American Economic Association

A Bayesian Approach to Longitudinal Social Relations Model (C1, I2)

Li Tan

,

Arizona State University

Xingchen Xu

,

Arizona State University

View Abstract

Abstract

The Social Relations Model (SRM) serves as a comprehensive conceptual and methodological framework for analyzing voluntary or involuntary interpersonal relationships and interactions among individuals within groups. Widely employed in social sciences research, SRM has gained particular relevance in the economics of education, given the increasing focus on team learning and collaborative project-based learning.
Longitudinal team learning data is abundant in practice. For instance, the Comprehensive Assessment of Team Member Effectiveness (CATME) system has been utilized by over 1.4 million students and 17,000 instructors. Instructors frequently gather multiple rounds of round-robin peer evaluation and assessment data within student teams throughout a semester or even across several semesters. Although previous research has applied the standard SRM to longitudinal social relations data, the model was originally developed based on cross-sectional data. Consequently, its application to longitudinal data, where the same individuals are observed across multiple periods, has demonstrated critical limitations.
To address the scarcity of methods for analyzing the Longitudinal Social Relations Model (LSRM), we present a Bayesian approach, aimed at developing a general and flexible framework to bridge this gap. Utilizing a simulation-based framework, we first demonstrate that the Bayesian method significantly outperforms two alternative approaches—a two-step method and the Social Relations Structural Equation Model—in terms of estimation accuracy. Furthermore, the Bayesian method is applied on an empirical application to illustrate its additional benefits: it easily incorporates covariates, such as demographic variables, into the Bayesian LSRM, and effectively handles missing data, a common issue in round-robin peer-evaluation datasets.
By introducing the Bayesian LSRM, we pave the way for analyzing a range of impactful and policy-relevant questions in social sciences research. This advanced method enhances the understanding and interpretation of longitudinal interpersonal relationships and interactions within groups, contributing to more effective policy-making and interventions in educational and other social contexts.

A Behavioral Heterogeneous Agent New Keynesian Model (E5, E3)

Fabian Seyrich

,

Berlin School of Economics

Oliver Pfäuti

,

University of Texas-Austin

Abstract

Recent empirical evidence led to a rethinking of the transmission mechanisms of monetary policy: (i) monetary policy influences consumption largely through indirect effects; (ii) households with higher marginal propensities to consume are more exposed to monetary policy; (iii) forward guidance has relatively weak effects on current economic activity; and (iv) the lower bound on nominal interest rates does not cause significant instability. To account for these facts simultaneously, we develop a New Keynesian model with cognitive discounting and household heterogeneity. In contrast to existing models, we jointly account for these facts without having to rely on a specific monetary or fiscal policy. The interaction of household heterogeneity and cognitive discounting leads to a novel amplification channel of inflationary supply shocks. To stabilize inflation, monetary policy needs to raise interest rates more substantially than under rational expectations, which leads to a stronger increase in consumption inequality and larger fiscal implications.

A Generalized Model of Growth in the Data Economy (O3, O4)

Danxia Xie

,

Tsinghua University

Longtian Zhang

,

Central University of Finance and Economics

Abstract

In this paper, we build a generalized model to describe the long-run effects of different types of data from consumption-generated to production-generated. We first introduce two new types of data which are generated from the production process and the nature, i.e., producer data and nature data, respectively. Labor will be employed for collecting and cleaning the data before they can be used; however, there are no privacy concerns as assumed in previous studies, which all focus on consumer data. We explore along the balanced growth path and find a higher growth rate of the economy than that can be derived when only using data from consumers---a result that we have not found in existing studies. We then go further to the generalized model using all the types of data simultaneously, and provide a full picture on the transformations of different growth regimes. This framework has new and important implications for the data economy.

A Historical Housing Price Index from South China: 1570–1949 (N9, R3)

Siyu Pan

,

Northern Arizona University

Abstract

Economic studies on premodern housing markets have earned more attention in recent years, but such work is still restricted by the scarcity of data and the scope is limited to a few urban areas scattered across the globe. In this paper, I collect a novel housing transaction dataset with over 1000 observations from the historic Huizhou prefecture, a relatively rural area in south China that had a strong economic influence in late imperial China. This paper develops a housing price index that covers the late Ming Dynasty, the Qing Dynasty, and the Republic of China era (1570--1949) using hedonic regression models. Besides the size of the property, my results also reveal housing price determinants that are unique to the historical and cultural background of this market: the orientation of the property, familial ties between sellers and buyers, adjacency to the ancestral halls, and whether the seller is in an urgent financial situation.

A Real Theory of Aggregate Demand Shortages (E3, D6)

Ehsan Ebrahimy

,

International Monetary Fund

View Abstract

Abstract

How can low consumer spending cause an inefficient recession or recovery? We propose a theory, in which aggregate demand shortages are the result of aggregate demand externalities stemming from financial frictions rather than nominal rigidities. In our theory, an economy can be demand constrained when, i) consumption goods and assets used as saving vehicles are not perfect substitutes, ii) productivity in the consumption goods sector can be improved or maintained via investment, and iii) external financing of this investment is subject to a tight enough borrowing constraint. In a demand-constrained equilibrium, a saving tax can increase consumption, investment, employment, output, and welfare. A tighter firms' borrowing constraint implies more severe aggregate demand shortages in a demand-constrained equilibrium. We use stylized extensions of our benchmark model to show how financial shocks can create demand-constrained recessions. Our theory suggests that fiscal rather than monetary policy might become the primary policy tool to restore the efficiency under certain conditions. Demand management policies might be necessary over a long period of time. The "potential", i.e., the constrained efficient allocation, may depend on both supply and demand factors.

A Unified Framework to Estimate Macroeconomic Stars (E3, C5)

Saeed Zaman

,

Federal Reserve Bank of Cleveland

Abstract

We develop a flexible semi-structural time-series model to estimate jointly several macroeconomic “stars” — i.e., unobserved long-run equilibrium levels of output (and growth rate of output), the unemployment rate, the real rate of interest, productivity growth, the price inflation, and wage inflation. The ingredients of the model are in part motivated by economic theory and in part by the empirical features necessitated by the changing economic environment. Following the recent literature on inflation and interest rate modeling, we explicitly model the links between long-run survey expectations and stars to improve the stars’ econometric estimation. Our approach permits time variation in the relationships between various components, including time variation in error variances. To tractably estimate the large multivariate model, we use a recently developed precision sampler that relies on Bayesian methods. The by-products of this approach are the time-varying estimates of the wage and price Phillips curves, and the pass-through between prices and wages, both of which provide new insights into these empirical relationships’ instability in US data. Furthermore, our estimates of the stars are among the most precise. Lastly, we document the competitive real-time forecasting properties of the model and, separately, the usefulness of stars’ estimates as steady-state values in external models.

Access to High-Income Countries and Product Innovation: Evidence from China (F1, O3)

Wenshuang Yu

,

University of Calgary

Abstract

In this paper, I study the effect of access to markets in high-income countries on firms' product innovation in developing countries. To do so, I exploit a reform that exogenously increased market access for firms in the world's largest developing country, China. Applying an instrumental variables strategy to rich firm-level administrative data, I show that exporting to high-income countries leads firms to innovate new products. These gains were driven primarily by privately-owned firms. My estimates show that a ten percentage point increase in market access to richer countries caused a 1.9 percentage point increase in product innovation for privately-owned firms. These effects are induced by the increase in revenue produced by exporting to high-income countries.

Adaptive Diarchy: Authority between Two Heads (D8, H1)

Yucheng Qiu

,

Peking University

Zanhui Liu

,

Tsinghua University

View Abstract

Abstract

An organization is referred to as diarchy if it appoints two executives with the same responsibility. This paper explores rationale behind diarchial organizations, starting from their most intriguing feature: A diarchy is often composed of two agents with ambiguous authority allocation. We study a model where a principal must rely on a biased agent for information and execution. To seek extra advice, she can appoint a supervisor with intrinsic preferences over dissenting voice. However, the supervisor may collude with the agent and advocate the latter's advice in communication. For the principal, the only way to prevent potential collusion is to make the supervisor roving. Nevertheless, the signal disclosed by the roving supervisor may become completely uninformative. We show that when vertical conflict of interest is moderate, adaptive diarchy emerges in equilibrium, and is characterized by three features. First, there exists a stationary supervisor. Second, although ex ante contingent delegation is impossible with the principal's lack of necessary expertise, the allocation of real authority is adaptive ex post. Third, the supervisor and the agent are of slight difference of opinion and similar relative status. We then embed our baseline setting in an infinite-horizon model to explore the historical regime shift from roving to stationary supervisors. We show that adaptive diarchy emerges in the proceed of the principal's learning about the extent to which enhancement of control erodes agents' participation.

Affirmative Action Policies with Race-Neutral Preferences (I2, J1)

Kendall Kennedy

,

Mississippi State University

View Abstract

Abstract

The role of structural biases in the university admissions process has been the subject of much attention, but the economics literature has only modeled the existence of affirmative action policies as the result of university preferences for diverse student populations, or as affecting endogenous effort choices. This study presents a simple university admissions model that predicts affirmative action even in the absence of explicit preferences for racial diversity.

First, human capital production theory implies that Black students have greater returns to university attendance than White students of equal human capital. In a standard human capital production framework, mean human capital is greater for White students than for Black students due to well-known inequality in educational investment. However, if comparing two students with similar human capital, the Black student has a lower expected level of investment but a higher expected innate ability than the White student. If ability and investments are complements in human capital production, then attending a university increases the human capital of Black students by more than White students in expectation.

Second, following the canonical Arrow-Phelps models of statistical discrimination, Black students have a lower mean and higher variance in their human capital. Students generate noisy signals of their true human capital, but these signals are biased against Black students both additively and through unequal signal precision. Admissions officers observe race-specific signal distributions but cannot distinguish between signal bias and Black-White human capital differences in observed Black-White signal gaps.

Combining these two channels, affirmative action is optimal if any of the following are true: (i) there is a Black-White education investment gap and complementarity between innate ability and investment; (ii) there is additive bias disfavoring Black students in signaling; or (iii) signals are less precise for Black students and the university’s objective function is convex in human capital.

Agricultural Minimum Wage and U.S. Agricultural Employment (Q1, J3)

SongYi Paik

,

University of Minnesota

View Abstract

Abstract

American farmers have faced widespread labor shortages for years, and foreign-born workers have become an essential labor resource. Due to concerns of a depressive effect on the wages of American farmworkers, the US government requires employers who hire foreign workers through the H-2A program to pay at least the minimum wage, called the Adverse Effect Wage Rates (AEWRs) around 40-50% higher than the state minimum wages.

This paper examines the relationship between the AEWRs and agricultural employment, and I narrow agricultural employment down to the employment of three different groups of workers who are most likely affected by the AEWRs: less-educated domestic farmworkers, undocumented farmworkers, and foreign workers.

Using the Department of Labor's H-2A program data along with the 2005-2019 American Community Survey, I estimate the changed agricultural employment at the Public Use Microdata Areas (PUMAs) level based on a two-way fixed effects specification. I find the disemployment effect among employed agricultural workers and those with less education. The AEWRs also have heterogeneous effects on agricultural workers: one dollar increase in the AEWR is associated with a decrease of 93 domestic workers with less education (22 percent reduction of the mean) and a decrease of 41 undocumented workers (30 percent reduction of the mean) while there is little impact on foreign workers. Also, one dollar increase in AEWR reduces 126 non-Hispanic White workers while increasing 31 Hispanic workers.

Although the AEWR policy is designed to avoid any adverse effects on American agricultural workers, it raises concerns about its negative consequences of harming job opportunities for less-educated domestic agricultural workers. Moreover, an increase in farmers’ labor costs disproportionally affects job opportunities for non-Hispanic White. However, the jobs left by them appear not to be sufficiently filled by other workers such as foreign workers or domestic Hispanic workers.

Aid, Coup D'état, and Conflict: Evidence from Myanmar (O1, F3)

Changwei Zhan

,

National University of Singapore

View Abstract

Abstract

This study analyzes the aid-politics-conflict nexus by empirically examining aid-related conflict occurrence in the context of political turmoil. Specifically, I focus on the aftermath of the 2021 military coup d'état in Myanmar, which resulted in sustained nationwide conflicts. I combine aid actions and conflict events datasets and construct a village-year-month panel covering over 13,000 villages. I employ the Difference-in-Differences design and control for fine granular fixed effects and control variables derived from high-resolution geographical data. I find that villages that received pre-coup aid experienced a 1.2 percentage points higher post-coup conflict incidence than adjacent villages without pre-coup assistance. Aid is mainly associated with anti-authoritarian movements by protesters and political militias against the military force. The aid-coup-conflict relationship can be explained by the feeling of relative deprivation and ideologically winning hearts and minds. I also include other discussions such as cash transfer, dependency on local economy, foreign-controlled enterprises, the Rohingya crisis, and China’s power. This study adds to the longstanding debate on the relationship between aid and conflicts in the context of intense political upheaval.

An Affine Term Structure Model with Fed Chairs' Speeches (E5, G1)

Eunmi Ko

,

Rochester Institute of Technology

Abstract

This paper analyzes the impact of the sentiment expressed in speeches by Fed Chairs on bond yields, using an affine term structure model similar to the framework by Ang and Piazzesi (2003). The speeches by Fed Chairs are evaluated using finBERT (Huang, Wang, Yang, 2022) to generate three sentiment factors: negative, neutral, and positive. The variance decomposition analysis indicates that sentiment factors account for a considerable portion of the variance in short-term and long-term yield forecasts. Also, the inclusion of sentiment factors is helpful to improve out-of-sample forecast performance for relatively longer-term yields.

Analysis of the Bank of Japan’s Monetary Easing Policy since the 2000s (E5, G1)

Hideaki Ohta

,

Ritsumeikan University

Abstract

This paper empirically analyses and evaluates the impact of the Bank of Japan's (BOJ) series of monetary easing policies (the world's first Quantitative Easing (QE) [2001-06], Comprehensive Monetary Easing (CME) and Quantitative and Qualitative Easing (QQE)) on the real economy, exchange rates and financial markets over the past two decades from April 2001 to March 2020. The analysis based on a Bayesian vector autoregressive model (BVAR) shows that during QE, monetary easing had a clear and positive impact on the real economy and markets (exchange rate, interest rates and bank lending), including real GDP growth, business indices (Bank of Japan Tankan and Business Watchers Survey), industrial production and bank lending The CME also functioned relatively effectively during the worst period after the Global Financial Crisis and had a positive impact on the economy and markets, including business confidence indices, the real effective exchange rate and bank lending. The QQE, on the other hand, had little impact on Japan's real economy and financial markets, except for a slight rise in stock prices, despite the substantial expansion of the monetary base. The analysis in this paper therefore clearly shows that QQE failed to achieve its original objective of bringing the Japanese economy back from long-term stagnation of the economy.

Antitrust and (Foreign) Innovation: Evidence from the Xerox Case (O3, L4)

Robin Mamrak

,

Ludwig Maximilian University of Munich

Abstract

How does antitrust enforcement against patent-based monopolies affect innovation? I address this question by empirically studying the US antitrust case against Xerox, the monopolist in the market for plain-paper copiers. In 1975, Xerox was ordered to license all its copier-technology patents in the US and abroad. I show that this promoted innovation by other firms in the copier industry, measured by a disproportionate increase in patenting in technologies where Xerox patents became available for licensing. This positive effect is driven by increased innovation by Japanese competitors. They started developing smaller desktop copiers and their innovation became more diverse.

Artificial Intelligence and Dual Contract (D0, G3)

Qian Qi

,

Peking University

View Abstract

Abstract

With the dramatic progress of artificial intelligence algorithms in recent times, it is hoped that algorithms will soon supplant human decision-makers in various fields, such as contract design. We analyze the possible consequences by experimentally studying the behavior of algorithms powered by Artificial Intelligence (Multi-agent Q-learning) in a workhorse dual contract model for dual-principal-agent problems. We find that the AI algorithms autonomously learn to design incentive-compatible contracts without external guidance or communication among themselves. We em- phasize that the principal, powered by distinct AI algorithms, can play mixed-sum behavior such as collusion and competition. We find that the more intelligent princi- pals tend to become cooperative, and the less intelligent principals are endogenizing myopia and tend to become competitive. Under the optimal contract, the lower con- tract incentive to the agent is sustained by collusive strategies between the principals. This finding is robust to principal heterogeneity, changes in the number of players involved in the contract, and various forms of uncertainty.