Cross-sectional Time-series Analysis Research Articles

Beyond Policy Diffusion: Spatial Econometric Models of Public Administration

AbstractInterdependence in the decision-making or behaviors of various organizations and administrators is often neglected in the study of public administration. Failing to account for such interdependence risks an incomplete understanding of the choices made by these actors and agencies. As such, we show how researchers analyzing cross-sectional or time-series-cross-sectional (TSCS) data can utilize spatial econometric methods to improve inference on existing questions and, more interestingly, engage a new set of theoretical questions. Specifically, we articulate several general mechanisms for spatial dependence that are likely to appear in research on public administration (isomorphism, competition, benchmarking, and common exposure). We then demonstrate how these mechanisms can be tested using spatial econometric models in two applications: first, a cross-sectional study of district-level bilingual education spending and, second, a TSCS analysis on state-level healthcare administration. In our presentation, we also briefly discuss many of the practical challenges confronted in estimating spatial models (e.g., weights specification, model selection, effects calculation) and offer some guidance on each.

Analysis of Trends in Insulin Utilization and Spending Across Canada From 2010 to 2015.

To describe insulin utilization and spending across Canada and investigate how interprovincial variations in long-acting insulin uptake impact provincial spending. We conducted a cross-sectional time-series analysis of insulin products dispensed nationally from January 1, 2010, to December 31, 2015, using data from IQVIA (Durham, North Carolina, United States). Analysis was stratified according to insulin type, payer and province. We report annual numbers for national insulin dispensing and spending and provincial numbers for publicly funded long-acting insulin dispensing and spending rates that are standardized by public drug beneficiary enrolment and diabetes prevalence. We report the percent of change of an annual provincial cost-to-utilization index of total insulin spending to total insulin dispensing between 2010 and 2015. Between 2010 and 2015, total insulin utilization increased 21% (4.4 million to 5.3 million prescriptions), and total insulin costs increased 54% ($345 million to $530 million) nationally. The national dispensing rate of long-acting insulin (+96%) and rapid-acting insulin (+38%) increased, while the national dispensing rate for intermediate-acting (-23%), short-acting (-37%) and premixed (-28%) insulins declined. Large interprovincial variation was observed for the rate of long-acting insulin uptake (range, Alberta, +1,505%; British Columbia, +27%) and the rate of long-acting insulin spending (Alberta, +2,177%; British Columbia, +44%) between 2010 and 2015 after standardization. Provinces with higher rates of long-acting insulin uptake experienced faster increases in their cost-to-utilization index (Alberta, +78%; British Columbia, +24%). Overall, the rate of uptake of long-acting insulins has increased nationally. Uptake varies widely among provinces and is likely to be associated with differential cost increases across public payers in Canada.

Social expenditure, social rights, and benefit receipt as indicators of welfare state generosity

PurposeRevitalising the debate about how to operationalise and measure the extent of welfare states – the so-called dependent variable problem – recent research claims a close theoretical interaction between three different indicators: aggregated data on social expenditure, social rights and social benefit receipt. It is suggested that they all serve as an approximation of welfare state generosity as a dependent variable and help understand variation between as well as change of welfare states. The purpose of this paper is to investigate how these three indicators statistically relate to each other, using data on unemployment cash benefits.Design/methodology/approachTo this end, a time series cross-sectional analysis is carried out, covering 16 European countries for the period 2003–2011.FindingsResults confirm theoretical reflections on the link between the different indicators, whereby higher levels of social expenditure are positively associated with more generous social rights as well as higher levels of benefit receipt. Additionally, the study points to an ambivalent relation between benefit access and benefit levels within indicators as well as across them. This suggests competing policy choices in European welfare states, whereby more generous benefit access implies lower benefit levels and vice versa.Originality/valueThe study contributes to the existing dependent variable literature in a twofold way. First, the conceptual link between the three different indicators and the assumptions they are associated with are critically reviewed. Second, by providing a statistical analysis of the relation between the different indicators across 16 countries, the study goes beyond theoretical elaborations about their association as well as existing small-N or medium-N case time trend studies.

Determinants of State Spending Patterns in Arab League Member States: a Post-Arab Spring Analysis, 1996–2014

What is the purpose of the modern nation-state and what drives its main functions? This analysis attempts to provide an answer by examining the correlates of state spending on military and public health in the Middle East and North Africa. I draw on theories of states and power, identifying the dual nature of modern nation-states, and argue that spending on military and social policy is shaped largely by domestic institutional arrangements. Theories from comparative politics, sociology, and political economy are operationalized. I consider hypotheses derived from each using pooled cross-sectional time series analysis with macroeconomic data and sociopolitical indicators. Analysis shows that democratic institutions generate spending in both areas, while governing capacity is associated with decreases in military and increases in public health spending. The regression also suggests the unrest of 2011 generated increases in military spending. Overall, these findings provide support for the notion that institutional arrangements largely determine state spending priorities and provide other insights relevant to the theories being tested. This paper extends our understanding of state behaviors vis-a-vis state spending in the context of social unrest. In light of these findings, I theorize that state behavior is molded by the triangular relationship between state, economy, and civil society, and I discuss the need to consider military expenditures in institutional analyses.

Always Running: Candidate Emergence among Women of Color over Time

The number of women seeking congressional office in the United States has dramatically increased since 1980. Previous research on women candidates explores why women run, but new research on candidate emergence shows that women face different challenges and advantages based on their race and ethnicity. We investigate these differences by disaggregating data on women’s candidate emergence by race and ethnicity to examine how these theories work when explicitly considering race and ethnicity. We focus our examination on women running in House primaries between 1980 and 2012. We argue that theories of candidate emergence are conditional to the racial and/or ethnic identification of the candidate. We employ a cross-sectional time series analysis with the intuition that examining congressional elections over time will allow us to make general comments about the participation of women in congressional elections. We find that many of the conditions thought necessary for women’s emergence as candidates are contextual and temporally specific. Moreover, conditions that encourage women to run do not necessarily apply to women of color.

Measuring and explaining the EU’s effect on national climate performance

ABSTRACTTo what extent has the European Union (EU) had a benign or retarding effect on what its member states would have undertaken in the absence of EU climate policies during 2008–2012? A measurement tool for the EU policy’s effect is developed and shows a benign average EU effect with considerable variation across countries. The EU’s policy effectiveness vis-à-vis its member states is explained by the EU’s non-compliance mechanism, the degree of usage of the Kyoto flexible mechanisms, and national pre-Kyoto emission reduction goals. Time-series cross-sectional analyses show that the EU’s non-compliance mechanism has no effect, while the ex-ante plans for using Kyoto flexible mechanisms and/or the ambitious pre-Kyoto emission reduction targets allow member states to escape constraints imposed by EU climate policy.

Insured flood damage in Sweden, 1987–2013

This study uses insurance claims as a proxy for property damage to analyse flood damage in Sweden between the years 1987 and 2013. The number of compensated insurance claims per year has risen rapidly during this period. As much as 70% of the claims are caused by flood damage occurring during the summer months June, July, and August, when intense rain with low predictability is common. To explore the damage trend a time series cross sectional analysis using four different fixed effect models was applied to the data set. Due to data scarcity, the time series had to be limited to 16 years and contain a total of 304 damage observations. The potentially explanatory climate related factor extreme rain, defined as >6 mm/15 min, and the socioeconomic factors gross regional product (GRP) per capita and housing stock were tested as explanatory factors. The GRP per capita and housing stock were found to be significant in two regression models. The estimated effect of extreme rainfall events exceeded the effects of GRP per capita and housing stock in the models. Extreme rain was robust to model specification and was found to have a highly significant impact on Swedish flood damage.

Explaining Women’s parliamentary representation: Are there differences between democracies and autocracies?

Do the same constituents trigger high women’s representation in democracies and autocracies? I try to address this research question with the help of a large-scale dataset containing the dependent variable women’s representation in national parliaments and six relevant independent variables—legally binding quotas, party quotas, the district magnitude, the years women have had parliamentary representation, development, workforce participation and a time trend—for 147 countries over a 25-year period. I find through cross-sectional time series analysis that the influence of the three institutional factors—legally binding quotas, and party quotas, as well as the district magnitude—differs between democracies and autocracies. Legally binding quotas have a somewhat larger influence in autocracies, whereas party quotas and a large district magnitude increase women’s representation more in democracies.

Assessing the relationship between healthcare market competition and medical care quality under Taiwan's National Health Insurance programme.

There is still significant uncertainty as to whether market competition raises or lowers clinical quality in publicly funded healthcare systems. We attempted to assess the effects of market competition on inpatient care quality of stroke patients in a retrospective study of the universal single-payer health insurance system in Taiwan. In this 11-year population-based study, we conducted a pooled time-series cross-sectional analysis with a fixed-effects model and the Hausman test approach by utilizing two nationwide datasets: the National Health Insurance Research Database and the National Hospital and Services Survey in Taiwan. Patients who were admitted to a hospital for ischemic or hemorrhagic stroke were enrolled. After excluding patients with a previous history of stroke and those with different types of stroke, 247 379 ischemic and 79 741 hemorrhagic stroke patients were included in our analysis. Four outcome indicators were applied: the in-hospital mortality rate, 30-day post-operative complication rate, 14-day re-admission rate and 30-day re-admission rate. Market competition exerted a negative or negligible effect on the medical care quality of stroke patients. Compared to hospitals located in a highly competitive market, in-hospital mortality rates for hemorrhagic stroke patients were significantly lower in moderately (β = -0.05, P < 0.01) and less competitive markets (β = -0.05, P < 0.01). Conversely, the impact of market competition on the quality of care of ischemic stroke patients was insignificant. Simply fostering market competition might not achieve the objective of improving the quality of health care. Other health policy actions need to be contemplated.

International Development NGOs and Bureaucratic Capacity: Facilitator or Destroyer?

Do development international nongovernmental organizations (INGOs) facilitate or destroy the bureaucratic capacity of the states in which they operate? The literature is split on this question. Some scholars argue that development INGOs weaken state capacity by delivering social services that the government is supposed to provide. Others argue that by increasing a country’s domestic demand for improved human rights, development INGOs improve a government’s capacity to fulfill them. In this paper, we show that the effect of development INGOs on state capacity depends on whether a state is democratic or nondemocratic. In our cross-sectional time-series analysis, we find that development INGO presence has a significant positive relationship with state capacity in democracies but no relationship with state capacity in nondemocratic states. These findings help explain the inconsistent claims in the existing INGO literature and are also relevant for development INGOs and the policymakers that support them.

Marching Orders? U.S. Party Platforms and Legislative Agenda Setting 1948–2014

What is the relationship between the priorities expressed in party platforms before an election and the subsequent legislative agenda? The agenda setting literature often deemphasizes the role of political parties in agenda setting, instead focusing on the importance of problems bubbling up to the surface and demanding attention from policymakers. However, parties will often express different issue priorities during elections, and compete based on those priorities. If those promises are credible, voters should be able to choose between different sets of priorities during elections. The paper utilizes new data from the U.S. Policy Agendas Project and Wolbrecht on policy attention in U.S. party platforms to study the relationship between U.S. parties and legislative activities in Congress. A time-series cross-sectional analysis finds strong evidence to support the proposition that legislative agendas are influenced by the platform of the President’s party in the short term, although the relationship differs for different types of agendas and by issue, and fades over time.

LO90: Epidemiologic trends in substance and opioid misuse in Alberta: a cross-sectional, time-series analysis

Introduction: Substance and opioid misuse are growing public health concerns in Canada. Substance use disorders affect 21.6% of Canadians and accounted for $267 million in healthcare costs in 2011. Opioid misuse is a current public health crisis. The extent of the rise in substance and opioid misuse-related Emergency Department (ED) visits in Canada and the demographic groups in which the rise is concentrated have not been elaborated. Alberta has one of the most complete provincial ED visit records and provides an important understanding of national trends. The objective of this study was to evaluate trends in substance and opioid misuse-related ED visits in Alberta from 2010/11 to 2014/5 within demographic cross-sections of the population using administrative ED visit data from the National Ambulatory Care Reporting System (NACRS). Methods: All visits made by adult patients (18 years old) to any of more than 100 Albertan EDs for a substance misuse-related presentation between 2010/11 and 2014/15 were analyzed. Visits were classified as being related to substance or opioid misuse if the primary and/or secondary visit diagnoses were among an a priori determined group of ICD-10 codes. Annual substance misuse-related visits were compared as visits per 100,000 adult population in Alberta to standardize for population growth. Linear regression was used to assess whether ED visits increased significantly over time. A cross-sectional time-series analysis was employed to examine trends within subgroups defined by sex and age categories (18-29, 30-39, 40-49, 50-59, and 60 years) over a 60-month period. Results: 149,719 substance misuse-related visits were made by 65,089 patients and 8768 opioid misuse-related visits were made by 5763 patients. From 2010/11 to 2014/15, substance misuse-related ED visits in Alberta increased by 38% from 811 to 1119 visits per 100,000 population. Opioid misuse-related ED visits increased significantly (64%) from 44 to 72 per 100,000 population. Conversely, total ED visits per 100,000 population did not increase significantly. Substance and opioid misuse-related visits rose more in non-rural than rural areas. Cross-sectional time-series analysis showed that the greatest increase in substance and opioid misuse-related ED visits occurred in males and in the 18-29 year age category, in which visit increases for opioid misuse appeared exponential. Conclusion: Substance and opioid misuse-related ED visits increased significantly from 2010/11 to 2014/15 in Alberta, with the most dramatic increases occurring in young patients and males. These findings have important implications for targeting urgent preventative public health interventions to stem the rise of this epidemic.

The political agenda-setting power of the media: the Europeanization nexus

ABSTRACTPrevious studies have demonstrated that the extent to which media coverage influences the issue priorities of policy-makers is contingent on the type of issues, media, and political agendas. This article contributes to this literature by elaborating on a factor that has been surprisingly neglected so far: the domestic or Europeanized character of the political issue covered in the news. Empirically, we apply time-series cross-section analyses to a dataset on media and parliamentary agendas during the years 1995–2003 in Switzerland. We find that the media’s political agenda-setting power mainly stems from news coverage of domestic issues. News on Europeanized issues have a weak impact on so-called symbolic parliamentary agendas, and no impact at all on the more substantial parliamentary agendas that may initiate decision-making processes.

Epidemiologic trends in substance and opioid misuse-related emergency department visits in Alberta: a cross-sectional time-series analysis.

Substance and opioid misuse are growing public health concerns. This study's objectives were to evaluate trends in substance and opioid misuse-related emergency department (ED) visits in Alberta, Canada. This is a cross-sectional time-series analysis utilizing National Ambulatory Care Reporting System ED data from Alberta, Canada. All substance and opioid misuse-related visits made by adults (≥ 18years) from 2010/11 to 2014/15 were analyzed. Acuity was measured by the Canadian Triage and Acuity Scale (CTAS). Relevant visits were identified by ICD-10 diagnostic coding. Substance and opioid visits over 60months were compared to all ED visits per 100,000 adult population using regression analysis, while controlling for temporal and seasonal variation. Trends among age and sex subgroups were also evaluated. From 2010/11 to 2014/15, substance and opioid misuse-related visits increased by 38.0% and 57.3% to 1119 and 118 visits per 100,000 population, respectively. Annual growth rates for substance and opioid visits were 4.4% higher (95% CI: 2.2, 6.7) and 10.6% higher (95% CI: 6.8, 14.6) than all ED visits. The 18-29year-old category experienced the highest annual growth rate of all age groups, and the annual opioid visit growth rate was 5.6% higher among males than females. Compared to all visits, substance misuse-related visits arrived more frequently by ambulance, were higher acuity, and were hospitalized more often. Substance and opioid misuse-related ED visits increased significantly from 2010 to 2015, especially among younger patients. Future research should elaborate causes and evaluate interventions to curb the growth of this issue.

Political and legal antecedents of affirmative action: a comparative framework

AbstractMuch of the literature on affirmative action is normative. Further, in scholarship that takes an empirical approach to examine this topic, the object of inquiry is typically the ramifications of such provisions – most notably the extent to which they foster social transformation. Yet, we know surprisingly little about the antecedents of affirmative action. This work examines what variables systematically predict affirmative action. We focus on the policy feedback literature and compensatory justice frameworks to examine the effects of democracy, modernisation and globalisation on affirmative action programmes. Time-series cross-sectional analyses of data for hundreds of groups from all over the globe for the period 1985–2003 confirm our hypotheses. This is the first work to examine affirmative action programmes in a large-N framework of such scale. We find that such programmes systematically correlate with democracy, modernisation and globalisation.

Performance of China’s advertising agencies: a time series cross-sectional analysis

ABSTRACTAlthough China has been the world’s second largest advertising market after the United States in terms of advertising spending since 2006, the performance of advertising agencies in China and the factors that contribute to this performance have been understudied. However, by incorporating the structure-conduct-performance model and agency theory into the integrative framework and conducting a time series cross-sectional analysis, we determine that the degree of concentration in the advertising agency industry and number of regulations in the advertising industry have had significant negative effects on the financial performance of agencies. In addition, agencies with mainly foreign capital have performed better than those with only Chinese capital. Agencies adopting strategies of initial public offering (IPOs) or engaging in name changes and mergers have performed better than those that have done nothing. Implications are also discussed.

Effects of pitch and timing expectancy on musical emotion.

Pitch and timing information work hand in hand to create a coherent piece of music; but what happens when this information goes against the norm? Relationships between musical expectancy and emotional responses were investigated in a study conducted with 40 participants: 20 musicians and 20 non-musicians. Participants took part in one of two behavioural paradigms measuring continuous expectancy or emotional responses (arousal and valence) while listening to folk melodies that exhibited either high or low pitch predictability and high or low onset predictability. The causal influence of pitch predictability was investigated in an additional condition where pitch was artificially manipulated and a comparison conducted between original and manipulated forms; the dynamic correlative influence of pitch and timing information and its perception on emotional change during listening was evaluated using cross-sectional time series analysis. The results indicate that pitch and onset predictability are consistent predictors of perceived expectancy and emotional response, with onset carrying more weight than pitch. In addition, musicians and non-musicians do not differ in their responses, possibly due to shared cultural background and knowledge. The results demonstrate in a controlled lab-based setting a precise, quantitative relationship between the predictability of musical structure, expectation and emotional response.

Compromising on Values? Human Rights Pressure and Competing Interests of the European Union in the Former Soviet States

What influences the European Union’s (EU’s) policy towards human rights abuse in third countries? What effects do the EU’s punitive measures have on the subsequent human rights situation? In literature, there is no consensus about the effects of such an instrument; moreover, scholars often question the consistency of its application in regions where the EU has strong strategic interests. Utilizing time-series cross-section analysis of twelve former Soviet republics over two decades, the present article demonstrates that the severity of imposed measures is guided by the actual human rights situation, and the presence of competing interests does not determine their intensity. Additionally, coercive measures are shown to have a positive effect on the subsequent human rights situation. At the same time, competing interests of the EU prove able to undermine the instrument’s ex-ante credibility and, thereby, limit its potential impact.

Electoral Rules and Partisan Control of Government: A Replication Study

Does the electoral system affect government partisanship? Iversen and Soskice answer this question in the affirmative. These authors argue that center-right governments dominate in majoritarian systems, whereas proportional representation systems see more center-left governments. They explain this difference by the strategic voting behavior of the middle class under alternative electoral rules. In this study, we test the robustness of their empirical results to alternative measures of the main variables as well as to a completed version of the original data set. Our replication does not corroborate Iversen and Soskice’s empirical findings. First, we cannot substantiate the notion that center-right governments emerge more frequently in majoritarian systems. Second, a time-series cross-section analysis does not support the hypothesis that the electoral system is a significant determinant of partisan control of government.

Catastrophic drug coverage: utilization insights from the Ontario Trillium Drug Program.

Catastrophic drug coverage programs help those with high drug-costs to reduce the burden of out-of-pocket expenses. We set out to measure changes in utilization, spending and demographic profiles of people accessing Ontario's catastrophic drug program, the Trillium Drug Program. We conducted a cross-sectional time-series analysis examining quarterly utilization and spending trends among medications reimbursed by the Trillium Drug Program in Ontario, Canada from Jan. 1, 2000, to Dec. 31, 2016. In each of 2000, 2005, 2010 and 2015, we described the population of beneficiaries, including demographic information, health care utilization and medication utilization. Over our study period, use of the Trillium Drug Program increased threefold from 3.6 beneficiaries per 1000 to 10.9 beneficiaries per 1000 Ontarians, and total government spending on the program increased by over 700%, reaching $487 million in 2016. Between 2000 and 2015, there was an increase in the number of beneficiaries who were under the age of 35 years (19.6% to 25.3%; p < 0.0001), did not have a hospital admission (68.3% to 80.5%; p < 0.0001) and had medium to high deductibles (2.3% to 8.0%; p < 0.0001). Further, there was a large increase in the percentage of users with drug claims greater than $1000 (3.4% to 10.4%; p < 0.0001) and those dispensed a high-cost biologic drug (1.6% to 5.5%; p < 0.0001). Increasing use of Ontario's catastrophic drug program highlights the growing burden of high drug prices for Canadians. With a growing number of expensive drugs being approved in Canada, we anticipate that spending and use of the catastrophic drug program will continue to expand.