ATLANTIC ECONOMIC JOURNAL最新文献

This paper investigates the determinants of railway usage for passengers using traditional econometric and machine-enhanced methods. The analysis employs European Union data retrieved from The World Bank and Eurostat, spanning the years from 2000 to 2018 and covering 25 European Union countries. The size of the rail network, the gross domestic product per capita, and the fertility rate positively affect passenger traffic. Unemployment has a negative impact. Country-specific effects also appear. Policymakers, regulators, and private stakeholders in transportation, tourism, environmental protection, energy use, and economic development should focus on expanding the rail network, reducing unemployment, increasing GDP per capita, and adopting family-friendly practices. Such policies should be adapted to country-specific needs.

Adopting a test-optional admissions policy alters admissions, enrollment and, potentially, pricing by four-year colleges. With less academic information about applicants and test-optional market segmentation, those admitted under a test-optional policy may face higher prices and accumulate more debt. Combining student debt data from The Institute for College Access & Success from 2000 and 2003-2015 with test-optional admissions data from the National Center for Fair and Open Testing for 2000-2015 in a two-way fixed effects estimator with institution-specific trends and controls, we find that private college graduates admitted under a test-optional policy borrow $1,022 (2016$), or 4.1 %, more than those required to submit their scores. Using the Callaway and Sant'Anna (Journal of Econometrics, 225(2), 200–230, 2021) estimator with multiple time-period treatments, debt is 4.1 % to 8.3 % higher. The evidence suggests that this larger debt is not a result of pricing power through market segmentation, but by selective institutions using price as an additional screening process in response to less information about applicants. Applicants and their families must weigh the ease of applying test-optional against the higher net price and, potentially, larger debt associated with attending test-optional institutions.

Previous research shows that culture can influence economic preferences even when individual characteristics and environment are held constant. This study adds to that literature by measuring the influence of culture on preferences for economic redistribution in the United States. In following the epidemiological approach to assessing the role of culture, the study uses the 1972–2022 cumulative file of the General Social Survey and the 2002–2023 European Social Survey to link the preferences of individual Americans with the preferences that may have been passed down to them from their ancestral countries in Europe. Depending on the choice of covariates in each model, a one-unit increase (on a 1 to 5 scale) in a European country’s average preference for redistribution is associated with a significant 0.20 (p = 0.011) to 0.30 (p = 0.002) increase in the redistribution preference of the average American who has ancestry from that country. Similar effect sizes were obtained even when limiting the sample to respondents who are fourth-generation and higher Americans. Ancestral preference for redistribution also appears to influence political behavior, as it predicts political liberalism and Democratic party identification across all models. The findings suggest a meaningful, robust, and persistent role for culture in determining attitudes toward redistribution in the United States.

American college football attendance in Division III, the lowest level of National Collegiate Athletic Association-sanctioned play, has been ignored. Attendance and associated financial flows are far lower than the highest levels of competition. Although the scale of operations is lower, the relative indirect benefits to Division III colleges can be substantial. This study takes a long-run approach using a balanced panel (208 teams over 20 years) to discover primary driving forces on average season attendance from among economic, team performance and demographic measures. Fixed- and random-effects models are estimated. The state’s monthly unemployment rate and real travel cost exert no impact. Attendance decreases as real incomes rise, suggesting these games are inferior goods. Winning games matters. More wins in the current season and the previous two seasons, and more playoff appearances in the last ten years all increase attendance. Habit persistence is present although its magnitude decreases as the lag increases. An increase in lifetime winning percentage also increases attendance in the random-effects model. Greater male undergraduate enrollment increases attendance. Higher county populations reduce attendance in the random-effects model. Findings suggest that building a successful team, as measured by winning games, is the most important determinant of fan attendance in Division III. This relationship holds in the short run (i.e., current season, recent previous seasons) and over longer periods (last ten years and lifetime of the program). Even at this lower level of competition, there is pressure on football coaches to win.

Correlation among financial assets during the 2007–2008 financial crisis is a topic of interest among practitioners and academics, as the increase in correlation among previously considered uncorrelated classes during the crisis has negative implications for portfolio diversification. Previous research considered the rise in correlation throughout the crisis to be a result of liquidity. This paper aims to reexamine those results and test for evidence that a rise in correlation can be found by examining daily data from the Federal Reserve Economic Data database and Yahoo Finance. Correlation coefficients conditional on heteroskedasticity for stock and real estate returns are compared with each other, as well as with bill, note, and bond yields. By examining cross-asset correlations during the two years prior, during, and after the 2007–2008 financial crisis, the rise in cross-asset correlation during the crisis period can be attributed to a rise in volatility in financial markets at that time, and not due to a rise in the underlying linkage between asset classes. This rise in volatility can be linked to the lack of liquidity from the 2007–2008 financial crisis and supports a broader theoretical relationship between low liquidity, high volatility, and elevated measured correlations. These findings suggest that policy interventions aimed at restoring market liquidity during crises can help stabilize not only volatility but also investor confidence in diversification strategies. Perceived breakdowns in diversification may only be reflecting temporary volatility distortions, rather than structural failures in asset class separation.

Studies on regional trade agreements find heterogeneous trade-promoting effects that depend on the terms of the agreements, sectors and products covered, and the characteristics of signatory countries. The agreements also have different enforcement dates and implementation periods, affecting the identification, estimation, size, and interpretation of their effect on trade. This paper builds on recent advances in the difference-in-differences estimation to explicitly account for any heterogeneous and time-varying effects and multiple implementation periods of trade agreements. Using bilateral trade data between 1962 and 2020 from the Centre d'Études Prospectives et d'Informations Internationales, the estimates show an effect of about 25–30% 10 years after implementation, which is larger than the standard two-way fixed estimates. The estimates also show phasing-in effects, in which the effects become larger over time. Furthermore, accounting for anticipation of the agreements affects the timing, size, and significance of these phasing-in effects. Overall, these results highlight the adjustment process for trade policies to take effect.

The conventional method of estimating the effect of education on economic growth is to estimate cross-country growth regressions using the annual growth of gross domestic product as a dependent variable and average number of years of schooling in a population and other macroeconomic factors as explanatory variables. Since students in different countries with the same number of years in school have different learning outcomes, cross-country studies that use quantity of schooling instead of quality of educational attainment and/or individual skills fail to capture the impact of human capital on economic growth. This study attempts to fill this gap using recently developed quality-adjusted educational attainment data from 53 countries over a 40-year period (1970–2010) and examines the significance of educational skills on economic growth. Several macroeconomic factors (e.g., government expenditure, growth of fixed capital, average longevity of the population, and trade as a percentage of gross domestic product) were used as control variables. The primary data source for the macroeconomic variables was the World Bank World Development Indicators. Cross country data on quality-adjusted educational attainment data came from three diverse sources: Lee and Lee, Barro and Lee, and Altinok and Diebolt. The empirical study used the dynamic generalized method of moments model that addresses the potential problem of endogeneity and causality, common issues in growth regressions. The study found that when human capital is measured by skill and learning, it is strongly associated with economic growth. Since educational institutions are the major source for developing cognitive skills, policies to improve educational attainment through investment in education would promote economic growth.

This study analyzes the relationship between income inequality and water pollution in the United States. Unlike prior studies that rely on cross-country comparisons or primarily focus on air pollution, this study makes a novel contribution by examining how income inequality influences water pollution over time within the United States, providing insights relevant to national policy interventions. Using the autoregressive distributed lag model, both the short- and long-run effects of urbanization, income, and income inequality on water quality are examined. The analysis employs nationally aggregated data spanning from 1969 to 2020, using dissolved oxygen as the primary indicator of water quality. Data sources include the United States Census Bureau, the World Bank, the Federal Reserve Bank of St. Louis, and the Global Environmental Monitoring System. Our results indicate that a more equitable distribution of income has adverse effects on water quality in the long term. Conversely, the proportion of income accounted for by the top five percent of the population and urbanization both have adverse impacts on water quality, suggesting that income concentration at the top and fast urban growth are responsible for worsening water conditions. Increases in gross domestic product per capita improve water quality both in the short and long run. These findings highlight the importance of tailoring environmental policy to account for disparities in income structure. Improving water infrastructure in less affluent communities and regulating pollution-intensive activities in high-income areas could help mitigate inequality-driven environmental risks and support public health.

How does students’ time using digital devices on fun activities like gaming, social media, and watching videos for entertainment affect their academic outcomes? Data from the Organization for Economic Co-operation and Development’s 2015 Programme for International Student Assessment test of students’ math, science, and reading knowledge for 43 countries is used to investigate these questions. Estimation results of three-level hierarchical linear models with random intercept and random slope show that marginal increases in social media use frequency are associated with increases in science, math, and reading PISA scores (3.915, 4.506, and 5.705 points, respectively, all at the 1% statistical significance level). Marginal increases in gaming are associated with increases in science and math scores (2.216 and 1.972 points respectively, both at the 1% statistical significance level). Marginal increases in digital device use frequency for entertainment raise science (1.371 points at the 5% level) and reading (1.997 points at the 1% level) scores. These findings suggest that in general using digital devices for various fun uses is positively associated with academic outcomes, which should mitigate some parental and policymaker concerns over screen time.