Research

Abstract: Private tutoring is widely used by parents to enhance their children’s academic performance. However, its effectiveness depends on both the quality of instruction and students’ study effort, which parents cannot directly control. Excessive workload may also lead to depressive symptoms, reducing learning efficiency. This paper focuses on two questions: (1) Is tutoring an efficient educational investment ; (2) How do depressive symptoms influence its efficiency. I develop a two-period dynamic structural model with two sequential decision makers—parents and children. Parents choose tutoring expenditure based on grades, expected human capital, and their preferences; children choose study effort considering mental health and perceived returns to effort. Using CFPS data on high school students, This paper jointly estimates educational investment and effort, incorporating the interaction between mental health and study effort.

Download Paper

Abstract: Background: Demand for mental health services, particularly publicly provided services, varies substantially at the local level. Accurate local es- timates of both demand and supply are essential for e¤ective planning and resource allocation. However, no data source currently provides county- level prevalence rates of mental illness within any U.S. state. Aims of the Study: In this paper, we estimate the prevalence of men- tal illness and identify potential shortages in public mental health service provision across counties in New York State. Methods: We estimate the probability of having a mental illness using a probit regression model based on demographic characteristics from the National Health Interview Survey (NHIS). These estimates are applied to population data from the American Community Survey (ACS) to gener- ate county-level prevalence estimates. We calculate mental health service de…cits as the di¤erence between estimated demand for public services and actual supply, using data from the New York O¢ ce of Mental Health 2019 (OMH). Results: Estimated prevalence of mental illness varies considerably across counties, ranging from 1.8% to 5.8%. Counties with higher average family incomes tend to exhibit lower prevalence rates. Overall, in most parts of New York, the supply of public mental health services generally meets county-level demand. Discussion and Limitations: Our prevalence-based demand estimates may understate true demand for public services, as not all individuals with mental illness seek care. Due to data limitations, we are unable to account for the size and service capacity of individual programs or to as- sess their cost-e¤ectiveness. Implications: Policymakers need accurate estimates of local mental health service needs and current provision to allocate resources e¤ectively. Better public information on local program capacity and outcomes is critical for evaluating coverage and cost-e¤ectiveness of mental health interventions.

Download Paper

Abstract: We consider many ways to use K6 as an explanatory variable in equations for marriage and employment. We find that a) it is inappropriate transform K6 into a binary variable; b) it is inappropriate to add the K6 variables into a single sum of Lykert scales; c) there are serious issues associated with using each of the K6 variables separately; and d) each possible way to use K6 has critical issues making its use problematic.

Download Paper

Abstract: Reducing the CO2 emissions of China's high energy-consuming industry is one of the most important prerequisites in achieving the target of peaking its total CO2 before 2030. In this paper, we build a theoretical model on identifying the determinants of CO2 emissions by incorporating the energy-augmenting technical progress in the analytical framework. Based on this work, the growth rates of CO2 emissions in China's six high energy-consuming industries are decomposed into four aspects including energy mix effect, scale effect, factor substitution effect, and energy-saving technical progress effect; and the peak paths of each industry's CO2 emissions are also studied using scenario analysis method. The results indicate that:1) the scale effect plays a dominant role in increasing the CO2 emissions for all the industries, especially during the tenth five-year plan; on the contrary, energy-saving technical progress effect and factor substitution effect have limited the increase of industrial CO2 emissions to a certain extent; the energy mix effect is extremely minor. 2) The energy-saving technical progress is the most effective way to reduce the CO2 emissions for four industries although the effect decreased gradually, and the rest two mainly rely on factor substitution effect. 3) In the high emission scenario, the total CO2 emissions of the six high energy-consuming industries continue to grow rapidly, and it is difficult to reach the peak before 2030; in the medium and low emission scenarios, the total CO2 emissions of China's high energy-consuming industry will peak around 2017 and 2023, respectively.

Download Paper