论文检索

基于最优传输的贝叶斯后验采样生成模型AI 解读

Abstract We investigate the problem of sampling from posterior distributions with intractable normalizing constants in Bayesian inference. Building on transport map-based generative models for posterior sampling, we propose an optimal transport (OT)-constrained transport map class that learns a deterministic map from a reference distribution to the target posterior through constrained optimization. The proposed class exploits structural properties of OT maps and allows…

重新审视石油供给新闻冲击：代理变量与非高斯结构向量自回归AI 解读

ABSTRACT We replicate a study by Känzig ( American Economic Review , 111 (2021), 1092–1125), who employs structural vector autoregressive techniques to examine the impact of changes in oil supply expectations on the price of oil and other macroeconomic aggregates. Känzig identifies an oil supply news shock by constructing a proxy from OPEC announcements about their production plans. As an alternative approach, we use non‐Gaussianity and independence to identify structural…

大冲击宏观模型的快速高效估计AI 解读

We propose a new approach for efficiently estimating and analyzing macroeconomic models subject to large shocks. We apply the methodology to a two-sector model that captures the heterogeneous exposure of different sectors to the COVID-19 pandemic. We solve the model nonlinearly and propose a new nonlinear, non-Gaussian filter designed to handle large shocks and identify their source and time location. Monte Carlo experiments show that the estimation and identification of…

对 Suri (2011)《技术采纳中的选择与比较优势》的评论AI 解读

This paper illustrates and addresses weak identification in the correlated random coefficient (CRC) model that Suri (2011) uses to study agricultural technology adoption. Using the publicly available dataset, which differs slightly from the version used in Suri (2011), we obtain different point estimates for the paper’s main CRC model. To understand why these differences are as large as they are, we recast the CRC model as a more general random coefficient model in which the…

协变量漂移下外部有效性的简单稳健性度量AI 解读

This paper studies the robustness of estimated policy effects to changes in the distribution of covariates, a key determinant of the external validity of (quasi)experimental results. I propose a novel robustness metric δ ∗ which measures the smallest covariate shift needed to invalidate an empirical claim about the policy effect (e.g., AT E > 0). I estimate δ ∗ via de-biased GMM, achieving a parametric rate of convergence while accommodating machine-learning estimators of…

一个精确 t 检验AI 解读

Multivariate linear regression and randomization-based inference are two essential methods in statistics and econometrics. Nevertheless, the problem of producing a randomized test for the value of a single regression coefficient that is exactly valid when errors are exchangeable, and which is asymptotically valid for the best linear predictor, has remained elusive. In this paper, we produce a test that is exactly valid with exchangeable errors and which allows for general…

从预测算法到异常的自动生成AI 解读

How can we extract theoretical insights from machine learning algorithms? We take a familiar lesson: researchers often turn intuition into theoretical insight by constructing “anomalies” — specific examples highlighting hypothesized flaws in a theory, such as the Allais paradox. We develop procedures that replace researchers’ intuitions with predictive algorithms: given a predictive algorithm and a theory, our procedures automatically generate anomalies for that theory. We…

异质性主体下的极值理论AI 解读

Extreme value processes feature in any economic model in which agents receive a number of draws from some distribution and we examine the behavior of the maximum in the limit as the number of draws becomes large. This paper asks: Do the average outcomes of such processes change when different agents receive a different number of draws? To answer this, we allow the number of draws an agent receives from the underlying distribution (e.g. of productivities, ideas, or utility…

集合值预测与模型误设下的信息推断AI 解读

This paper proposes an information-based inference method for partially identified parameters in incomplete models that is valid both when the model is correctly specified and when it is misspecified. Key features of the method are: (i) it is based on minimizing a suitably defined Kullback-Leibler information criterion that accounts for incompleteness of the model and delivers a non-empty pseudotrue set; (ii) it is computationally tractable; (iii) its implementation is the…

最富者为何如此富有？新的纵向经验证据及其对财富不平等理论的启示AI 解读

We study the lifecycle dynamics of wealth inequality using 1993-2019 Norwegian administrative panel data on wealth and income. Employing a novel budget-constraint approach, we decompose the excess wealth of the top 0.1% households relative to the median between ages 45 and 64 into higher saving rates (36%), inheritances (31%), returns (28%), and labor income (5%). One-quarter of the wealthiest—the “New Money”—start with negative wealth on average but accumulate rapidly…

含要素增强型技术的生产函数估计：加成率的应用AI 解读

I develop a new method for estimating production functions with laboraugmenting technology and apply it to markup estimation. The method imposes limited parametric restrictions and generalizes prior approaches that rely on the CES production function. I first extend the canonical Olley-Pakes framework assumptions to multidimensional productivity and then develop an identification strategy based on a control variable approach and first-order conditions. I use this method to…

最优公式工具变量AI 解读

When estimating the effects of treatments defined by complex formulas, researchers often use simple functions of exogenous shocks as instruments. A leading example is “simulated instruments” for public policy eligibility, which capture variation in state-level policy generosity. We show how more powerful instruments can be constructed by incorporating heterogeneous shock exposure while using a recentering procedure to avoid bias. We characterize the asymptotically efficient…

固定缴费养老金计划的设计AI 解读

Defined contribution plans are a major vehicle for retirement savings in the US, holding almost $10 trillion in assets under management. In recent years, the quality and availability of these plans has been the subject of active policy attention and of several major lawsuits. This paper studies how employers and plan providers design these plans. We argue that low plan quality and limited provision can come from two sources. First, employer willingness to pay may be…

农业政策干预的规模化效应AI 解读

Policies to raise agricultural productivity have been central in the fight against global poverty. Their impacts are often measured in experiments that provide strong causal identification but may be too small-scale to capture general equilibrium effects that arise when policies are scaled up. We develop a quantitative model of agricultural trade, featuring a granular economic geography, that combines parameters estimated from small-scale experiments with rich administrative…

警力资源的有效部署：来自印度随机酒驾严打的理论与新证据AI 解读

Should police activity be narrowly focused and high force, or widely dispersed but of moderate intensity? Critics of intense “hotspot” policing argue it primarily displaces, not reduces, crime. But if learning about enforcement takes time, the police may take advantage of that period to intervene intensively in the most productive location. We propose a multi-armed bandit model of criminal learning and structurally estimate its parameters using data from a randomized…

有序随机效用模型中参数假设的可检验含义AI 解读

We establish the testable implications of ordered random utility models under nonparametric, semi-nonparametric, and parametric assumptions. After characterizing these models in a continuous setting and in the absence of fundamental parametric restrictions, we show that assumptions about the type distribution alone are immaterial, while assumptions about the map linking types to utilities are relevant only insofar as they restrict the class of utilities at stake.…

基于条件异方差识别日度油价冲击及其不确定性AI 解读

This paper develops a daily structural VAR for identifying oil price shocks from conditional heteroskedasticity and quantifies the transmission of their timevarying uncertainty. Identification exploits volatility clustering and spillovers in a structural GARCH specification and the complete identified set is derived. We derive covariance impulse response functions and introduce a time-varying forecast error variance decomposition (TV-FEVD) that allows for a decomposition of…

分解加成率的上升AI 解读

A recent literature argues that the rise in aggregate markups is due to economic activity reallocating toward high-markup firms rather than an increase in firmlevel markups. I show that standard decomposition methods used to analyze this phenomenon can estimate a non-zero reallocation effect even when the joint distribution of markups and market shares is time-invariant. This is due to churning in the firm-level distribution of markups and market shares. Empirically, I find…

紧识别SVAR中基于“软”符号限制的快速后验抽样AI 解读

We propose algorithms for conducting Bayesian inference in structural vector autoregressions identified using sign restrictions. The key feature of our approach is a sampling step based on ‘soft’ sign restrictions. This step draws from a target density that smoothly penalises parameter values that violate the restrictions, facilitating the use of computationally efficient Markov chain Monte Carlo sampling algorithms. An importance-sampling step yields draws conditional on…

利用成交价格进行有限样本拍卖推断AI 解读

We first provide inference on the quantile function of the bid distribution in firstprice or descending auctions, and the quantile function of the value distribution in second-price or ascending auctions, with a known number of bidders with symmetric independent private values. Specifically, we provide finite-sample, nonparametric, tuning parameter-free, uniform confidence bands for these quantile functions when observing only the transaction price. We also show how these…

有限状态马尔可夫链近似：一种隐马尔可夫方法AI 解读

This paper proposes a novel finite-state Markov chain approximation method for Markov processes with continuous support, providing both an optimal grid and a transition probability matrix. The method can be used for multivariate processes and processes with time-varying components. It is based on minimizing the information loss between a Hidden Markov Model and the true data-generating process. We provide sufficient conditions under which this information loss can be made…

生育、子女性别与父母迁移决策：来自中国的证据AI 解读

We study how the number and gender of children affect rural-tourban parental migration in China. Exploiting variation from the “1.5- child policy,” which allowed a second birth only if the firstborn was a girl, we separately identify the causal effects of fertility and child gender based on a novel semi-parametric method. Using the 2010 population census, we find that an additional child increases the likelihood of joint parental migration by 12.1 percent, while the presence…

用统计决策理论全面评估预测算法的样本外表现AI 解读

We argue that comprehensive out-of-sample (OOS) evaluation using statistical decision theory (SDT) should replace the current practice of K-fold and Common Task Framework validation in machine learning (ML) research on prediction. SDT provides a formal frequentist framework for performing comprehensive OOS evaluation across all possible (1) training samples, (2) populations that may generate training data, and (3) populations of prediction interest. Regarding feature (3), we…

即期相关过程日内变动的非参数检验AI 解读

The association between log-price increments of exchange-traded equities, as measured by their spot correlation estimated from high-frequency data, exhibits a pronounced upwardsloping and almost piecewise linear relationship at the intraday horizon. There is notably lower—on average less positive—correlation in the morning than in the afternoon. We develop a nonparametric testing procedure to detect such variation in a correlation process. The test statistic has a known…