price elasticity estimation

This paper addresses the lack of robust design foundations for sensitivity analysis in finite-population causal inference. Methodologically, it introduces a novel sensitivity analysis framework grounded in the experimental design distribution—first integrating design-based distributions with partial identification theory to construct model-free, non-asymptotic confidence intervals for the average treatment effect (ATE). It further reinterprets the role of randomization in sensitivity analysis and provides a new design-driven rationale for covariate balance checks. Key contributions include: (1) model-free, finite-population inference under heterogeneous treatment effects; (2) robust ATE confidence intervals with clear identification-theoretic interpretation; and (3) empirical validation across three real-world applications, demonstrating reliability and practicality in small-sample and highly heterogeneous settings.

Quantifying how input uncertainty propagates to model outputs remains a fundamental challenge in computational modeling. Method: This study systematically reviews and empirically compares prominent global and local sensitivity analysis (SA) techniques—including Sobol’, FAST, Morris screening, and local derivative-based methods—implemented via standard software packages, supporting both probabilistic modeling and distribution-free settings. Contribution/Results: We propose a practical decision framework that guides method selection based on problem characteristics, analytical objectives, and resource constraints—rejecting the notion of a universally “optimal” SA method and thereby addressing a critical gap in methodological implementation guidance. A reusable, open-source toolkit is developed to enhance the reliability and interpretability of uncertainty attribution. The framework and tools have been validated across multiple engineering and policy modeling applications, demonstrating robustness and scalability in real-world contexts.

This paper addresses the challenge of setting price bounds in data-driven pricing, proposing a bi-objective optimization framework that jointly maximizes revenue and ensures price reliability. Methodologically, it introduces the first integration of Bootstrap confidence interval estimation with Nelder-Mead black-box optimization, guided by expected total revenue under K-fold cross-validation. Demand prediction uncertainty is quantified via Bootstrap resampling, and robust price upper and lower bounds are efficiently computed using the simplex method. Experimental results demonstrate that the method significantly narrows price intervals while sustaining high revenue—particularly for small product catalogs or low-noise settings. Moreover, its revenue stability and bound tightness improve consistently with increasing data volume. The approach establishes a new paradigm for interpretable, production-ready intelligent pricing.

This paper identifies a fundamental bias in estimating average treatment effects (ATE) in continuous-parameter A/B tests—particularly price experiments—arising from interference among market participants. In pricing contexts, conventional estimators of profit change expectations can exhibit sign reversal, leading firms to adopt profit-damaging pricing policies. To address this, we propose a lightweight debiasing method requiring only equal partitioning of experimental units. We are the first to systematically characterize the “sign reversal” phenomenon and prove its ubiquity in two-sided markets and multi-category commission pricing. Through structural modeling and differential analysis, we derive an explicit closed-form expression for the bias and theoretically demonstrate that the classical estimator can indeed flip sign. Empirical evaluations across diverse market settings confirm that our method consistently restores correct decision directionality, thereby ensuring reliable causal inference.

This study investigates how limited attention systematically distorts the identification and estimation of structural parameters governing elasticities of substitution. By developing a sparse utility-maximization model that integrates covariance-moment estimation with supply-side orthogonality conditions, the paper demonstrates for the first time that although price and expenditure-share data under limited attention are observationally equivalent to those generated by a fully rational model, the estimated elasticities correspond to lower bounds of the true elasticities—attenuated by attention weights. This bias stems from fundamental identification constraints and cannot be mitigated by finite-sample corrections or weak-instrument robust methods. In the absence of external information, estimators necessarily converge to the attenuated values rather than the true parameters, thereby establishing limited attention as a fundamental constraint on structural estimation.

This study addresses statistical inference challenges arising from weak identification, noise contamination, multiple constraints, and model misspecification by proposing a unified framework based on Lagrangian constrained optimization. The work innovatively introduces Individual Shadow Prices (ISPs) to quantify the information content of each constraint and designs platform rules to distinguish signal from noise. Incorporating a Stein-type risk criterion, the method employs a data-driven approach to select tolerance parameters and leverages Karush–Kuhn–Tucker (KKT) conditions to achieve debiased estimation. Theoretical analysis establishes the consistency and asymptotic normality of the resulting estimator. Numerical simulations and empirical application to the Solow growth model demonstrate that the proposed approach effectively captures model uncertainty and enhances inference accuracy.

This study addresses the challenge of accurately estimating demand functions for new products based on consumers’ willingness-to-pay (WTP). It proposes a novel paradigm that directly identifies demand functions from WTP data by constructing a general yet analytically tractable parametric demand model and designing a consistent estimation procedure. Theoretical analysis and Monte Carlo simulations demonstrate that the proposed method effectively recovers the true underlying demand function and exhibits robust performance across a variety of model specifications. Combining theoretical rigor with practical applicability, this approach offers a versatile tool for academic research, business decision-making, and policy evaluation.

This study addresses the sensitivity of conventional maximum likelihood estimation to outliers in modeling proportion data with boundary values, which often leads to biased inference. To overcome this limitation, the authors propose a robust inflated Beta regression estimator that exhibits strong robustness and favorable asymptotic properties. A Wald-type robust test is developed alongside the estimator, and a data-driven adaptive tuning algorithm is introduced to enhance performance. The proposed approach significantly improves robustness while preserving model simplicity and interpretability. Extensive simulations and empirical analyses demonstrate that the method substantially outperforms traditional maximum likelihood estimation in the presence of outliers, offering both theoretical rigor and practical utility.

This study addresses the challenge of demand estimation in differentiated product markets when the dimensionality of product characteristics exceeds the number of observations, a setting where the classical Berry–Levinsohn–Pakes (1995) model struggles. The authors extend the BLP framework by incorporating high-dimensional control variables and employing Neyman-orthogonal estimation combined with machine learning techniques—such as Lasso—to handle high-dimensional nuisance parameters. Under approximate sparsity conditions, the proposed approach ensures √T-asymptotic normality for key parameters of interest, such as price coefficients, even when nuisance parameters converge at slower rates. Monte Carlo simulations demonstrate that the method yields accurate and robust estimates of price effects in finite samples under high-dimensional settings, substantially broadening the applicability of the BLP model.