Administrative Panel Data Research Articles

Differential responses to market competition by private and public hospitals in China: a longitudinal analysis

BackgroundChina launched a nationwide health-care reform in 2009. An important objective of the reform was to improve the health-care service quality by encouraging the entry and development of private hospitals and instilling market competition. The reform has led to a large increase in private hospitals, yet the effect on the hospital market and health-care service quality has not been systematically studied. This study evaluates the effect of the increase of private hospitals on the operations and health-care service quality of incumbent hospitals, and further explores differential responses to market competition between private and public hospitals. MethodsWe obtained administrative hospital panel data from the annual report of medical and health institutions in Chengdu, Sichuan province, China. The data are collected by the National Health and Family Planning Commission and cover all hospitals in Chengdu from 2007 to 2016. We excluded hospitals that recorded zero or missing values for either total revenue or total patient visits. The sample included a total of 603 hospitals, of which 208 were public hospitals. We used fixed-effects estimation by controlling for fixed effects of hospital, year, and district-by-year. We investigated six categories of outcome variables (hospital revenue, patient visits, the number of medical staff, mean yearly hospital expenditure, the number of medical devices, and inpatient mortality rate) and their subcategories. The independent variable was the ratio of the number of new private hospitals in a 10 km radius in a given year to the number of existing hospitals in a 10 km radius in the base year (2007). This variable measures the increase in intensity of competition from private hospital entries to the local hospital market. FindingsWhen the ratio of the number of private hospital entries to existing hospitals increased by 100%, there was a decrease in total revenue of (21·4%, p<0·0001) and patient visits to (23·7%, p<0·0001) incumbent private hospitals. By contrast, for incumbent public hospitals, total revenue only decreased by 1·3% (p=0·839) and patient visits increased by 1·6% (p=0·826). Furthermore, private hospitals cut total expenditure by 18·9% (p<0·0001). Specifically, the number of medical staff decreased by 4·8% (p=0·015) and total wage expenditure decreased by 13·5% (p=0·013). By contrast, public hospitals raised the number of medical staff by 5·3% (p=0·092) and increased the number of medical devices by 14·9% (p=0·048). Finally, inpatient mortality rate decreased in public hospitals by 3·43 deaths per 1000 admissions (p=0·001). By contrast, inpatient mortality rate decreases by 0·119 deaths per 1000 admissions (p=0·855) in private hospitals. InterpretationThe results show that the 2009 health-care reform has intensified hospital competition. Private and public hospitals respond differently to more intense competition: private hospitals reduce expenditure, whereas public hospitals raise health-care service quality. Our findings suggest that future pro-competition reforms need to consider differential responses to market competition between private and public hospitals. FundingJY acknowledges financial support from the Singapore Ministry of Education Academic Research Fund (no. FY2018-FRC3-006).

How Does Improvement in Commuting Affect Employees? Evidence from a Natural Experiment

We collect rich worker month-level administrative panel data from two companies for a two-year period prior to and after the opening of a nearby subway station, which significantly improved public transportation commutes for a subset of workers. Consistent with a simple principal-agent model where improvement in commute reduces the cost of effort for workers, we find a significant difference-in-differences (DID) increase (12.6% of the standard deviation) in bonus pay (which is strongly correlated to worker-level performance measures) for affected workers relative to coworkers not impacted by the subway. The bonus increase is larger for workers with more variance in performance measures (marketing personnel and non-managers), is positively correlated with commute time saved, and lower for workers with access to technology that facilitated telecommuting. We do not find that improved performance is simply a result of affected workers spending extra time at the workplace. Additionally, we find a significant relative decline in monthly exit hazard (by about 50%) for affected workers, and evidence for higher quality (conditional on wages) of new hires from affected areas.

The effect of primary care availability on antibiotic consumption in Hungary: a population based panel study using unfilled general practices

ObjectiveWe analyse the effect of primary care availability on antibiotic consumption and on the quality of antibiotic prescribing behaviour.DesignRetrospective panel design, secondary analysis of settlement-level administrative panel data (n=2320 settlements,...

Human capital spillovers and the churning phenomenon: Analysing wage effects from gross in- and outflows of high-skilled workers

The article estimates human capital externalities on wages originating from internal gross migration flows of high-skilled workers. We draw on rich administrative micro panel data that allow us to disentangle externalities from sorting and labour market supply and demand effects through an extensive set of time-varying fixed effects. We show that regional inflows and outflows of high-skilled workers occur simultaneously and that both are positively correlated. Given the existence of such a churning phenomenon, looking only at net migration flows might be misleading. Our econometric analysis indicates that inflows of high-skilled workers increase the wages of locals, whereas outflows decrease those wages. Although externalities from outflows outweigh those from inflows in the short run, the opposite holds in the long run. Our results suggest that human capital externalities are transmitted through the productivity effects of local personal networks, which, for newcomers, develop over time.

Opportunities and competition in thick labor markets: Evidence from plant closures

AbstractSince Marshall (1890), it has been widely held in urban economic theory that cities insure workers against the risk of unemployment by offering a larger pool of potential jobs. Using a large administrative panel data set on workers displaced as a result of plant closures, we examine whether positive effects from a higher urban job density are offset by more intense competition between workers. When controlling for the sorting of workers between regions, we find robust evidence that the effect of job competition on unemployment duration exceeds that of job opportunities in absolute value. Our results put the idea of urban risk‐sharing into perspective and provide an explanation for observed longer unemployment durations in cities.

Measuring Job Creation, Growth, and Survival among the Universe of Start-ups in the United States Using a Combined Start-up Panel Data Set

The field of entrepreneurship is growing rapidly and expanding into new areas. This article presents a new compilation of administrative panel data on the universe of business start-ups in the United States, which will be useful for future research in entrepreneurship. To create the US start-up panel data set, the authors link the universe of non-employer firms to the universe of employer firms in the Longitudinal Business Database (LBD). Start-up cohorts of more than five million new businesses per year, which create roughly three million jobs, can be tracked over time. To illustrate the potential of the new start-up panel data set for future research, the authors provide descriptive statistics for a few examples of research topics using a representative start-up cohort.

Consumption Network Effects

AbstractIn this article we study consumption network effects. Does the consumption of our peers affect our own consumption? How large is such effect? What are the economic mechanisms behind it? We use administrative panel data on Danish households to construct a measure of consumption based on tax records on income and assets. We combine tax record data with matched employer–employee data to identify peer groups based on workplace, which gives us a much tighter and credible definition of networks than used in previous literature. We use the non-overlapping network structure of one’s peers group, as well as firm-level shocks, to build valid instruments for peer consumption. We estimate non-negligible and statistically significant network effects, capable of generating sizable multiplier effect at the macro-level. We also investigate what mechanisms generate such effects, distinguishing between intertemporal and intratemporal consumption effects as well as a more traditional risk sharing view.

Employment and Wages over the Business Cycle in Worker-Owned Firms: Evidence from Spain

This paper compares worker-owned firms and mainstream capital-owned enterprises over the business cycle. Specifically, I study whether conventional employees in worker-owned firms enjoy greater employment stability than similar workers in traditional enterprises over the business cycle, and investigate whether this stability is associated with greater volatility of working-time or wages. Unlike the literature that has compared partners of cooperatives to wage-earners of mainstream firms, I compare wage-earners across both type of organizations along the three margins of adjustment. To perform the econometric analysis, I rely on rich Spanish administrative data and panel data methods to account for composition differences between the two types of organizations. The results show that worker-owned firms offer higher job security because they do not adjust employment levels over the business cycle as much as mainstream enterprises. Wages and working-time, instead, are equally responsive across the two types of firms. The findings can be rationalized by the presence of similar labor regulations and differences in the objectives of the two type of organizations. Namely, both types of firms are constrained by regulations, such as the national Labor Code and collective bargaining, on the adjustments they can impose on wages and working-time. However, the social nature of worker-owned firms mitigates employment volatility in these organizations.

MPC Heterogeneity and Household Balance Sheets

We use sizeable lottery prizes in Norwegian administrative panel data to characterize households’ marginal propensities to consume (MPCs). Our main contribution is to document how MPCs vary with household characteristics and prize size, and how lottery prizes are spent and saved over time. We find that spending spikes in the year of winning, and reverts to normal within 5 years. Controlling for all items on households’ balance sheets and characteristics such as education and income, it is the amount won, age and liquid assets that vary systematically with MPCs. Low-liquidity winners of the smallest prizes (around USD 1,500) are estimated to spend all within the year of winning. The same estimate for high-liquidity winners of large prizes (USD 8,300-150,000) is slightly below one half. While conventional models will struggle to account for such high MPC levels, we show that a two-asset life-cycle model with a realistic earnings profile and a luxury bequest motive can account for both the time profile of consumption responses and their systematic co-variation with observables.

How School Socioeconomic Status Affects Achievement Growth across School Transitions in Early Educational Careers

Our study investigates how changing socioeconomic status (SES) composition, measured as percentage free and reduced priced lunch (FRL), affects students’ math achievement growth after the transition to middle school. Using the life course framework of cumulative advantage, we investigate how timing, individual FRL status, and legacy effects of a student’s elementary school SES composition each affect a student’s math achievement growth. We advance research on school transitions by considering how changing contexts affect achievement growth across school transitions. Furthermore, we improve on school context research by measuring the ways that SES compositions across school transitions may be interconnected. Using state administrative panel data for third through eighth graders from 2009 to 2015, we use fixed effects to estimate math achievement growth by the end of eighth grade. Findings suggest that a student’s elementary SES composition has a legacy effect on middle school achievement growth net of his or her own achievement growth and middle school SES composition. In addition, SES composition effects differ depending on the timing of exposure and a student’s individual FRL status. Our study has important implications for both educational transition research and school effects research, which are discussed.

Teachers, Electoral Cycles, and Learning in India

Teachers are central for learning, but their hiring and management can be influenced by political factors, especially when they are civil servants. Using an administrative school-level panel data set for India, we find significant increases in teacher transfers and moderate increases in hiring after state elections. The timing of these elections is pre-determined and staggered across states. The reorganization of teachers is partly related to political turnover. Pupils’ test scores are also up to 0.15 standard deviations lower in the post-election phase. Following a range of estimations for alternative explanations, we conclude that the reorganization of teachers can disrupt learning. [ABSTRACT FROM AUTHOR]

Negative home equity reduces household mobility: Evidence from administrative data

We investigate the impact of negative home equity on household residential mobility. We employ a unique administrative panel data set covering nearly the whole population of Dutch homeowners. The data contains a large set of personal characteristics of more than 2,000,000 homeowners. We identify the effect of negative home equity on mobility by exploiting the quasi-exogenous regional variation of the unanticipated housing market shock during 2008–2012, in which real house prices declined by 25%. Households plunging into negative home equity due to unanticipated declining house prices are 21% less likely to move compared to households upholding positive home equity.

Does class size matter for school tracking outcomes after elementary school? Quasi-experimental evidence using administrative panel data from Germany

We use administrative panel data on about a quarter of a million students in the German state of Hesse to estimate the causal effect of class size on school tracking outcomes after elementary school. Our identification strategy relies on the quasi-random assignment of students to different class sizes based on maximum class size rules. In Germany, students are tracked into more or less academic middle school types at about age ten based, to a large extent, on academic achievement in elementary school. We mostly find no or small effects of class size in elementary school on receiving a recommendation or on the actual choice to attend the more academic middle school type. For male students, we find that an increase in class size by 10 students would reduce their chance of attending the higher school track—which more than 40% of students attend—by 3 percentage points.

Motivating emissions cleanup: Absolute vs. relative performance standards

This article empirically compares the effectiveness of relative versus absolute performance standards in motivating compliance actions. By leveraging a unique set of Chilean administrative panel data, I examine a natural experiment created by a change in the performance standard used to incentivize the reduction of particulate matter in the atmospheric emissions of stationary pollution sources in the Santiago Metropolitan Region. I find that the absolute standard drove ∼21% less emissions cleanup than did the relative standard. I also demonstrate how sharp heterogeneity in responses to the change in standards is predictable based on an ex-ante identifiable type-categorization, which leads to the broadly-applicable conclusion that stricter regimes drive more compliance actions when imposed via an absolute performance standard compared to a relative standard. The extension of this framework provides a general means of anticipating whether an absolute or relative performance standard will drive higher rates of compliance actions in other settings.

Who are the Socially Responsible Investors?

We use a large administrative panel data set to study which characteristics are related to households' investments in socially responsible investing (SRI) mutual funds. To isolate financial and non-financial preferences for SRI investments, we distinguish between two types of SRI funds; ESG and charitable funds. We also analyze to what extent affluence and age play a role in SRI. We find that participation in SRI funds is lower for young and retired investors. Moreover, we find that young adults are more often ESG investors, while retired investors are more likely to participate in charitable funds. Further, financial wealth is important for SRI participation but becomes less relevant once investors are sufficiently rich. For the exposure to SRI funds, the overall picture is quite different as it is negatively related to wealth, income, risky share, and education.

Do boys benefit from male teachers in elementary school? Evidence from administrative panel data

With girls having overtaken boys in many education indicators, the “feminization” of elementary school teaching is generating debate about its disadvantages for male students. Hence, using administrative panel data on the universe of students, teachers, and schools in one German state, I exploit within-school and within-teacher variation to test for teacher gender effects on the type of middle school recommended and then chosen in the German multitrack educational system. The results indicate barely any effect of teacher gender on either recommended or chosen school type for either boys or girls. Even when the analysis follows students into middle school, no effects of elementary school teacher gender are discernible on either subsequent changes in school type or grade repetition.

Value added, wages and labour market flows at the establishment level

ABSTRACTIn this article, we analyse the connection between value added, wages and labour market flows at the establishment level. We develop a simple model to illustrate the expected comovement of these variables. For the empirical analysis, we link the new German Administrative Wage and Labor Market Flow Panel data set to the IAB Establishment Panel. We show that establishments’ hires rates have a positive and separations rates a negative comovement with establishment-specific value added, whereby hires react by more than separations. In addition, we provide evidence that establishments’ partial equilibrium reaction is an important driver for aggregate labour market dynamics.

How does retirement affect healthcare expenditures? Evidence from a change in the retirement age.

Using individual-level administrative panel data from Hungary, we estimate causal effects of retirement on outpatient and inpatient care expenditures and pharmaceutical expenditures. Our identification strategy is based on an increase in the official early retirement age of women, using that the majority of women retire upon reaching that age. According to our descriptive results, people who are working before the early retirement age have substantially lower healthcare expenditures than nonworkers, but the expenditure gap declines after retirement. Our causal estimates from a two-part (hurdle) model show that the shares of women with positive outpatient care, inpatient care, and pharmaceutical expenditures, respectively, decrease by 3.0, 1.4, and 1.3 percentage points in the short run due to retirement. These results are driven by the relatively healthy, by those who spent some time on sick leave and by the less educated. The effect of retirement on the size of positive healthcare expenditures is generally not significant.

The Impact of the Financial Crisis on Homeowners' Portfolio Choice

This paper examines the mediating effect of the recent financial crisis on the relationship between house value fluctuation and households’ liquid portfolio choice. To isolate exogenous variation in homeowners' home equity and mortgage debt before and after the crisis, I use a new regional level construction cost index along with other commonly used house price determinants as instruments. Using an administrative register-based panel data for the entire Danish population in the period 2004 – 2012, I find that the effect of housing on households' risky asset holdings is asymmetric before and after the crisis. The elasticity of risky asset shares with respect to mortgage debt and home equity is -0.31 and 0.28 respectively in pre-crisis period. On the other hand, whereas for the post-crisis period, the elasticity of risky asset shares with respect to mortgage debt and home equity is -0.37 and 0.34 respectively. Suppose an average household had spent 10% more on housing, the estimates suggest that they would hold 6% less in risky shares pre-crisis and 7% less post-crisis. Homeowners rebalance their liquid portfolio to a larger extent in response to their housing value fluctuation after the financial crisis, adding an additional effect that is likely to exacerbate the instability of an already tumultuous financial system.

Financial Literacy and Portfolio Dynamics

ABSTRACTWe match administrative panel data on portfolio choices with survey measures of financial literacy. When we control for portfolio risk, the most literate households experience 0.4% higher annual returns than the least literate households. Distinct portfolio dynamics are the key determinant of this difference. More literate households hold riskier positions when expected returns are higher, they more actively rebalance their portfolios and do so in a way that holds their risk exposure relatively constant over time, and they are more likely to buy assets that provide higher returns than the assets that they sell.