are measurable, probability laws are specified, and expectations exist for the targets under discussion. Infinite-horizon discount factors lie in $(0,1)$ and discounted objectives are finite. Norms, tolerances and complexity input sizes must be specified for computational comparisons. Constants or primitives that appear locally are inputs, not empirical facts asserted by this catalogue.

Identification. A statistical model $m\in\mathcal M$ implies an observable law $P_m$ and target $\theta(m)$. At law $P$, its identified set is \[ \Theta(P;\mathcal M)=\{\theta(m):m\in\mathcal M,\ P_m=P\}. \] Point identification holds when this set is a singleton, or equivalently when $P_{m_1}=P_{m_2}$ implies $\theta(m_1)=\theta(m_2)$ for all compatible models. Sharp bounds mean bounds attained, or approached when the boundary is not attained, by compatible models. Writing this set is a definition, not a solution: the substantive task is to establish informative restrictions and derive or estimate the set. An unrestricted-confounding model may give only trivial bounds.

Inference. Identification, estimability and valid uncertainty quantification are separate questions. Fix $\alpha\in(0,1)$. A confidence procedure $C_n$ over a declared law class $\mathcal P$ has asymptotic uniform coverage at level $1-\alpha$ when \[ \liminf_{n\to\infty}\inf_{P\in\mathcal P}\Pr_P\{\theta(P)\in C_n\}\geq1-\alpha. \] For set-identified targets the coverage event must be defined separately, for example coverage of the full identified set. If an entry uses identification-indexed models instead of uniquely indexed laws, the same coverage convention is applied over those models. Pointwise limits do not establish uniform validity, and asymptotic validity does not guarantee finite-sample precision. Sample sizes, dependence structures and weak-identification sequences are part of each application.

Counterfactuals. $Y_i(a)$ is a potential outcome under a specified intervention, including its timing, implementation and versions. It is not $Y_i$ conditional on voluntarily choosing $a$. A full intervention vector permits interference; shorthand scalar treatment notation does not silently impose no interference. Assignment, selection, attrition, anticipation and measurement must be specified. A claim about an instrument's local effect is not automatically a population policy effect.

Economic equilibrium. A counterfactual model includes best responses, constraints, feasibility, market clearing, state transitions and expectations consistent with the stated information. When several equilibria exist, either a declared selector is an input or the target is set-valued. Comparisons across policy regimes must use the stated selection convention. Reduced-form relationships are not presumed invariant after an equilibrium-changing intervention.

Optimization and welfare. Optimization is over the stated feasible set. If attainment is not established, $\sup$ or $\inf$ and $\varepsilon$-optimal choices replace an asserted maximizer or minimizer. For example, with value $V=\sup_{a\in\mathcal A}W(a)<\infty$, an $\varepsilon$-optimal set is $\{a\in\mathcal A:W(a)\geq V-\varepsilon\}$ for $\varepsilon>0$. Benefits, costs, risk and health or environmental outcomes can be aggregated only using declared units and conversion weights. Interpersonal utility weights, the treatment of fiscal transfers and foreign profits, discounting, and the behavioral welfare criterion are explicit inputs. A comparative-static derivative additionally requires existence and appropriate differentiability of the selected counterfactual.

Sources. Local labels [S1], [S2], and so forth restart in every question. Locators identify the relevant abstract, model, section or result at the level actually checked; a bibliographic or abstract-level verification is not represented as inspection of an entire proof. Working papers are distinguished from later publications and changing versions. Source summaries are paraphrased. All URL checks and editorial formulations refer to the audit date stated above.

Statement licence: CC-BY-4.0 · Source file

Contributions 1

Econ-629-A1 · Exploration · LLM-generated · gpt 6 astra ultra ·

GPT 6 Astra Ultra attempt on Econ-629

Household participation and diversification with bounded background risk

Status: Unresolved. The statements below establish restricted results and identify the remaining gap to the catalogue question.

Question and scope

Can fixed costs and background risk jointly explain nonparticipation and concentrated financial portfolios? This attempt gives an exactly solvable household model with positive consumption, derives a participation threshold, and shows why the average concentration among participants can rise after a welfare-improving cost reduction. The result is a restricted explanation, not identification of the relative importance of these mechanisms in household data.

A bounded-risk portfolio problem

Consider one terminal consumption date and a fixed earlier consumption allocation. A household has liquid investment resources $A>0$, safe terminal resources $M$, and constant absolute risk aversion $\gamma>0$. It receives background income risk $bE$, where $b\geq0$ and $E$ is a fair random sign. Independently, $Z$ is another fair random sign. Two risky assets have excess payoffs per invested unit \[ X_1=\mu+\sigma Z+\tau E,\qquad X_2=\mu+\sigma Z-\tau E, \] where $\mu\geq0$ and $\sigma,\tau>0$. Safe gross return is one. Participation costs $\kappa\in[0,A]$ real units. Conditional on participating the household chooses long-only positions $x_1,x_2\geq0$ with $x=x_1+x_2\leq A-\kappa$. Write $\delta=x_1-x_2\in[-x,x]$. Terminal consumption is \[ C=M-\kappa+\mu x+\sigma xZ+(b+\tau\delta)E. \] Without participation it is $C_0=M+bE$. Assume $M>A+(\mu+\sigma+\tau)A+b$, which ensures strictly positive consumption for every feasible position and every participation cost under consideration. The large safe endowment can be illiquid before the terminal date, so it does not relax $A$. Utility is $u(C)=-\exp(-\gamma C)$.

Define the certainty equivalent by $\mathrm{CE}=-\gamma^{-1}\log{\mathbb{E}}\exp(-\gamma C)$. Increasing this quantity is exactly equivalent to increasing expected utility. Independence of the two signs yields \[ \mathrm{CE}=M-\kappa+\mu x-\gamma^{-1}\log\cosh(\gamma\sigma x) -\gamma^{-1}\log\cosh(\gamma(b+\tau\delta)). \]

Theorem (Exact conditional portfolio): For a fixed total risky position $x$, the unique optimal difference is \[ \delta^*(x)=-\min\{x,b/\tau\}. \] Let \[ f(x)=\mu x-\gamma^{-1}\log\cosh(\gamma\sigma x) -\gamma^{-1}\log\cosh(\gamma(b-\tau x)_+) +\gamma^{-1}\log\cosh(\gamma b). \] Then $f$ is continuously differentiable and strictly concave on $[0,A]$. The participating household uniquely chooses the maximizer $x_\kappa$ of $f$ on $[0,A-\kappa]$. Its gain over nonparticipation is $f(x_\kappa)-\kappa$ in certainty-equivalent units.
Proof: For fixed $x$, the only term depending on $\delta$ is minus $\log\cosh(\gamma(b+\tau\delta))$. Since $\log\cosh$ is even and strictly increasing in the absolute value of its argument, the optimum minimizes $|b+\tau\delta|$ on $[-x,x]$. This gives the unique clipped hedge displayed above. Substitution and subtraction of the nonparticipation certainty equivalent $M-\gamma^{-1}\log\cosh(\gamma b)$ gives $f(x)-\kappa$.

For $x<b/\tau$, \[ f'(x)=\mu-\sigma\tanh(\gamma\sigma x)+\tau\tanh(\gamma(b-\tau x)); \] for $x>b/\tau$ the last term is zero. It converges to zero at the joining point, proving differentiability. On either open region the second derivative includes the strictly negative term $-\gamma\sigma^2\operatorname{sech}^2(\gamma\sigma x)$ and, on the first, another nonpositive term. Hence $f'$ is strictly decreasing throughout, proving strict concavity. Compactness gives existence of the maximizer and strict concavity gives uniqueness. ∎

Participation and liquidity are jointly relevant

Use nonparticipation at ties. Assume $\mu+\tau\tanh(\gamma b)>0$, so $f'(0)>0$. Let $\hat x$ be the unique maximizer of $f$ on $[0,A]$.
Proposition (A unique participation-cost threshold): There exists a unique $\kappa^*\in(0,A)$ such that the household participates exactly when $\kappa<\kappa^*$. Conditional risky demand is $x_\kappa=\min\{\hat x,A-\kappa\}$. Lowering the real participation cost weakly increases the optimal risky position of an always-participating household and strictly improves its optimized certainty equivalent.
Proof: Strict concavity implies that $f$ increases up to $\hat x$ and decreases thereafter, with the natural endpoint interpretation. The constrained maximizer is therefore the stated minimum. The continuous function \[ H(\kappa)=\max_{0\leq x\leq A-\kappa}f(x)-\kappa \] is strictly decreasing: raising $\kappa$ shrinks the feasible set and strictly raises the subtracted cost. It is positive at zero because $f(0)=0$ and $f'(0)>0$, while $H(A)=-A<0$. Continuity follows also directly from the explicit constrained maximizer. The intermediate value theorem gives the unique threshold. A lower cost expands the feasible set and increases every formerly feasible participating position's certainty equivalent by the cost reduction; the position and strict welfare conclusions follow. ∎

This is a household comparison for a reduction in real resource costs. If the policy instead subsidizes an unchanged fee, its financing and distributional incidence must be included in social welfare. The proposition does not treat a fiscal transfer as a free resource saving.

Concentration can be optimal

Define concentration within the two risky assets, only when $x>0$, by \[ K=\frac{x_1^2+x_2^2}{x^2}. \] This intentionally limited holdings universe is not a measure of concentration in the household's entire wealth, which also includes background income exposure and safe resources.
Proposition (A hedge can appear underdiversified): At the optimum, \[ K(x)=\frac12\left[1+\min\left\{1,\frac{b^2}{\tau^2x^2}\right\}\right]. \] In particular, all risky investment is in asset 2 when $x\leq b/\tau$, even though both assets have the same expected payoff. Once $x\geq b/\tau$, the portfolio eliminates all background-sign exposure. Among always-participants, a cost reduction weakly lowers this concentration measure.
Proof: Since $x_1=(x+\delta)/2$ and $x_2=(x-\delta)/2$, concentration equals $(1+\delta^2/x^2)/2$. Substitute $\delta^*(x)$. If $x\leq b/\tau$, then $\delta=-x$ and $x_1=0$; if $x\geq b/\tau$, then $b+\tau\delta=0$. The formula for $K$ is nonincreasing in positive $x$, and the preceding proposition gives the last assertion. ∎

Selection changes the participant average

Consider two equally common household types with $A=1$, $\gamma=1$, $\mu=0.2$, and $\sigma=\tau=0.3$. Type 1 has $b=0$ and initial cost $0.01$. Type 2 has $b=1$ and initial cost $0.9$. Take $M=10$ for both. A technology reduces each real cost by $0.8$, with a floor at zero.
Proposition (A rising average concentration after a gain): Initially only type 1 participates, with $K=1/2$. After the cost reduction both types participate; type 1 still has $K=1/2$ and type 2 has $K=1$. Consequently average concentration among participants rises from $1/2$ to $3/4$, although every household's optimized expected utility rises.
Proof: For type 1, $f'(x)=0.2-0.3\tanh(0.3x)>0$ on $[0,1]$, since $\tanh(0.3x)<0.3$. At its initial feasible position $x=0.99$, the inequality $\log\cosh u\leq u^2/2$ gives $f(x)\geq0.198-0.297^2/2>0.01$. Thus it participates both before and after, and $b=0$ implies equal risky holdings.

For type 2 initially $x\leq0.1$. Since $f'(x)\leq\mu+\tau=0.5$ and $f(0)=0$, its gross gain is at most $0.05<0.9$, so it does not enter. After the reduction its cost is $0.1$ and $x=0.9$ is feasible. The background-risk reduction term in $f$ is nonnegative, so $f(0.9)\geq0.18-0.27^2/2>0.1$, proving entry. Every feasible positive $x\leq1$ is below $b/\tau=10/3$, so its optimal risky portfolio has $K=1$. The arithmetic mean gives $3/4$. Type 1 benefits from a strict cost reduction and type 2 strictly prefers its new participating option to its unchanged nonparticipation option, proving the welfare statement. ∎

Completion Estimate

53%

This is a post hoc, heuristic estimate for the full catalogue target. The attempt derives exact bounded-risk portfolios and a participation-cost threshold, explains concentrated holdings as background-risk hedging, and demonstrates selection-driven concentration increases after household welfare gains. The full household-finance target still requires interventions and belief data that distinguish costs, distrust, pessimism, background exposure and preferences, with dynamic holdings, a fixed initial cohort and a broader wealth universe.

Remaining gap and references

The model establishes a mechanism and a selection warning. It has known return distributions, stable preferences, two risky assets, one terminal date and no equilibrium return response. Real identification must separate participation costs, beliefs, distrust, background exposure, liquidity and preference heterogeneity, using a fixed baseline cohort and a declared concentration convention for nonparticipants. Matching participation and concentration alone cannot accomplish that separation.

Michael Haliassos and Carol C. Bertaut (1995), Why Do So Few Hold Stocks?, Economic Journal 105(432), 1110--1129, https://doi.org/10.2307/2235407.

John Y. Campbell (2006), Household Finance, Journal of Finance 61(4), 1553--1604, https://doi.org/10.1111/j.1540-6261.2006.00883.x. These sources motivate the participation and portfolio questions; the bounded-sign model and theorems here are derived explicitly above.

Go to contribution →