Regression Kink Design Research Articles

Excellence for All? University Honors Programs and Human Capital Formation

Can public university Honors programs deliver the benefits of selective undergraduate education within otherwise nonselective institutions? We evaluate the impact of admission to the Honors College at Oregon State University, a large nonselective public university. Admission to the Honors College depends heavily on a numerical application score. Nonlinearities in admissions probabilities as a function of this score allow us to compare applicants with similar scores, but different admissions outcomes, via a fuzzy regression kink design. The first stage is strong, with take-up of Honors College programming closely following nonlinearities in admissions probabilities. To estimate the causal effect of Honors College admission on human capital formation, we use these nonlinearities in the admissions function as instruments, combined with course-section fixed effects to account for strategic course selection. Honors College admission increases course grades by 0.10 grade points on the 0 to 4 scale, or 0.14 standard deviations. Effects are concentrated at the top of the course grade distribution. However, course grades of first-generation students decrease in response to Honors admission, driven by low performance in natural science courses. Results suggest that selective Honors programs can accelerate skill acquisition for high-achieving students at public universities, but not all students benefit from Honors admission.

The impact of parental benefits on disadvantaged households

AbstractOver the past 25 years, the Government of Quebec (Canada) has introduced a number of relatively novel policies aimed at fighting poverty such as the Universal Child Care Program (UCCP) in 1997 and the Quebec Parental Insurance Program (QPIP) in 2006. Since its inception, the QPIP has provided a means‐tested supplementary benefits scheme for disadvantaged households. The scheme yields a well‐defined kink in the benefits schedule with respect to two entirely exogenous criteria. Using the QPIP administrative data files from 2006 to 2017, we estimate the causal impact of the supplemental benefits on leave duration and participation of poor households within a sharp Regression Kink Design (RKD) approach. Our results indicate that single mothers are relatively responsive to additional benefits. Conversely, partnered mothers are not found to respond to the supplemental benefits, irrespective of fathers’ own participation in the parental leave. The Canadian government is currently considering introducing a similar parental leave program. Our results may prove useful for the design of the program.

The direct and spillover effects of diabetes diagnosis on lifestyle behaviours.

Using blood sample data we exploit an arbitrary cut-off of diabetes risk and through a fuzzy regression kink design we estimate the effect of a diabetes diagnosis on own and partner health-related behaviours. Diabetes diagnosis increases the probability of exercising, both for those diagnosed with diabetes and their partner. We also conduct mediation analysis which suggests that joint household participation is the channel behind this effect. Our results have significant implications for the understanding of the channels that induce behavioural change, and household decision making, as well as, for the evaluation of diabetes related policies.

The impact of increase in block pricing on electricity demand responsiveness: Evidence from Ghana

The implementation of Increase in Block Pricing (IBP) policy is premised on the assumption that consumers adjust their behavior to save energy in response to electricity price increases. However, consumers often lack a comprehensive understanding of the pricing schedules, leading to inaccurate consumption predictions. This study utilizes monthly electricity billing data from the administrative records of urban residential consumers of the Electricity Company of Ghana Limited to estimate the causal effect of IBP policy schedules on electricity consumption. To achieve this, we employed the Regression Kink (RK) design identification strategy to explore the kink structure of the policy function. This paper finds an inelastic causal relationship between IBP policy and electricity consumption. The results show a treatment effect of 0.41 KWh with an associated p-value of 0.473 at a robust 95 % confidence interval. Thus, IBP policy has no causal effect on electricity consumption at the 50 KWh lower block threshold. Notably, electricity consumers do not respond significantly to price increments, rendering IBP policy ineffective in achieving energy saving at the lower block threshold. This research contributes to a deeper understanding of the effectiveness of IBP policy in promoting energy savings and informs researchers, policymakers, and stakeholders in the energy sector. Results can be used to develop evidence-based policies that contribute to energy conservation and climate change mitigation.

Estimating the effect of land use regulation on land price: At the kink point of building height limits in Fukuoka

This study estimates the effect of land use regulation on land price by exploiting the feature of building height limits of the aviation law in Fukuoka, Japan. The law limits the height of a building that is within 4000 meters of an airport to 54.1 m, but when the distance exceeds 4000 meters, the limits are relaxed. Exploiting this regulation feature, we estimate the effect of the regulation on land price using the regression kink design. We find that building height restriction has a significantly negative effect on land price and the magnitude of the effects depends on the stringency of regulation.

Moral Hazard versus Liquidity in Household Bankruptcy

ABSTRACTThis paper studies the role of moral hazard and liquidity in driving household bankruptcy. First, I estimate that increases in potential debt forgiveness have a positive, but small, effect on filing using a regression kink design. Second, exploiting quasi‐experimental variation in mortgage payment reductions, I estimate that filing is five times more responsive to cash‐on‐hand than relief generosity. Using a sufficient statistic, I show the estimates imply large consumption‐smoothing benefits of bankruptcy for the marginal filer. Finally, I conclude that 83% of the filing response to dischargeable debt comes from liquidity effects rather than a moral hazard response to financial incentives.

Age at marriage and marital stability: evidence from China

Many studies showed that marrying younger is associated with a higher risk of divorce. We investigate the causal effect of marrying at an earlier age on women’s divorce risk. We exploit the introduction of the 1981 reform in China, which facilitated legal marriage for urban women younger than 25 years old, using the Chinese Census data. We show that the reform generated a kink in the mean age at marriage for women, which we use in a fuzzy regression kink design (RKD) to assess the causal effect of marrying younger on the probability of divorce. First, we confirm in our data the existence of a negative (in fact, a U-shaped) association between age at marriage and divorce, as commonly observed in previous studies from the USA. Then, we show that this association disappears in our analyses based on RKD. This finding suggests that the well-documented association between early marriage and divorce is in fact attributable to unobservable factors driving both marriage timing and the likelihood of divorce. We discuss the implications.

Price elasticities revisited: The effect of price changes caused by taxation on the amount of alcohol consumed

Price elasticity is an important indicator of alcohol consumption. The success of an alcohol policy largely depends on how sensitive the demand is to the price. The purpose of this paper is to determine the effect of vodka price changes on the amount of vodka and pure alcohol consumed by Russian citizens exploiting the variation caused by price regulations, and in particular, taxation. Demand for alcohol is sensitive to its price, as abundant research has shown. However, the price of alcohol, in turn, may depend on demand. To overcome the endogeneity, we used a sharp change in excise tax observed in Russia in 2011 to model the price of vodka using the regression kink design. Based on nationally representative households and individual survey data, we also used the Arellano–Bond method to take into account habit formation. Our results are close to those obtained in other countries but differ from price elasticity values calculated in some work based on Russian data. The results confirm the effectiveness of the tax policy in those years when it increased the real prices of alcohol.

Moral Hazard Versus Liquidity and the Optimal Timing of Unemployment Benefits

Abstract We develop a novel way of identifying the liquidity and moral hazard effects of unemployment insurance exclusively from how job-finding rates respond to unemployment benefits that vary over an unemployment spell. We derive a sufficient statistics formula for the dynamically optimal level of unemployment benefits based on these two effects. Using a regression kink design (RKD) that simultaneously exploits two kinks in the schedule of unemployment benefits, we apply our method to Spain for the years 1992–2012 and find that moral hazard effects dominated liquidity effects, suggesting that Spanish unemployment benefits exceeded the optimal level in that period.

The effects of a tax deduction for lifelong learning expenditures

We study the effects of a tax deduction for lifelong learning, exploiting exogenous variation in the effective costs of lifelong learning due to jumps in tax bracket rates. We use a regression kink design and tax return data on the universe of Dutch taxpayers. Low-income individuals show no response, but high-income individuals are more likely to report lifelong learning expenditures (though not a higher amount) when net costs are lower. Furthermore, for high-income individuals the effect peaks at the age interval 40–45 years of age.

The effect of changing mortgage payments on default and prepayment: Evidence from HAMP resets

AbstractAfter 5 years of subsidy, interest rates for mortgages modified under the Home Affordable Modification Program (HAMP) reset, increasing annually in increments of up to 1 percentage point until they reach a predetermined rate. We identify the causal effect of increasing interest rates on default, delinquency, and prepayment using an event study design and confirm the effects using a regression kink design. New delinquency and default spike at time of reset among those whose rates increase, with smaller persistent effects continuing. These delinquency and default effects are smaller for lower balance loans in the second reset, for which additional subsidies at the time of second reset are relatively more valuable. We find that a 1 percentage point increase in the mortgage rate led to a roughly 20% increase in the default hazard. This implies that the reset rate increase is cost‐effective for lenders given the default environment at the time of reset: increased revenue from higher rates exceeds estimated losses from additional defaults.

The impact of cost-sharing on prescription drug demand: evidence from a double-difference regression kink design

Pharmaceuticals represent the third-largest expenditure item in health care spending in the OECD countries, and cost growth is around 5% per year in many OECD countries. One possible way to contain the rise in pharmaceutical spending is the use of cost-sharing schemes that makes insured individuals directly bear parts of the cost of a drug. This study estimates the price sensitivity of demand for prescription drugs using data on all prescription drug purchases from a random sample of 400,000 Swedes followed from 2010 to 2013. We use a regression kink design (RKD) by exploiting the kinked Swedish cost-sharing scheme to assess the price elasticity. Further, since the cost-sharing scheme has changed over time, we also use a double-difference RKD to account for potential confounding nonlinearities around the kink. Our results indicate that the standard RKD results are biased and exaggerate the price sensitivity. Our preferred double-difference RKD specifications show no or minor price sensitivity (95% CI price elasticity from − 0.12 to 0.02). The results are similar in several sub-group analyses across age groups, sexes, and income quartiles.

Размер поддержки и продолжительность регистрируемой безработицы в России: метод регрессии с изломом

The unemployment insurance (UI) system is one of the key labor market institutes. UI benefits allow to smooth consumption during unemployment. Moreover, people receiving UI benefits can search for a new position more selectively, and thus find a job that with more appropriate qualification requirements and salary. Nevertheless, generous UI benefits lead to a decrease in the intensity of job search and an increase in the duration of registered unemployment. The optimal UI system design balances these effects. This paper is based on administrative microdata on registered unemployment in Russia in 2019 (collected by The Federal Service for Labor and Employment of Russia) to estimate the causal effect of UI benefits on the duration of registered unemployment. Regression kink design and the maximum UI benefits levels are exploited to identify the causal effect. Estimated elasticities of UI duration with respect to average daily benefits are positive and significant with values from 0.2 to 5.3 due to specification which corresponds to the estimates obtained for other countries.

Did the Hospital Readmissions Reduction Program Reduce Readmissions? An Assessment of Prior Evidence and New Estimates.

In this article, we provide a comprehensive, empirical assessment of the hypothesis that the Hospital Readmissions Reduction Program (HRRP) affected hospital readmissions. In doing so, we provide evidence as to the validity of prior empirical approaches used to evaluate the HRRP and we present results from a previously unused approach to study this research question-a regression-kink design. Results of our analysis document that the empirical approaches used in most prior research assessing the efficacy of the HRRP often lack internal validity. Therefore, results from these studies may not be informative about the causal consequences of the HRRP. Results from our regression-kink analysis, which we validate, suggest that the HRRP had little effect on hospital readmissions. This finding contrasts with the results of most prior studies, which report that the HRRP significantly reduced readmissions. Our finding is consistent with conceptual considerations related to the assumptions underlying HRRP penalty: in particular, the difficulty of identifying preventable readmissions, the highly imperfect risk adjustment that affects the penalty determination, and the absence of proven tools to reduce readmissions.

Using Nonlinear Budget Sets to Estimate Extensive Margin Responses: Method and Evidence from the Earnings Test

We estimate the impact of the Social Security Annual Earnings Test (AET) on older workers’ employment. The AET reduces social security claimants’ current benefits in proportion to their earnings in excess of an exempt amount. Using a regression kink design and Social Security Administration data, we document that the discontinuous change in the benefit reduction rate at the exempt amount causes a corresponding change in the slope of the employment rate, suggesting that the extensive margin of labor supply is more sensitive to this policy than commonly thought. We develop a model and method that allow us to translate the behavioral responses into a lower bound estimate of 0.49 for the extensive margin elasticity, which implies more than a 1 percentage point increase in work in the absence of the AET. (JEL H55, J14, J22, J26, J31)

Board change and firm risk: Do new directors mean unstable corporate policies?

AbstractResearch Question/IssueGiven the contentious debate over whether the appointment of new directors reduces firm risk, this study explores the effect of board changes on firm risk.Research Findings/InsightsAppointing new directors leads to firm risk. A regression kink design shows that boards with more than 30% of new directors experience a significant increase in firm risk for approximately two years. Moreover, the effect of new directors on firm risk is more pronounced for firms with weaker corporate governance mechanisms but is attenuated if the demographic gaps between the new and existing directors are relatively larger. Additionally, the effect is moderated when new directors have interlocking and academic experience. Further analysis reveals that the appointment of new directors is associated with less consistent corporate policies.Theoretical/Academic ImplicationsWe extend the theory of dynamic board governance, providing a qualitative and quantitative description of the effect of board change via a regression kink design, which shows that board change is an important variable that upsets the balance of board governance, leads to higher volatility of corporate policies, and increases the risk of corporate operations.Practitioner/Policy ImplicationsThe findings suggest that corporate management should carefully assess the risk of board change. A large proportion of new director involvement can create challenges in communication, understanding, and cooperation, leading to inconsistent corporate strategies and policies, thus increasing operating volatility. Moreover, this study offers insights to policymakers rethinking board spills in their countries.

The effect of unemployment benefits on health and living standards in Turkey: evidence from structural equation modelling and regression discontinuity design

PurposeUnemployment can negatively affect individuals, their families and communities in various ways. When individuals are out of work may experience mental and physical health problems, material deprivation and poverty. This study aims to examine the impact of unemployment benefits on health and living standards in Turkey.Design/methodology/approachThe authors employ a structural equation modelling (SEM) to consider the simultaneous relationships among the unemployment benefits and the latent variables of health and Standard of Living (SoL). We propose a fuzzy Regression Discontinuity Design (FRDD) and a Regression Kink Design (RKD) within the SEM framework to infer causality. For the empirical analysis, the authors employ the panel Income and Living Conditions Survey (ILCS) in 2007–2015.FindingsThe authors’ findings suggest that those who receive these benefits are more likely to report higher levels of health and improve their living standards compared to the non-recipients. Furthermore, unemployment benefits replacement rates are associated with improved levels in health and living standards. The authors’ results indicate a substantial heterogeneity on the impact of unemployment benefits since males, low educated individuals and those belonging to the lower levels of income are affected more in terms of their health status and living standards.Originality/valueThe majority of earlier studies have focused on the impact of unemployment benefits on labor outcomes. The originality of this study is that we implement the FRDD and RKD within the SEM framework to explore, simultaneously, the impact of unemployment insurance on health and living standards. Moreover, future research studies can implement this framework to infer causality and explore the impact of related policies and reforms.

The Costs and Benefits of Liquidity Regulations: Lessons from an Idle Monetary Policy Tool

Abstract We investigate how liquidity regulations affect banks by examining a dormant monetary policy tool that functions as a liquidity regulation. For causal inference, we use a regression kink design that relies on the variation in a marginal high-quality liquid asset requirement around an exogenous threshold. We show that mandated increases in liquidity cause banks to reduce credit supply. Liquidity requirements also depress banks’ profitability, though some of the regulatory costs are passed on to liability holders. We document a prudential benefit of liquidity requirements by showing that banks subject to a higher requirement just before the financial crisis had lower odds of failure.

Fostering entrepreneurship: the crucial role of monitoring schemes

PurposeThe purpose of this paper is to investigate the effect of government-led programs on the engagement of individuals in entrepreneurship.Design/methodology/approachThe authors worked with government-led programs of 16 European countries between 2003 and 2014 and were able to benefit from the 2008 natural experiment (i.e. the global financial crisis) to produce a robust investigation using a regression kink design (RKD).FindingsThe work shows that government-led programs that are designed to include monitoring schemes can significantly increase individuals' engagement in opportunity-driven entrepreneurship. The authors found that monitoring schemes do not have the same relevance for necessity-driven entrepreneurship. Therefore, the authors believe the difference occurs because monitoring design avoids problems related to moral hazard and adverse selection when it comes to individuals choosing whether to participate (or not) in government-led programs.Originality/valueWhile it is important for governments to provide an enabling environment for entrepreneurship, this study showed that not all types of public program have positive results. In fact, it has been demonstrated that poorly-designed programs can actually decrease the likelihood of individuals engaging in entrepreneurial activities. The efficiency of programs is substantially improved, however, when they are designed to include monitoring schemes.

Local Polynomial Order in Regression Discontinuity Designs

Treatment effect estimates in regression discontinuity (RD) designs are often sensitive to the choice of bandwidth and polynomial order, the two important ingredients of widely used local regression methods. While Imbens and Kalyanaraman and Calonico, Cattaneo, and Titiunik provided guidance on bandwidth, the sensitivity to polynomial order still poses a conundrum to RD practitioners. It is understood in the econometric literature that applying the argument of bias reduction does not help resolve this conundrum, since it would always lead to preferring higher orders. We therefore extend the frameworks of Imbens and Kalyanaraman and Calonico, Cattaneo, and Titiunik and use the asymptotic mean squared error of the local regression RD estimator as the criterion to guide polynomial order selection. We show in Monte Carlo simulations that the proposed order selection procedure performs well, particularly in large sample sizes typically found in empirical RD applications. This procedure extends easily to fuzzy regression discontinuity and regression kink designs.