Long-run general-equilibrium effects of microfinance
LLM status Partially solved unverified
G21 — Banks - Depository Institutions - Micro Finance Institutions - Mortgages Economics catalogue openaigpt 6 astra ultra 45% cataloguednot recordedPartially solvedStatement
Original atlas ID: D1 (entry 31). Classification: Editorial saturation, long-run treatment, and equilibrium problem.
Original atlas question: Does expanded credit transform productivity and occupational structure rather than merely smooth household finance?
Definitions and assumptions
Fix markets $m$, nonempty finite baseline-household sets $\mathcal I_m$, households $i\in\mathcal I_m$, and horizon $H$. Treatment is an offer of a specified credit contract, including interest, liability, repayment, and access rules, rather than borrowing itself. Let $\mathbf D_m$ be the market's vector of offers. Potential outcomes $Y_{imt}(\mathbf D_m)$ may depend on other households' offers through prices, employment, and competition. Assume partial interference: arbitrary effects within a market are permitted, but no effects cross market boundaries; alternatively replace this assumption with a specified cross-market model. For saturation $s\in[0,1]$, let $\pi_s$ be an announced randomized offer-allocation rule satisfying $\mathbb E_{\pi_s}[|\mathcal I_m|^{-1}\sum_i D_{im}]=s$ in every market. Thus saturation is the expected offer share, not the borrowing share. Define outcomes separately for real consumption, business value added, enterprise entry, and occupation; assume finite expectations and consistent measurement.Formal research question
For each outcome $Y$, define the market policy response \[ \mu_{Y,t}(s)=\mathbb E_m\mathbb E_{\mathbf D_m\sim\pi_s} \left[\frac1{|\mathcal I_m|}\sum_{i\in\mathcal I_m} Y_{imt}(\mathbf D_m)\right], \qquad \Delta_{Y,t}(s)=\mu_{Y,t}(s)-\mu_{Y,t}(0). \] Identify or sharply bound $\Delta_{Y,t}(s)$ for relevant saturations and $t\leq H$, especially when an entire market receives access. Determine whether output and occupational changes persist after a stated subsidy or introductory contract expires. For national expansion, construct a model of lender entry, default, capital funding, wages, and product demand, and evaluate its transition separately from these experimental market contrasts.Interpretation and scope
Individual take-up effects differ from offer effects. An encouragement instrument identifies borrower effects only with an exclusion restriction that may fail if the offer changes expectations or competing lenders' behavior. A low-saturation experiment need not identify $s=1$: positive design support or an explicit extrapolation model is required. Occupational transformation must be defined, for example as persistent changes in enterprise value added and employment shares, rather than inferred from a single positive borrowing coefficient. Attrition, enterprise closure, and household migration are outcomes to be tracked for the baseline cohort. Consumption smoothing may improve welfare without raising measured productivity, and gains for borrowers can coexist with losses for competing businesses. Welfare analysis must include repayment, lender resource costs, risk, and distributional weights.Source audit and status
[S1] is the verified Hyderabad evaluation, [S2] a six-study synthesis, and [S3] a Bayesian hierarchical analysis. These already establish context-specific and pooled evidence. They do not identify every saturation, national equilibrium, or decade-long contract design. The long-horizon scale-up estimand is editorial, not a sourced conjecture or a certified universal 2026 open problem.References
- },leftmargin=*]
- Abhijit Banerjee, Esther Duflo, Rachel Glennerster, and Cynthia Kinnan (2015). The Miracle of Microfinance? Evidence from a Randomized Evaluation. American Economic Journal: Applied Economics 7(1), 22--53. https://doi.org/10.1257/app.20130533. Locator: Primary publisher, working-paper, or author record: bibliographic information and abstract; full-text theorem verification is not claimed. Role: Hyderabad randomized microcredit access; thematic reconstruction of the abbreviated citation.
- Abhijit Banerjee, Dean Karlan, and Jonathan Zinman (2015). Six Randomized Evaluations of Microcredit: Introduction and Further Steps. American Economic Journal: Applied Economics 7(1), 1--21. https://doi.org/10.1257/app.20140287. Locator: Primary publisher, working-paper, or author record: bibliographic information and abstract; full-text theorem verification is not claimed. Role: Synthesis of six studies; explains why program and context matter.
- Rachael Meager (2019). Understanding the Average Impact of Microcredit Expansions: A Bayesian Hierarchical Analysis of Seven Randomized Experiments. American Economic Journal: Applied Economics 11(1). https://doi.org/10.1257/app.20170299. Locator: Primary publisher, working-paper, or author record: bibliographic information and abstract; full-text theorem verification is not claimed. Role: Hierarchical analysis of treatment-effect heterogeneity; not a scale-up identification theorem.
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
GPT 6 Astra Ultra attempt on Econ-551
Credit offer laws, market entry saturation, and persistent effects
Status: Partial; the full catalogue target remains unresolved.Abstract
Identical expected credit coverage can produce different outcomes under interference, so saturation must specify an assignment law. A solved entry economy shows more enterprises and output alongside lower aggregate producer profits. A separate persistence model derives long-horizon effects after temporary credit access ends. Market-level randomization identifies these implemented policies, while a national lending equilibrium remains outside the result.Coverage is not an intervention by itself
Fix a baseline market population and let $\mathbf D$ be its vector of credit offers. Each offer specifies interest, repayment, liability, and access rules. Potential outcomes depend on the entire vector. Saturation $s$ is an expected offer share under a declared law $\pi_s$, not actual borrowing or an unconditional number describing every possible experiment.For two households, suppose both potential outcomes equal $D_1D_2$, perhaps because their projects are complementary. Three assignment laws all have expected offer share $1/2$. Independent half-probability offers give expected outcome $1/4$. Offering to both or neither with equal probability gives expected outcome $1/2$. Offering to exactly one gives outcome zero. All calculations follow directly from the probability of $\mathbf D=(1,1)$.
Thus $\mu(s)$ is undefined without $\pi_s$ when arbitrary interference is admitted. At the endpoints this ambiguity disappears: an expected offer share of one forces every household to receive an offer almost surely, and a share of zero forces no offers. At intermediate values, treatment correlation is part of the policy.
An exact market entry economy
Now consider a continuum economy with initial entrepreneur mass $n_0\ge0$ and a mass of additionally credit-eligible potential entrants. Each enterprise supplies one output unit. Inverse demand is $P(n)=A-bn$, with $b>0$, where $n$ is total enterprise mass. Each project uses real resources and foregone labor worth $w$, and assume \[ n_0<n^*=(A-w)/b\le1. \] Credit access enables a mass $s\in[0,1-n_0]$ of potential entrants to start, and they enter whenever operating profit is positive. Existing enterprises stay while profit is nonnegative; fix a stay/entry tie rule supporting the equilibrium at zero profit.Output, profits, and welfare need not have the same sign
Aggregate producer operating profit in this model is \[ \Pi(n)=n(A-w-bn). \] Its derivative is $A-w-2bn$, which becomes negative before entry stops at $n^*$. With $A-w=1$, $b=1$, and $n_0=0.6$, expanding access enough for $0.2$ new entrants raises output from $0.6$ to $0.8$ but lowers producer profit from $0.24$ to $0.16$.Each new enterprise earns $0.2$, so their total gain is $0.04$. Existing enterprises' per-firm profit falls from $0.4$ to $0.2$, a loss of $0.12$ across their mass. The net producer-profit change is therefore $-0.08$, agreeing with the direct calculation. Tracking borrowers alone would miss the larger incumbent loss.
If the inverse demand curve represents monetary consumer value, consumer surplus is $bn^2/2$ and total resource surplus is \[ W(n)=(A-w)n-bn^2/2. \] It rises from $0.42$ to $0.48$ in the same example. Higher output, lower producer profits, and higher total welfare coexist. Real intermediation and program costs would be subtracted separately. Loan principal and repayments are transfers within a boundary containing both lender and borrower, while screening effort, funding opportunity costs, and default resource losses are real costs.
This distinction matters for the catalogue's separate outcomes: consumption smoothing, business value added, employment, and welfare cannot be inferred from one another by renaming a borrowing coefficient.
A restricted persistence calculation
For a separate dynamic benchmark, suppose initial credit access raises by $\Delta\in[0,1]$ the probability that a household establishes a durable productive asset at date one. Once created, the asset survives each period independently with probability $r\in[0,1]$. After the introductory contract ends there is no further treatment effect on transitions, and no replacement creation in this benchmark. A surviving asset generates earnings increment $y>0$.The effect on productive-asset probability at horizon $h\ge1$ is $\Delta r^{h-1}$. This follows by multiplying initial probability by survival through $h-1$ transitions. For discount $\delta\in(0,1)$ and earnings paid from date one, the discounted earnings effect is \[ \Delta V=\sum_{h=1}^{\infty}\delta^h y\Delta r^{h-1} =\frac{\delta y\Delta}{1-\delta r}. \tag{2} \] The geometric series converges because $\delta r<1$.
An initial effect alone does not identify this value. For $\Delta=0.2$, $y=1$, and $\delta=0.9$, values are approximately $0.3273$ when $r=0.5$ and $1.2414$ when $r=0.95$. Both models have the same date-one effect $0.2$. Follow-up transition evidence or a justified persistence restriction is necessary. Temporary consumption insurance without productive-asset creation can improve utility yet have $\Delta=0$ in this specific productivity measure.
A design for implemented market policies
Suppose independent markets are randomly assigned one of finitely many pairs consisting of a contract version and an offer law $\pi_s$. Follow every baseline household, including enterprise closures and movers. Under no cross-market interference, consistency, and assignment independence, the expected observed market-average outcome in each arm equals \[ {\mathbb{E}}_m{\mathbb{E}}_{\mathbf D\sim\pi_s}\left[ |\mathcal I_m|^{-1}\sum_iY_{ih}(\mathbf D)\right]. \] This is exactly the catalogue response for that law and horizon. The equality follows by first conditioning on the full potential-outcome table, averaging over offers according to $\pi_s$, and then averaging over randomized market assignment.A whole-market all-offer arm directly identifies the full-access response relative to a no-offer arm. Intermediate-arm comparisons do not automatically identify that endpoint without positive support or an extrapolation model. If offers are independent Bernoulli within markets, endpoint assignment vectors technically have positive probability, but their rarity may make inference impractical. The distinction is precision versus absent support.
Randomized offers do not automatically instrument borrowing. An offer can change expectations, outside-lender terms, or relatives' behavior even if the recipient does not borrow, violating a receipt-effect exclusion. The policy intention-to-treat outcome remains well defined without that exclusion.
Remaining general-equilibrium problem
The examples establish three precise points: assignment correlation matters, market competition limits occupational expansion, and persistence determines long-run value. They do not solve a national expansion with endogenous lender entry, capital funding, default, labor supply, and cross-market product demand. Parameters governing those responses are not learned from a single low-saturation trial by assumption.The broad catalogue target remains unresolved. The restricted models are illustrative equilibrium and transition calculations, not estimates or universal conclusions about microfinance effectiveness.
Completion Estimate
45%This is a post hoc, heuristic estimate for the full catalogue target. The attempt proves assignment-law dependence under interference, solves an entry-market benchmark, and derives post-credit asset-persistence value and market-policy identification. Full microfinance transformation still requires supported long-horizon saturation responses, tracked closures and movers, and a national transition with endogenous lending, default, funding, wages, and demand.
References
- [micro] A. Banerjee, E. Duflo, R. Glennerster, and C. Kinnan (2015). The Miracle of Microfinance? Evidence from a Randomized Evaluation. https://www.aeaweb.org/articles?id=10.1257/app.20130533. The primary publication supplies the randomized-access context; no numerical empirical findings are imported into the models above.