关于双重差分法DID的32篇精选Articles专辑！

凡是搞计量经济的，都关注这个号了

邮箱：[e-mail]

所有计量经济圈方法论丛的code程序, 宏微观数据库和各种软件都放在社群里.欢迎到计量经济圈社群交流访问.

前些日，咱们圈子引荐了①“实证研究中用到的200篇文章, 社科学者常备toolkit”、②实证文章写作常用到的50篇名家经验帖, 学者必读系列、③过去10年AER上关于中国主题的Articles专辑、④AEA公布2017-19年度最受关注的十大研究话题, 给你的选题方向，受到各位学者欢迎和热议，很多博士生导师纷纷推荐给指导的学生参阅。

继上次，腾讯公司相关部门与因果推断研究小组开展了还算友好的交流后（“BATJ巨头急需大批经济学博士, 望奔走相告”），最近，阿里巴巴相关部门人员也希望在因果推断研究小组交流访问（因果推断研究小组惊动了阿里巴巴！）。经济学博士在BATJ公司有啥用呢? 难不成比IT程序员还有能耐，正如上文所讲，因果推断在将来很长一段时间里都是科技公司和社科学者使用的主流方法。我们会一如既往地在小组和社群探讨主流的因果推断方法，同时也欢迎大型科技公司与咱们学者保持更紧密的互动。

之前，咱们小组引荐了1.DID运用经典文献，强制性许可:来自对敌贸易法的证据，2.连续DID经典文献, 土豆成就了旧世界的文明，3.截面数据DID讲述, 截面做双重差分政策评估的范式，4.RDD经典文献, RDD模型有效性稳健性检验，5.事件研究法用于DID的经典文献"环境规制"论文数据和程序，6.广义DID方法运用得非常经典的JHE文献，7.DID的经典文献"强制许可"论文数据和do程序，8.传销活动对经济发展影响, AER上截面数据分析经典文，9.多期DID的经典文献big bad banks数据和do文件，10.因果推断IV方法经典文献，究竟是制度还是人力资本促进了经济的发展？，11.AER上因果关系确立, 敏感性检验, 异质性分析和跨数据使用经典文章，12.第二篇因果推断经典，工作中断对工人随后生产效率的影响？，13.密度经济学:来自柏林墙的自然实验, 最佳Econometrica论文，14.AER上以DID, DDD为识别策略的劳动和健康经济学，15.一个使用截面数据的政策评估方法, 也可以发AER，16.多期DID模型的经典文献，big bad banks讲解","，17.多期DID的经典文献big bad banks数据和do文件，18.非线性DID, 双重变换模型CIC, 分位数DID，受到博士生导师普遍欢迎，并分享给其指导的学生学习。

与合成控制法一样（关于合成控制法SCM的33篇精选Articles专辑！），DID也是当下非常流行的因果推断方法，在英文和中文顶刊中频繁出现。基于此，咱们小组引荐32篇使用双重差分法DID做实证研究的社科文章，感兴趣的学者可以在社群下载交流和讨论。下面每一篇文章都值得年轻学者在新型肺炎期间认真研习，毕竟每个个体在特殊时期都有自己的角色和相应责任。若你不能改变新型肺炎给自己带来的焦虑不安局面，那就宅在家里好好做平时没时间做的事情。

另外，有不少学者给圈子的文章打赏，对此甚是感谢（都记录在册）。

Baltagi, B. H. (2014). Panel Data and Difference-in-Differences Estimation. Encyclopedia of Health Economics. A. J. Culyer. San Diego, Elsevier: 425-433.

This article reviews some of the benefits and limitations of panel data. It also discusses fixed and random effects estimation and reviews some health economics applications that use these methods. Next, the difference-in-differences estimator, the Hausman test and the Hausman and Taylor estimation method are discussed and illustrated with empirical health applications. Dynamic panel data as well as limited dependent variable panel data models are discussed and once again illustrated with applications from health economics. This article ends by referring the reader to further extensions of these basic panel models and providing the appropriate references.

Chagas, A. L. S., et al. (2016). "A spatial difference-in-differences analysis of the impact of sugarcane production on respiratory diseases." Regional Science and Urban Economics 59: 24-36.

Sugarcane production represents around 10% of the agricultural area and 1% of GDP in Brazil, and has grown substantially in recent years. The traditional harvest method involves burning the field to facilitate access to the canes, resulting in well-documented negative effects on health. The existing studies do not consider the effects on health in the surrounding areas. This article presents a new variety of a spatial diff-in-diff model to control for the effects of sugarcane production in neighboring non-producing regions. This method is an addition to the Spatial Econometrics literature, as it includes spatial effects on treated and untreated regions, so that the effects on both producing and surrounding non-producing regions can be properly estimated. The results indicate that the effects on the producing regions are 78% larger than if the effects on the surrounding areas were ignored. Moreover, the effects on the surrounding areas, typically ignored in other studies, are relevant, and almost as large as the effects on the producing areas.

Chen, R., et al. (2018). "The causal impact of HACCP on seafood imports in the U.S.: An application of difference-in-differences within the gravity model." Food Policy 79: 166-178.

We offer a difference-in-differences (DID) approach to assess the causal effect of Hazard Analysis by Critical Control Points (HACCP), a food safety regulation, on U.S. seafood imports with a gravity model and event-specific changes. With a panel database of U.S. seafood imports from 217 partner countries from 1991 to 2006, we employ a causal framework of analysis with a “treatment group” of U.S. seafood imports and two alternative control groups (U.S. food imports outside of HACCP rules and the European Union's 15 seafood imports). Further, we assess the effects of HACCP on the intensive and extensive margins of U.S. seafood imports. Contrary to previous work, we find that HAACP implementation has no effect on the flow of U.S. seafood imports, while the estimates of the other key variables are consistent across the models seen in previous work. Thus, we find evidence that non-tariff measures like HACCP had net null effect on imports, though the distribution of imports shifted.

Clò, S. and E. Fumagalli (2019). "The effect of price regulation on energy imbalances: A Difference in Differences design." Energy Economics 81: 754-764.

Providing adequate incentives to schedule energy programs accurately is a critical feature of liberalized electricity markets, particularly those with large shares of intermittent, renewable energy resources. In this regard, two main regulatory approaches are widely adopted in Europe. The single pricing scheme rewards or penalizes market agents according to the impact of their individual imbalances on the system imbalance. The dual pricing scheme penalizes (at best does not reward) all individual energy imbalances. This study theoretically identifies and then provides supporting empirical evidence of potential inconsistencies between market agents' balancing responsibility and the economic incentives provided by these pricing rules (de facto, opportunities for arbitrage in sequential markets). The causal effect of imbalance price regulations on the volume of the energy imbalances is investigated by exploiting a quasi-experimental change in regulation in the Italian power system. A difference-in-differences design provides robust evidence that the volume of intentional imbalances significantly decreases when moving from a single to dual pricing scheme. We conclude that the economic incentives of a dual pricing scheme are better aligned with a market agent's responsibility to be balanced and worth of further consideration from a policy perspective.

Delaney, J. A. and T. D. Kearney (2015). "The impact of guaranteed tuition policies on postsecondary tuition levels: A difference-in-difference approach." Economics of Education Review 47: 80-99.

This study considers the impact of state-level guaranteed tuition programs on postsecondary tuition levels. The analytic framework argues that state-level laws requiring flat tuition rates for four years contain inflationary risk, which encourages institutions to set tuition higher than they otherwise would with annual adjustments. To empirically test this idea, this study uses a national panel dataset and a quasi-experimental difference-in-difference methodological approach, with Illinois’ Truth-in-Tuition law serving as the treatment condition. On average, institutions subject to this law increased annual tuition by approximately 26–30% and aggregate four-year tuition by approximately 6–7% in excess of the amount predicted by the trend for institutions not subject to the law. These findings are robust to multiple alternative specifications and support the idea that state-level guaranteed tuition programs encourage large institutional tuition increases. Implications of these findings for state policymakers, higher education institutional leaders, and college-age students and their families are also discussed.

Delgado, M. S. and R. J. G. M. Florax (2015). "Difference-in-differences techniques for spatial data: Local autocorrelation and spatial interaction." Economics Letters 137: 123-126.

We consider treatment effect estimation via a difference-in-difference approach for spatial data with local spatial interaction such that the potential outcome of observed units depends on their own treatment as well as on the treatment status of proximate neighbors. We show that under standard assumptions (common trend and ignorability) a straightforward spatially explicit version of the benchmark difference-in-differences regression is capable of identifying both direct and indirect treatment effects. We demonstrate the finite sample performance of our spatial estimator via Monte Carlo simulations.

Dempsey, J. A. and A. J. Plantinga (2013). "How well do urban growth boundaries contain development? Results for Oregon using a difference-in-difference estimator." Regional Science and Urban Economics 43(6): 996-1007.

Urban containment policies, including urban growth boundaries (UGBs), are a common tool used by city planners to promote compact development. We analyze how well UGBs do in containing development using fine-scale GIS data on cities in Oregon. Earlier studies on UGBs yield mixed results, with some authors finding no effects of UGBs on housing market variables and urbanization rates and others finding significant effects. A challenge in measuring these effects is that the location of the UGB is unlikely to be an exogenous determinant of a land parcel's value for development. The panel structure of our dataset allows us to estimate the UGB's effect on the probability of development using a difference-in-difference estimator applied to a narrow band of plots along each city's UGB. This estimator controls for time-invariant unobservable variables and common temporal effects among plots, thereby mitigating the potential for biased estimates due to the endogeneity of the UGB's location. We also pursue a novel approach to controlling for time-varying factors that exploits our fine-scale data. We find that UGBs contain development in many of the Oregon cities we examine, although there are some cities in which development rates are the same inside and outside of the UGB. Our results reveal that, in most cities, the effect of the UGB is small relative to pre-existing differences in development probabilities. This suggests that it may be difficult to identify UGB effects with cross-sectional data, the approach commonly taken in previous studies.

Deschacht, N. and K. Goeman (2015). "The effect of blended learning on course persistence and performance of adult learners: A difference-in-differences analysis." Computers & Education 87: 83-89.

This article examines the effect of blended learning on adult learners' academic success. Using a large administrative data set we test the impact of introducing a blended learning format within the first year of a business education curriculum on course persistence and performance. Our difference-in-difference research design minimizes the potential bias resulting from the selectivity of learners enrolled in blended programmes. We find out that blended learning improves exam results. Although we observe a negative effect on the course persistence of adult learners (increased drop-out due to blended learning), the overall effect on course pass rates remains positive. Implications for practice and follow-up studies are discussed.

Diao, M., et al. (2017). "Spatial-difference-in-differences models for impact of new mass rapid transit line on private housing values." Regional Science and Urban Economics 67: 64-77.

This study uses the opening of the new Circle Line (CCL) in Singapore as a natural experiment to test the effects of urban rail transit networks on non-landed private housing values. We use a network distance measure and a local-polynomial-regression approach to identify the CCL impact zone with discontinuity in housing price gradient between a treatment zone and a control zone. We then estimate the spatial difference-in-differences models that account for spatial autocorrelation in housing price changes in the two zones “before and after” the opening of the CCL. We find that the opening of the CCL increases housing value in the treated neighborhoods located within the 600-metre network distance from the new CCL stations by approximately 8.6%, relative to other properties in the untreated neighborhoods controlling for heterogeneities in housing attributes and local amenities, and spatial and temporal fixed effects. We find significant “anticipation” effects as early as 1 year prior to the opening of the CCL line, but the effects diminish closer to the actual opening date. The results imply that the inter-dependent spatial structure between the treated and the untreated neighborhoods, if neglected, may lead to over-estimation of the capitalization effects of the new transit lines on housing values.

Douglas, I. and D. Tan (2017). "Global airline alliances and profitability: A difference-in-difference analysis." Transportation Research Part A: Policy and Practice 103: 432-443.

Alliances and partnerships between airlines are entered into to create competitive advantage, reduce costs, and expand network reach. Three global alliance clusters emerged, with founding partners located in the major geographic regions, and often already involved in bilateral partnerships with other founders. This study examines whether the formation of global airline alliances, with its related expansion of network reach, resulted in an increase in profitability for the founding members. Employing difference-in-difference regressions, this study has found no evidence that the formation of global alliances improved the profitability of founding member airlines, or conferred an economic advantage over airlines that were not founding members. This result is robust across geographic regions, individual global alliances, and alternative event dates. The results of this research suggest that regulators should be less anxious that by sanctioning these closer relationships they are providing major carriers with opportunities for excessive profits.

Dubé, J., et al. (2014). "A spatial Difference-in-Differences estimator to evaluate the effect of change in public mass transit systems on house prices." Transportation Research Part B: Methodological 64: 24-40.

Evaluating the impact of public mass transit systems on real-estate values is an important application of the hedonic price model (HPM). Recently, a mathematical transformation of this approach has been proposed to account for the potential omission of latent spatial variables that may overestimate the impact of accessibility to mass transit systems on values. The development of a Difference-in-Differences (DID) estimator, based on the repeat-sales approach, is a move in the right direction. However, such an estimator neglects the possibility that specification of the price equation may follow a spatial autoregressive process with respect to the dependent variable. The objective of this paper is to propose a spatial Difference-in-Differences (SDID) estimator accounting for possible spatial spillover effects. Particular emphasis is placed on the development of a suitable weights matrix accounting for spatial links between observations. Finally, an empirical application of the SDID estimator based on the development of a new commuter rail transit system for the suburban agglomeration of Montréal (Canada) is presented and compared to the usual DID estimator.

Girma, S. and H. Görg (2007). "Evaluating the foreign ownership wage premium using a difference-in-differences matching approach." Journal of International Economics 72(1): 97-112.

This paper seeks to identify the causal effect of foreign acquisitions on wages of skilled and unskilled workers, using difference-in-differences propensity score matching estimators. Our results suggest that there is substantial heterogeneity in the post-acquisition wage effect depending on the nationality of the foreign acquirer and the skill group of workers. We find sizable post acquisition wage effects on skilled and unskilled wages following an acquisition by a US firm. No such impacts result from acquisitions by EU multinationals. Also we discern some positive wage effects for unskilled workers resulting from acquisitions by multinationals from the rest of the world.

Grafova, I. B., et al. (2014). "The difference-in-difference method: Assessing the selection bias in the effects of neighborhood environment on health." Economics & Human Biology 13: 20-33.

This paper uses the difference-in-difference estimation approach to explore the self-selection bias in estimating the effect of neighborhood economic environment on self-assessed health among older adults. The results indicate that there is evidence of downward bias in the conventional estimates of the effect of neighborhood economic disadvantage on self-reported health, representing a lower bound of the true effect.

Guignet, D. B. and A. L. Martinez-Cruz (2018). "The impacts of underground petroleum releases on a homeowner's decision to sell: A difference-in-differences approach." Regional Science and Urban Economics 69: 11-24.

Actual and perceived damages from environmental disamenities can disrupt a utility maximizing household's otherwise optimal decision of when to sell their home. This study examines this relatively under-investigated topic with an empirical application to petroleum releases from leaking underground storage tanks, like those commonly found at gas stations. The ubiquity and relative homogeneity of this disamenity facilitates a difference-in-differences methodology. The results reveal that the timing of home sales is impacted by leak and cleanup events at these disamenities; leading to both selling sooner and delaying a sale, depending on the event, presence of the primary exposure pathway, and the quality of the home. The implications of these results are discussed.

Ikenwilo, D. (2013). "A difference-in-differences analysis of the effect of free dental check-ups in Scotland." Social Science & Medicine 83: 10-18.

The Scottish Government introduced free NHS dental check-ups in April 2006 as a way of encouraging utilisation and improving the oral health of residents. We use data from the British Household Panel Survey (BHPS), a nationally representative data of 117761 individual respondents in the United Kingdom covering the period between 2001 and 2008 to evaluate the impact of this policy on utilisation of NHS dental check-ups in Scotland, using a difference-in-difference approach. Results show that there was a 3–4 percent increase in NHS dental check-up in Scotland, compared to the rest of the UK. Results suggest that a removal of financial barrier to dental check-ups does indeed lead to a modest increase in utilisation, and may have wider implications for the delivery of dental care in Scotland.

Ionescu-Ittu, R., et al. (2015). "A difference-in-differences approach to estimate the effect of income-supplementation on food insecurity." Preventive Medicine 70: 108-116.

Objective The Universal Child Care Benefit (UCCB) is a 2006 Canadian federal policy of income supplementation that provides parents with $100 monthly in Canadian dollars for each child aged <6years. The study main objective was to estimate the causal effect of UCCB on self-reported food insecurity overall and in vulnerable subgroups. Method The Canadian Community Health Survey (2001–2009) was used to conduct a difference-in-differences (DID) regression analysis for the effect of the UCCB on self-reported food insecurity. Respondents were ages ≥12 in families with at least one child aged <6years (UCCB-eligible, n=22,737) or a child aged 6–11 but no child <6years (control group, n=17,664). Results Over the study period 16.3% of respondents experienced some level of food insecurity. Overall, UCCB reduced the proportion of respondents reporting food insecurity by 2.4% (95% CI: −4.0%, −0.9%). There was a significantly stronger impact on food insecurity for respondents from households with yearly income below the population median (−4.3%, 95% CI: −7.2%, −1.4%) and respondents from single parent families (−5.4%, 95% CI: −10.3%, −0.6%). Conclusion We found that a relatively small monthly income supplementation results in a significant reduction in food insecurity at the population level, with larger effects in vulnerable groups.

Li, H., et al. (2012). "The effects of congestion charging on road traffic casualties: A causal analysis using difference-in-difference estimation." Accident Analysis & Prevention 49: 366-377.

This paper aims to identify the impacts of the London congestion charge on road casualties within the central London charging zone. It develops a full difference-in-difference (DID) model that is integrated with generalized linear models, such as Poisson and Negative Binomial regression models. Covariates are included in the model to adjust for factors that violate the parallel trend assumption, which is critical in the DID model. The lower Bayesian Information Criterion value suggests that the full difference-in-difference model performs well in evaluating the relationship between road accidents and the London congestion charge as well as other socio-economic factors. After adjusting for a time trend and regional effects, the results show that the introduction of the London congestion charge has a significant influence on the incidence of road casualties. The congestion charge reduces the total number of car accidents, but is associated with an increase in two wheeled vehicle accidents.

Lin, Y.-H. and Y.-c. Chang (2017). "Does mandating cumulative voting weaken controlling shareholders? A difference-in-differences approach." International Review of Law and Economics 52: 111-123.

Corporate scholars have long championed the use of mandatory cumulative voting in developing countries. Yet, in comparison to majority or plurality voting, we know very little about its effectiveness. Even though cumulative voting is allowed in most jurisdictions, in practice it is not widely used. Taiwan stands out as a unique jurisdiction which mandates cumulative voting on all companies. Therefore, Taiwan is the only jurisdiction, to the best of our knowledge, that can be used to test the causal effect of cumulative voting on director election. Taking advantage of an exogenous legal change that occurred in Taiwanese corporate law in December 2011, we use panel data on 640 publicly traded companies from 2009 to 2015 in a difference-in-differences framework to tease out the effect of cumulative voting. From 2001 to 2011, cumulative voting was the default rule, and 20 companies opted for majority voting. While directors and supervisors are elected every three years, not all companies change boards in the same year. Fixed-effect panel regression models show that in the 2012 election—about six months after the legal reform—the cumulative voting rule appears to have weakened the controlling shareholders’ control of the companies that had previously opted for majority voting. The controlling shareholders’ control in the 2013, 2014, and 2015 elections, however, did not decrease. The take-away lesson is that mandating cumulative voting may not create a long-term effect because controlling shareholders find other means to maintain influence. Policymakers should leave the governance decisions to the firm and focus on rules that could restrain private benefits of control and enhance transparency to rein in controlling shareholders.

Lindlbauer, I., et al. (2016). "Changes in technical efficiency after quality management certification: A DEA approach using difference-in-difference estimation with genetic matching in the hospital industry." European Journal of Operational Research 250(3): 1026-1036.

Hospitals in Germany have been required to have an internal quality management system (QMS) since 2000. Although formal certification of such systems is voluntary, the number of certifications has increased steadily. The most common standards in Germany are ISO 9001, which is also widely used internationally, and KTQ (Kooperation für Transparenz und Qualität im Gesundheitswesen), which was developed specifically for the German health care sector. While a large body of literature has investigated the impact of QMS certification on performance in many industries, there is only scarce evidence on the causal link between QMS certification and technical efficiency. In the present study, we seek to elucidate this relationship using administrative data from all German hospitals from 2000 through 2010 combined with information on certification. Our analysis has three steps: First, we calculated efficiency scores for each hospital using a bootstrapped data envelopment analysis. Second, we used genetic matching to ensure that any differences observed could be attributed to certification and were not due to differences in sample characteristics between the intervention and control groups. Third, we employed a difference-in-difference specification within a truncated regression to examine whether certification had an impact on hospital efficiency. To shed light on a potential time lag between certification and efficiency gains, we used various periods for comparison. Our results indicate that hospital efficiency was negatively related to ISO 9001 certification and positively related to KTQ certification. Moreover, coefficients were always larger in the period between first certification and recertification.

Miyawaki, A., et al. (2017). "Impact of medical subsidy disqualification on children's healthcare utilization: A difference-in-differences analysis from Japan." Social Science & Medicine 191: 89-98.

Financial support for children's medical expenses has been introduced in many countries. Limited work has been done on price elasticity in children's healthcare demand, especially in countries other than the United States. Moreover, it remains unclear how the effects of a change in the cost sharing rate on healthcare demand would differ by medical condition. We investigated the impact of an increase in the cost sharing rate on medical service utilization among school children as a whole and for each of nine common conditions, applying a difference-in-differences approach. The study period ranged from April 1, 2012, to March 30, 2014. Participants were elementary school children in an urban area who were eligible for National Health Insurance (a community-based public insurance) during the study period and who were enrolled in the 2nd, 3rd, or 4th grade in April 2013. We collected observations from 2896 persons and 69,504 (2896 × 24 months) person-months. When elementary school children were promoted to the 4th grade, they became disqualified for a municipal medical subsidy. The control group was the children promoted to the 2nd or the 3rd grade, who remained eligible for the subsidy. All data were obtained from health insurance claims. We identified the nine most common medical conditions among the subject children, and stratified the analyses by the condition diagnosed. We found that an increase in the cost sharing rate reduced outpatient service utilization as a whole. Also, we observed an increase in inpatient service utilization, not because of worsened health conditions, but rather due to substitution of inpatient service for outpatient service. The reductions in outpatient service were heterogeneous across medical conditions; declines were sharper for mild or chronic conditions. These findings may help to characterize how a change in cost sharing rate affects health outcomes in children.

Nykiforuk, C. I. J., et al. (2019). "Advocacy coalition impacts on healthy public policy-oriented learning in Alberta, Canada (2009–2016): A difference-in-differences analysis." Social Science & Medicine 220: 31-40.

Since 2009, the Alberta Policy Coalition for Chronic Disease Prevention (APCCP) has pursued policy, systems, and environmental change strategies engaging policy elites to promote healthy public policy for chronic disease prevention in Alberta, Canada. Employing Advocacy Coalition Framework (ACF) vocabulary to facilitate our analysis, we examined whether concerted advocacy by the APCCP shifted elites' belief system structures over an eight year period compared to the general public as a baseline, by fostering healthy public policy-oriented learning. As data for the study, we employed a trend design series of cross-sectional Chronic Disease Prevention Surveys in Alberta, Canada between 2009 and 2016, comparing policy elite responses in 2009 (n = 183) and 2016 (n = 174) with general public responses in 2010 (n = 1203) and 2016 (n = 1200). Drawing on four scales developed in a published exploratory factor analysis, we examined changes in elite versus public beliefs with respect to (i) behavioral etiology, (ii) socio-ecological etiology, (iii) individual responsibility, and (iv) societal responsibility. Each scale was analyzed for reliability using Cohen's alpha (α), tested for sample mean (μ) value differences with analysis of variance (ANOVA) (p < .05), and compared between groups over time using difference-in-differences analysis. Cohen's alphas above approximately 0.700 indicated acceptable scale reliability (0.692≤α ≤ 0.879). ANOVA testing indicated significant group mean difference for every scale but societal responsibility among elites (μ2009 = 13.2, μ2016 = 13.7; p = .06). Standardized beta coefficients (β) presented significant differences between elites and the public for three of four scales, excepting behavioral etiology (β = −0.009, p = .746). In ACF terms, transformation of elites' policy core beliefs is necessary, but not sufficient, for major policy change such as healthy public policy. Spanning provincial policy communities relevant to whole-of-government intervention for chronic disease prevention, our results provide evidence to support the plausibility of long term socio-ecological strategies aiming to foster policy-oriented learning among elites by advocacy coalitions like the APCCP.

Puhani, P. A. (2012). "The treatment effect, the cross difference, and the interaction term in nonlinear “difference-in-differences” models." Economics Letters 115(1): 85-87.

In any nonlinear “difference-in-differences” model with strictly monotonic transformation function, the treatment effect is the cross difference of the observed outcome minus the cross difference of the potential non-treatment outcome, which equals the incremental effect of the interaction term coefficient.

Ruhose, J. and G. Schwerdt (2016). "Does early educational tracking increase migrant-native achievement gaps? Differences-in-differences evidence across countries." Economics of Education Review 52: 134-154.

We study whether early tracking of students based on ability increases migrant-native achievement gaps. To eliminate confounding impacts of unobserved country traits, we employ a differences-in-differences strategy that exploits international variation in the age of tracking as well as student achievement before and after potential tracking. Based on pooled data from 12 large-scale international student assessments, we show that cross-sectional estimates are likely to be downward-biased. Our differences-in-differences estimates suggest that early tracking does not significantly affect overall migrant-native achievement gaps, but we find evidence for a detrimental impact for less integrated migrants.

Salinas, P. and A. Solé-Ollé (2018). "Partial fiscal decentralization reforms and educational outcomes: A difference-in-differences analysis for Spain." Journal of Urban Economics 107: 31-46.

Several arguments derived from fiscal federalism theory suggest that decentralization may improve the provision of public goods and services. However, theory remains inconclusive regarding these effects under partial decentralization. The aim of this study is to examine this hypothesis by evaluating the effects on educational outcomes of the partial fiscal decentralization reform that took place in Spain during the 1980s. Since education competences were devolved to the regions at different points in time, we can consistently estimate the effects of this reform by applying the differences-in-differences method and by using the non-decentralized regions as the comparison group. We find that the reform had a sizeable impact on the percentage of students dropping out early from school. The effects are much stronger for regions with a high level of revenues. We also find that the effects are concentrated in the high-school program and that the reform was not able to improve educational outcomes in the vocational program. We interpret these results as evidence that decentralization improved the match between education policy and population preferences.

Scott, A. and J. Yong (2015). "Do new workforce roles reduce waiting times in ED? A difference-in-difference evaluation using hospital administrative data." Health Policy 119(4): 488-493.

This paper evaluates the effect of introducing two new workforce roles under a pilot program conducted in Victoria, Australia. The trial took place at a regional hospital's emergency department (ED) between 1 July 2008 and 30 June 2009. The evaluation is based on three outcome measures: waiting time (in minutes) at ED before treatment; proportion of presentations with waiting time on target; and length of stay (in days), for ED presentations that led to in-patient admissions. The technique of difference-in-differences analysis is used. A total of 142,980 patient records from the pilot hospital and three comparison hospitals were extracted from the Victorian Emergency Minimum Dataset (VEMD). Further, 21,925 records of patients whose ED presentations led to in-patient admissions were extracted from the Victorian Admitted Episodes Dataset (VAED). The evaluation finds the piloted roles have lowered waiting time and raised the proportion of on-target presentations. These effects were found to be the strongest for less urgent triage categories. However, the evidence on in-patient length of stay was mixed. The results provide positive evidence that new workforce roles can be effective in improving the efficiency of emergency care delivery.

Slaughter, M. J. (2001). "Trade liberalization and per capita income convergence: a difference-in-differences analysis." Journal of International Economics 55(1): 203-228.

In this article I analyze whether trade liberalization contributes to per capita income convergence across countries. The analysis focuses on four post-1945 multilateral trade liberalizations. To identify trade’s effect on income dispersion, in each case I use a ‘difference in differences’ approach which compares the convergence pattern among the liberalizing countries before and after liberalization with the convergence pattern among control countries, chosen using a variety of methods, before and after liberalization. My main empirical result is I find no strong, systematic link between trade liberalization and convergence. In fact, much evidence suggests trade liberalization diverges incomes among liberalizers.

Solé-Ollé, A. and P. Sorribas-Navarro (2008). "The effects of partisan alignment on the allocation of intergovernmental transfers. Differences-in-differences estimates for Spain." Journal of Public Economics 92(12): 2302-2319.

In this paper we test the hypothesis that municipalities aligned with upper-tier grantor governments (i.e., controlled by the same party) will receive more grants than those that are unaligned. We use a rich Spanish database, which provides information on grants received by nearly 900 municipalities during the period 1993–2003 from three different upper-tier governments (i.e., Central, Regional and Upper-local). Since three elections were held at each tier during this period, we have enough within-municipality variation in partisan alignment to provide differences-in-differences estimates of the effects of alignment on the amount of grants coming from each source. Moreover, the fact that a municipality may simultaneously receive grants from aligned and unaligned grantors allows us to use a triple-differences estimator, which consists of estimating the effects of changing alignment status on the change in grants coming from the aligned grantors relative to the change in grants coming from the unaligned ones. The results suggest that partisan alignment has a sizeable positive effect on the amount of grants received by municipalities. For example, with majority governments at the two layers, aligned municipalities receive over 40% more grants than unaligned ones.

Srulovici, E., et al. (2019). "Effectiveness of Managing Diabetes During Ramadan Conversation Map intervention: A difference-in-differences (self-comparison) design." International Journal of Nursing Studies 95: 65-72.

Background Some individuals with diabetes fast during Ramadan despite medical concerns for risk of adverse outcomes. The Managing Diabetes During Ramadan Conversation Map is a self-management education group-based intervention for Muslim individuals with type 2 diabetes, specifically addressing diabetes management during Ramadan. Objective The aim of this study was to evaluate the effectiveness of the Managing Diabetes During Ramadan Conversation Map intervention in improving short-term clinical outcomes and reducing healthcare utilization following Ramadan. Design This was a retrospective rolling cohort study. Settings Participants were Clalit Health Services members with type 2 diabetes who participated in the intervention between 2014 and 2017 across Israel. Participants This study included 1732 participants who enrolled in the intervention over the five-year study period. The cohort was mainly between the ages of 45 and 74 years (83.3%), female (71.9%), of lower socioeconomic status (92.1%), with a diabetes duration of 10 years or more (51.7%), obese (64.0%), and had never smoked (73.8%). Methods The data used in this study came from Clalit Health Services’ electronic health records, which are integrated in a central data warehouse. We used a difference-in-differences (self-comparison) design to examine the effect of the intervention on changes in laboratory results and healthcare utilization over a six month baseline and follow-up. Mixed model linear regressions and Poisson regressions were used to estimate continuous and count outcomes, respectively. Results Post intervention, participants experienced a reduction of 8.61 mg/dL in their glucose levels (p = 0.005) and 0.34% in their HbA1c levels (p < 0.001). In a sub-group analysis of participants with HbA1c > 7%, larger reductions in glucose (17.02 mg/dL [p < 0.001]) and HbA1c (0.63% [p < 0.001]) levels were recorded. This sub-group also experienced a reduction of 4.83 mg/dL in LDL level (p = 0.007) and had 0.2 fewer primary care visits (p < 0.001). Conclusions Participation in the Managing Diabetes During Ramadan Conversation Map improved patient glucose and HbA1c levels. A greater benefit was reported in those individuals with HbA1c > 7%. These findings hold important global health implications for the millions of individuals with type 2 diabetes for whom Ramadan can pose a challenge in disease control.

Sunak, Y. and R. Madlener (2016). "The impact of wind farm visibility on property values: A spatial difference-in-differences analysis." Energy Economics 55: 79-91.

Today's investment decisions in large-scale onshore wind projects in Germany are no longer determined only by the investment's economic benefit, but also by concerns associated to social acceptance. Despite a mostly positive attitude towards the expansion of wind power, local public concerns often stem from the belief that the proximity to large-scale wind farms may lead to a decrease in property prices. In particular, the change in landscape caused by the construction of a wind farm may have an adverse impact on the view from some properties, and thus may negatively affect their price. To investigate the potential devaluation of properties in Germany due to wind farms, we use a quasi-experimental technique and apply a spatial difference-in-differences approach to various wind farm sites in the federal state of North Rhine-Westphalia. We adopt a quantitative visual impact assessment approach to account for the adverse environmental effects caused by the wind turbines. To properly account for spatial dependence and unobserved variables biases, we apply augmented spatial econometric models. The estimates indicate that the asking price for properties whose view was strongly affected by the construction of wind turbines decreased by about 9–14%. In contrast, properties with a minor or marginal view on the wind turbines experienced no devaluation.

Vermeulen, M. J., et al. (2016). "The Effect of Pay for Performance in the Emergency Department on Patient Waiting Times and Quality of Care in Ontario, Canada: A Difference-in-Differences Analysis." Annals of Emergency Medicine 67(4): 496-505.e497.

Study objective In 2008, a pay-for-performance program was implemented in sequential waves in Ontario emergency departments (EDs), with the aim of reducing length of stay. We seek to evaluate its effects on ED length of stay and quality of care. Methods This was a retrospective observational study of ED visits in Ontario from April 1, 2007, to March 31, 2011, using multivariable difference-in-differences analysis. Pay-for-performance hospitals and matched control sites were selected for each of 3 waves of the program. The primary outcome was 90th percentile ED length of stay; we also examined quality-of-care indicators. Results Pay-for-performance hospitals had a modest reduction in overall adjusted 90th percentile ED length of stay in wave 1 (–36 minutes; 95% confidence interval [CI] –50 to –21 minutes), but not in wave 2 (–14 minutes; 95% CI –30 to 2 minutes) or wave 3 (–7 minutes; 95% CI –23 to 8 minutes). ED admitted patients had a pronounced reduction in adjusted 90th percentile length of stay in wave 1 (–225 minutes; 95% CI –263 to –188 minutes) and wave 2 (–133 minutes; 95% CI –175 to –91 minutes). Nonadmitted low-acuity patients had reductions in adjusted 90th percentile ED length of stay in wave 1 (–24 minutes; 95% CI –29 to –18 minutes) and wave 3 (–19 minutes; 95% CI –24 to –14 minutes). The program did not negatively affect ED quality-of-care measures, such as 30-day mortality or readmission rates. Conclusion Pay-for-performance was associated with modest overall benefits for ED length of stay without adversely affecting quality of care.

Walsh, B., et al. (2019). "Did the expansion of free GP care impact demand for Emergency Department attendances? A difference-in-differences analysis." Social Science & Medicine 222: 101-111.

The removal of co-payments for General Practitioner (GP) services has been shown to increase utilisation of GP care. The introduction of free GP care may also have spillover effects on utilisation of other healthcare such as Emergency Department (ED) services, which often serve as substitutes for primary care, and where co-payments to attend exist for many. In Ireland, out-of-pocket payments are paid by the majority of the population to access GP care, and these costs are amongst the highest in Europe. However, in July 2015 all children in Ireland aged under 6 became eligible for free GP care. Using a large administrative dataset on 413,562 ED attendances between January 2015 and June 2016 we apply a difference-in-differences method, with treatment and control groups differentiated by age, to examine whether ED utilisation changed amongst younger children following the introduction of universal free GP care. In particular, we examine ED attendances following a GP referral, as referrals from GPs also afford access to the ED free of charge. We find that the expansion of free GP care did not reduce overall ED utilisation for under 6s. Additionally, we find that the proportion of ED attendances occurring through GP referrals increased by over 2 percentage points. This latter finding may be indicative of increased pressure placed on GPs from increased demand. Overall, this study finds that expanding free GP care to all young children did not reduce their ED utilisation.

Weber, C. E. (2014). "Toward obtaining a consistent estimate of the elasticity of taxable income using difference-in-differences." Journal of Public Economics 117: 90-103.

The elasticity of taxable income (ETI) is a central parameter for tax policy debates. This paper shows that mean reversion prevents most estimators employed in the literature from obtaining consistent estimates of the ETI. A new method is proposed that will resolve inconsistency due to mean reversion under testable assumptions regarding the degree of serial correlation in the error term. Using this procedure, I estimate an ETI of 0.858, which is about twice as large as the estimates found in the most frequently cited paper on this subject [13]. The corresponding elasticity of broad income is 0.475.

拓展性阅读

之前，咱们圈子引荐过一些数据库，如下：1.这40个微观数据库够你博士毕业了；2.中国工业企业数据库匹配160大步骤的完整程序和相应数据；3.中国省/地级市夜间灯光数据；4.1997-2014中国市场化指数权威版本；5.1998-2016年中国地级市年均PM2.5；6.计量经济圈经济社会等数据库合集；7.中国方言,官员, 行政审批和省长数据库开放；8.2005-2015中国分省分行业CO2数据；9.国际贸易研究中的数据演进与当代问题；10.经济学研究常用中国微观数据手册。

下面这些短链接文章属于合集，可以收藏起来阅读，不然以后都找不到了。

2年，计量经济圈公众号近1000篇文章，

Econometrics Circle

数据系列：空间矩阵 | 工企数据 | PM2.5 | 市场化指数 | CO2数据 |  夜间灯光 | 官员方言  | 微观数据 |

计量系列：匹配方法 | 内生性 | 工具变量 | DID | 面板数据 | 常用TOOL | 中介调节 | 时间序列 | RDD断点 | 合成控制 | 

数据处理：Stata | R | Python | 缺失值 | CHIP/ CHNS/CHARLS/CFPS/CGSS等 |

干货系列：能源环境 | 效率研究 | 空间计量 | 国际经贸 | 计量软件 | 商科研究 | 机器学习 | SSCI | CSSCI | SSCI查询 |

计量经济圈组织了一个计量社群，有如下特征：热情互助最多、前沿趋势最多、社科资料最多、社科数据最多、科研牛人最多、海外名校最多。因此，建议积极进取和有强烈研习激情的中青年学者到社群交流探讨，始终坚信优秀是通过感染优秀而互相成就彼此的。