Individual-specific Heterogeneity Research Articles

Cross-section asymptotic for random-effects panel data models with autoregressive errors

A useful strategy to examine serial correlation in panel regression models is to adopt serially correlated error terms. The random effect is also present to address the individual-specific heterogeneity. Familiar transformations can restructure the AR(1) process into a serially uncorrelated dynamic model. It may result in unfitting estimation results once the statistical inference is carried on the conditional likelihood. We discuss in the article the consequence of this action. In particular, in the case of large cross-section units and short time-sequences, we provide analytical expressions for the asymptotic bias of maximum likelihood estimates when the analyst naïvely neglects the initial conditions problem. Under several conditions, we compute the size of biases and measure the mean squared errors to illustrate this violation. We also construct a joint regression model of the initial and subsequent responses to handle the initial conditions. Then, we conduct simulation studies to determine the asymptotic behavior.

What are the triggers for arrears on debt over a business cycle? Evidence from panel data

AbstractThe paper investigates the triggers of arrears on debt in a full recourse euro area country. An extensive individual‐level quarterly panel dataset enables quarterly arrears on debt to be tracked while controlling for individual specific heterogeneity. The estimations show that lower income and higher debt service ratios are associated with a higher probability of arrears, confirming the “ability to pay” hypothesis. The results reveal different debt repayment behaviour for housing and consumer loans over a business cycle. Newly taken housing loans exhibit a lower probability of arrears and the same applies to loans granted during the period of easy credit conditions and high real estate prices, indicating that housing credit boom has not contributed to the rise in arrears on debt. Newly taken consumer loans are related to a higher probability of arrears, suggesting that consumer loans have been used to alleviate financial difficulties.

Precursors of intellectual property rights enforcement in East and Southeast Asia

In this study we identify the main determinants of perceived strength of intellectual property rights in four developed (Japan, Singapore, South Korea, and Taiwan) and five emerging (China, Indonesia, Malaysia, the Philippines, and Thailand) Asian countries over the period 2003–2016. We use a panel model with additive unobserved individual-specific heterogeneity in a high-dimensional setting. The setting allows the number of time-varying regressors to exceed the sample size. Based on the Cluster–Lasso approach, we found that (1) bribery and corruption (inverse), equal opportunity, administration of justice, knowledge transfer, personal security and private property rights, and qualified engineers are significant determinants of intellectual property rights for developed countries; (2) adaptability of government policy, bribery and corruption (inverse), bureaucracy (inverse), and science in schools are significant determinants of intellectual property rights for emerging countries. Policy makers may use these results to strengthen IP rights and thus encourage indigenous innovation as well as foreign direct investment.

Identifying heterogeneous factors for driver injury severity variations in snow-related rural single-vehicle crashes

Snowy weather is consistently considered as a hazardous factor due to its potential leading to severe fatal crashes. A seven-year crash dataset including rural highway single vehicle crashes from 2010 to 2016 in Washington State is applied in the present study. Pseudo elasticity analysis is conducted to investigate significant impact factors and the temporal stability of model specifications is tested via a likelihood ratio test. The proposed model based on the seven-year dataset is able to capture the individual-specific heterogeneity across crash records for four significant factors, i.e., surface ice, male, and airbag combine deployment for minor injury, and male for serious injury and fatality. Their estimated parameters were found to be normal distribution instead of fixed value over the observations. Other significant impact factors with fixed effects are: inroad object, animal, overturn, surface wet, surface snow, unusual horizontal design, medium and high speed limits, driver age, impaired condition, no belt usage, vehicle type, airbag deployment. Especially, when compared to significant factors for crashes under other weather conditions, male indicator and impaired condition show significant higher effects in snow-related crashes. The results of temporal stability test show that the model specification is generally not temporally stable for driver injury severity model based on the years of crash data that were used, especially for longer period (more than 3-year dataset). Models that allow the explanatory variables to track temporal heterogeneity, are of great interest and can be explored in future research.

Nonparametric estimation of the determinants of inefficiency in the presence of firm heterogeneity

We consider the application of the profile least-squares method to estimate the impact of the determinants of inefficiency in the presence of panel data and unobserved individual specific heterogeneity for the stochastic frontier model. This method has the advantage over previous approaches in that the effect of the determinants of inefficiency on output can be recovered nonparametrically in the presence of heterogeneity at the firm level. We describe the estimator and offer a small set of Monte Carlo exercises that showcases the value of the method for the stochastic frontier model. Application of the method to assess insights generated compared to a pooled cross-section approach using data from the Taiwan banking system is also provided.

Economic Determinants Affecting Military Expenditures : Panel Data Analysis

Our research revolves around the topic of considering the military expenditures per capita as a dependent variable and the GDP per capita and CO2 Emissions per capita as two explanatory variables. The study is made up of ten sections addressing several points, each of which clarifies the research method in order to reach a conclusion revealing the importance of the findings. Beginning with the basic statistical characteristics, such as averages, standard deviations, minimums, maximums and the Compound Annual Growth Rate (CAGR), a benefit use of the graph of each variable for each country has been highlighted for a better understanding of the rising and falling during its temporal evolution. The various aspects of the panel analysis have been completed as the questions of individual specific heterogeneity in panel data, the panel unit root tests using the most famous from the first and second generations, and the co-integration analysis according to the Pedroni’s approach, which has led to the rejection of the null hypothesis of no co-integration for each country and for the group as a whole.

Random parameter probit models to analyze pedestrian red-light violations and injury severity in pedestrian–motor vehicle crashes at signalized crossings

Pedestrian red-light violations at signalized crossings are an important traffic safety concern, particularly in densely populated cities. In this study, the authors quantitatively investigated the factors associated with pedestrian red-light violations and injury severity resulting from pedestrian–motor vehicle crashes at signalized crossings. Random parameter probit models are used to account for individual-specific heterogeneity that arises from a set of unmeasured factors related to traffic conditions and the pedestrians’ physical and mental status. Data for the analysis are based on the historical crash record maintained by the Hong Kong Transport Department. Children younger than age 11 years are not only associated with a higher likelihood of pedestrian red-light violations but also tend to have a higher probability of severe injuries. Factors including summer, dual carriageways with a central traffic island, and pedestrian age of 12 to 25 years are solely related to a higher likelihood of pedestrian red-light violations; meanwhile, variables solely associated with a higher probability of severe injuries include crashes that occur between 22:00 and 06:59, crashes occurring in rainy weather, crashes involving pedestrians older than age 46 years, and bus crashes. Appropriate countermeasures are recommended to curb pedestrian red-light violations and to reduce the injury severity of pedestrian crashes.

An analysis on perceived crowding level reported by domestic visitors of South Korean National Parks: a multilevel ordered logit approach

ABSTRACTPerceived Crowding Level (PCL) is an indicator of the social carrying capacity of recreational sites such as National Parks. Using a sample of national-level visitor survey data across 21 National Parks in South Korea, this paper aims to apply a multilevel ordered logit model as a method in testing for statistical relationships between the PCL and the covariates, accounting for site- and individual-specific heterogeneity. The results hint that levels of perceived crowding in South Korean National Parks are related to factors such as the behaviour of other users and the quality of park facilities such as comfort in accommodation. The measured level of heterogeneity can provide useful information about the level (nation-wide or park-specific) of intervention policies and management strategies.

Using fixed-parameter and random-parameter ordered regression models to identify significant factors that affect the severity of drivers’ injuries in vehicle-train collisions

This study attempts to identify significant factors that affect the severity of drivers’ injuries when colliding with trains at railroad-grade crossings by analyzing the individual-specific heterogeneity related to those factors over a period of 15 years. Both fixed-parameter and random-parameter ordered regression models were used to analyze records of all vehicle-train collisions that occurred in the United States from January 1, 2001 to December 31, 2015. For fixed-parameter ordered models, both probit and negative log–log link functions were used. The latter function accounts for the fact that lower injury severity levels are more probable than higher ones. Separate models were developed for heavy and light-duty vehicles. Higher train and vehicle speeds, female, and young drivers (below the age of 21 years) were found to be consistently associated with higher severity of drivers’ injuries for both heavy and light-duty vehicles. Furthermore, favorable weather, light-duty trucks (including pickup trucks, panel trucks, mini-vans, vans, and sports-utility vehicles), and senior drivers (above the age of 65 years) were found be consistently associated with higher severity of drivers’ injuries for light-duty vehicles only. All other factors (e.g. air temperature, the type of warning devices, darkness conditions, and highway pavement type) were found to be temporally unstable, which may explain the conflicting findings of previous studies related to those factors.

Life Satisfaction and Economic Position Relative to Neighbors: Perceptions Versus Reality

I examine the relationship between economic position relative to local neighbors and life satisfaction in rural Bangladesh. In particular, I exploit a novel household level census of three villages that includes the geospatial coordinates of every household and a perception measure of economic position relative to neighbors. This allows for exploring the sensitivity of the aforementioned relationship to (1) different objective definitions of neighborhood, and (2) the type of positional measure used, objective or perceived. I find that a higher perceived position improves life satisfaction while objective position has no statistically significant effect. This difference stems from a very low correlation between the objective and perceived positional measures. The low correlation can be explained by individual specific heterogeneity in definitions of neighbor. However, observable socioeconomic and demographic variables cannot explain the difference.

Risk factors affecting injury severity determined by the MAIS score

ABSTRACTObjective: Traffic crashes result in a loss of life but also impact the quality of life and productivity of crash survivors. Given the importance of traffic crash outcomes, the issue has received attention from researchers and practitioners as well as government institutions, such as the European Commission (EC). Thus, to obtain detailed information on the injury type and severity of crash victims, hospital data have been proposed for use alongside police crash records. A new injury severity classification based on hospital data, called the maximum abbreviated injury scale (MAIS), was developed and recently adopted by the EC. This study provides an in-depth analysis of the factors that affect injury severity as classified by the MAIS score.Method: In this study, the MAIS score was derived from the International Classification of Diseases. The European Union adopted an MAIS score equal to or greater than 3 as the definition for a serious traffic crash injury. Gains are expected from using both police and hospital data because the injury severities of the victims are detailed by medical staff and the characteristics of the crash and the site of its occurrence are also provided. The data were obtained by linking police and hospital data sets from the Porto metropolitan area of Portugal over a 6-year period (2006–2011). A mixed logit model was used to understand the factors that contribute to the injury severity of traffic victims and to explore the impact of these factors on injury severity. A random parameter approach offers methodological flexibility to capture individual-specific heterogeneity. Additionally, to understand the importance of using a reliable injury severity scale, we compared MAIS with length of hospital stay (LHS), a classification used by several countries, including Portugal, to officially report injury severity. To do so, the same statistical technique was applied using the same variables to analyze their impact on the injury severity classified according to LHS.Results: This study showed the impact of variables, such as the presence of blood alcohol, the use of protection devices, the type of crash, and the site characteristics, on the injury severity classified according to the MAIS score. Additionally, the sex and age of the victims were analyzed as risk factors, showing that elderly and male road users are highly associated with MAIS 3+ injuries. The comparison between the marginal effects of the variables estimated by the MAIS and LHS models showed significant differences. In addition to the differences in the magnitude of impact of each variable, we found that the impact of the road environment variable was dependent on the injury severity classification.Conclusions: The differences in the effects of risk factors between the classifications highlight the importance of using a reliable classification of injury severity. Additionally, the relationship between LHS and MAIS levels is quite different among countries, supporting the previous conclusion that bias is expected in the assessment of risk factors if an injury severity classification other than MAIS is used.

Housing Prices, Unemployment Rates, Disadvantage, and Progress Toward a Degree

Rising unemployment and housing price appreciation are associated with increased college enrollment. Enrollment does not, however, guarantee completion. We use a discrete time, competing hazard function that accommodates individual-specific heterogeneity to assess the impact changing unemployment and housing prices have on progress toward a college degree in the United States for students interviewed for the 1996-2001 Beginning Post-Secondary Survey. The results indicate that rising unemployment rates have at best a modest effect on six year graduation rates. Both boys and girls are, however, more likely to not be enrolled and less likely to have graduated at the six-year mark when housing prices appreciate, and this effect is more pronounced for more disadvantaged youth.

Evaluating the role of income, state dependence and individual specific heterogeneity in the determination of subjective health assessments

This paper investigates the role of various determinants of an individual's subjective self-assessment of own health. While the economics literature has focused primarily on the role of income on these assessments, we include an examination of the role of state dependence and unobserved individual specific time invariant heterogeneity. We employ a dynamic fixed effects ordered choice model to examine the responses of Australian residents. We find no statistically significant relationship between transitory income and health responses. We also find that while there is evidence of state dependence, this does not appear to be responsible for the distribution of responses. Our results suggest that the variation in the individual specific effects, comprising both observed and unobserved time invariant factors, is primarily responsible for the variation across individuals’ responses.

Mixed hidden Markov quantile regression models for longitudinal data with possibly incomplete sequences.

Quantile regression provides a detailed and robust picture of the distribution of a response variable, conditional on a set of observed covariates. Recently, it has be been extended to the analysis of longitudinal continuous outcomes using either time-constant or time-varying random parameters. However, in real-life data, we frequently observe both temporal shocks in the overall trend and individual-specific heterogeneity in model parameters. A benchmark dataset on HIV progression gives a clear example. Here, the evolution of the CD4 log counts exhibits both sudden temporal changes in the overall trend and heterogeneity in the effect of the time since seroconversion on the response dynamics. To accommodate such situations, we propose a quantile regression model, where time-varying and time-constant random coefficients are jointly considered. Since observed data may be incomplete due to early drop-out, we also extend the proposed model in a pattern mixture perspective. We assess the performance of the proposals via a large-scale simulation study and the analysis of the CD4 count data.

Inference in High-Dimensional Panel Models With an Application to Gun Control

We consider estimation and inference in panel data models with additive unobserved individual specific heterogeneity in a high-dimensional setting. The setting allows the number of time-varying regressors to be larger than the sample size. To make informative estimation and inference feasible, we require that the overall contribution of the time-varying variables after eliminating the individual specific heterogeneity can be captured by a relatively small number of the available variables whose identities are unknown. This restriction allows the problem of estimation to proceed as a variable selection problem. Importantly, we treat the individual specific heterogeneity as fixed effects which allows this heterogeneity to be related to the observed time-varying variables in an unspecified way and allows that this heterogeneity may differ for all individuals. Within this framework, we provide procedures that give uniformly valid inference over a fixed subset of parameters in the canonical linear fixed effects model and over coefficients on a fixed vector of endogenous variables in panel data instrumental variable models with fixed effects and many instruments. We present simulation results in support of the theoretical developments and illustrate the use of the methods in an application aimed at estimating the effect of gun prevalence on crime rates.

The Dynamics of Adolescent Depression: An Instrumental Variable Quantile Regression with Fixed Effects Approach

Summary The paper employs a recently developed instrumental variable approach for the estimation of dynamic quantile regression models with fixed effects to model the dynamics of health outcomes. Our proposed estimator not only allows us to control for individual-specific heterogeneity via fixed effects in the dynamic quantile regression framework but may also reduce the bias that exists in conventional fixed effects estimation of dynamic quantile regression models with small numbers of time periods. Using data on the children of the US National Longitudinal Survey of Youth 1979 cohort, we examine the extent of true state dependence in youth depression conditional on unobserved individual heterogeneity and family socio-economic status. Our results suggest that true state dependence in youth depression among the survey respondents is very low and the observed positive association between previous and current depression is mainly due to time invariant unobserved individual heterogeneity.

Time to burn (calories)? The impact of retirement on physical activity among mature Americans

Physical activity is crucial for maintaining and improving health, especially at advanced ages. While retirement increases the amount of time available for physical activity, there is only limited evidence regarding the causal effect of retirement on recommended levels of physical activity. Addressing this gap in the literature, we use data from the U.S. Health and Retirement Study to estimate the causal impact of retirement on meeting the federal government's 2008 Physical Activity Guidelines for Americans. Using official early and normal retirement ages as instruments for retirement, our causal IV analyses suggest significant positive effects of retirement on meeting the Guidelines. These effects are robust with regard to the treatment of unobserved individual-specific heterogeneity, the measurement of guideline compliance, the definition of retirement and respondents’ health insurance status. We also show that the effects of retirement on physical activity are larger for persons with higher levels of education and wealth.

Micro- and Macroeconomic Implications of Heterogeneity in the Production of Human Capital

We derive a tractable nonlinear earnings function that we estimate separately individual by individual using NLSY79 data. We obtain three ability measures, a rate of skill depreciation, a time discount rate, and a population-wide estimate of the human capital rental rate. We utilize these parameters to verify a number of heretofore untested theorems based on the life cycle model. We show how these human capital production function parameters relate to cognitive ability, personality traits, and family background. Finally, we show that accounting for individual-specific heterogeneity dramatically reduces estimates of population-wide persistence of permanent and transitory shocks by over 50 percent.

Income Inequality and Health: Panel Data Evidence from Canada

Abstract Using longitudinal data from the Canadian National Population Survey (1994–2006), this study examines the impact of income inequality on current health outcomes. The result suggests that once unobserved individual specific heterogeneity is controlled for, income inequality as measured by Gini Coefficient has no significant impact on current health status. This result holds true for contemporaneous income inequality as well as for lagged income inequalities. There are mixed results from the robustness check using various measures of income inequality. Decile Ratio (90P/10P) and Coefficient of Variation have no impacts on current health status. On the other hand, contemporaneous income inequality measured by Log Mean Deviation and Theil Index have significant negative effects on current health. All of the models suggest that absolute income has a significant positive effect on health status

The impact of recession on drinking and smoking behaviours in Canada

Using longitudinal data from the Canadian National Population Health Survey (1994-2009), this study examined the impact of macroeconomic conditions as measured by the provincial unemployment rate on individual alcohol drinking and smoking behaviour. After controlling for unobserved individual specific heterogeneity, the study found that for the overall sample, unemployment rate has a significant positive impact on weekly alcohol consumption as well as on the probability of being a binge drinker. The study further found that unemployment rate has a significant positive impact on the number of cigarettes smoked by the daily smokers. However, unemployment rate has no impact on the probability of being a smoker. The study further examined whether or not there is a gender difference in the impact of unemployment rate on drinking and smoking behaviour. The results suggest that the impact of unemployment rate on drinking and smoking behaviour is more pronounced for males than for females.