Propensity Score Matching Estimator Research Articles

The Nexus Between Financial Literacy and Saving Performance: Implications for Creating Self‐Financing Society

ABSTRACTMaking informed and effective savings decisions requires adequate knowledge and skills about saving products. This knowledge and skills are acquired through financial literacy (FL) training. Despite the existence of various local financial institutions and their efforts to encourage saving mobilization, domestic savings in many developing countries remain low due to low levels of FL. Using a nonrandomized, quasi‐experimental design method, with a sample of 328 saving and credit cooperatives (SACCOs) from Ethiopia, we investigated whether SACCOs providing FL for their education and Savings Committee members report improved saving performance than their counterparts not offering FL. We used administrative data and conducted our analysis at the institutional level. Our result from a combination of difference‐in‐difference and propensity score matching (DiD‐PSM) estimations reveals that FL training plays a decisive role in boosting SACCOs' saving performance. The DiD‐PSM result further shows, compared to controlled SACCOs, treated SACCOs have been successful in acquiring new net saver members and encouraging current members to increase their deposits. Besides, the implications of our findings for financial self‐reliance are discussed. To improve saving performance, policy initiatives promoting FL training in SACCOs are required.

Propensity score matching for estimation of pairwise marginal hazard ratios

. There is a growing interest in using observational studies to estimate the effects of treatments on survival or time-to-event outcomes. However, few standard approaches can adequately accommodate multiple treatment levels, which are common in observational comparative effectiveness research. We study the asymptotic properties of the generalized propensity score matching estimators of the marginal hazard ratios between pairs of treatment levels. The estimates are obtained by fitting a marginal Cox proportional hazard model on the matched dataset. We evaluate our approach in a simulation study and a case study where we analyze the IQVIA electronic medical records data.

Research on the impact of land rights certification on farmers’ operating behavior

This article aims to assess the impact of issuing land rights certification on farmers' operating behavior. To accomplish this objective, this paper establishes a theoretical framework for farmers' management behavior and thoroughly examines the stability, endowment, and security effects of land rights certification. Subsequently, using household survey data, the Probit and propensity score matching (PSM) estimation methods are employed to empirically substantiate the influence of farmland ownership confirmation. The verification results show that land right confirmation tends to motivate farmers to engage in agricultural management, and encourages female and older laborers to be more involved in agricultural production compared to males and younger laborers. Furthermore, the paper explores the mechanism by introducing such indicators as investment in agricultural machinery, quality of cultivated land, non-agricultural income of farmers, and laborers’ insurance status. Through analysis and testing, this paper holds that land right confirmation has a dual impact on farmer management, and when the reverse effect is limited, farmers are prompted to adjust the contradiction through labor force allocation within the family.

Does sovereignty help economic growth? A recent reassessment

There is relevant literature that has found that independence harmed economic growth in some former colonies, but there are few to no systematic and recent empirical analyses on the relationship between sovereignty and economic growth. We follow and enhance the Solow–Swan growth model to measure economic convergence with the United States. This model was estimated through feasible generalized least squares panel regressions and robust regressions. These estimations allow us to capture the experience of recent (since the 1950s) sovereignties vis-à-vis long-existing countries and the economic outcomes of democratic sovereignties. We then stratify economies by region and income level and show propensity score matching estimators of recent former colonies with other countries that share similar growth determinants. On average, our parameter estimates suggest that independence causes countries’ per capita income to converge with that of the United States. Initial democratic and economic conditions appear to be among the modifying factors between sovereignty and economic growth.

Long-term outcomes of DCB and DES for the treatment of in-stent restenosis in relation to the vessel size.

Percutaneous coronary intervention (PCI) with drug-coated balloons (DCB) or drug-eluting stents (DES) are well-established treatments for in-stent restenosis, however little is known about the impact of vessel size on the outcomes. The study aimed to evaluate the efficacy and safety profile of DCB versus DES in DES in-stent restenosis depending on the vessel size. Consecutive patients with DES in-stent restenosis who underwent PCI between January 2010 and February 2018 entered the registry with a long-term follow-up. Patients who received DCB at the index procedure were compared with those who received DES in three subgroups depending on the vessel size (≤2.5 mm; 2.5-3.5 mm; >3.5 mm). Data were analyzed using propensity score matching and Kaplan-Meier estimator plots. Among 1,374 patients with DES in-stent restenosis, 615 were treated with DES and 759 with DCB. After propensity score matching, we analyzed 752 patients in the DES and DCB groups at a long-term follow-up. The risk of DOCE did not differ significantly between the DES and DCB groups, both in the overall population (HR 0.85; 95%CI [0.58; 1.26], log-rank p = 0.41) and when divided into small (HR 0.84; 95%CI [0.36; 1.95], log-rank p = 0.70), medium-sized (HR 0.90; 95%CI [0.49; 1.65], log-rank p = 0.73), and large-sized (HR 0.81; 95%CI [0.42; 1.53], log-rank p = 0.50) coronary arteries. The incidence of all-cause death was significantly higher in the overall DES population (HR 4.03; 95%CI [2.40; 6.79], log-rank p < 0.001) and subgroup of small (HR 5.54; 95%CI [1.80; 17.02], log-rank p = 0.003), medium-sized (HR 4.37; 95%CI [1.92; 9.94], log-rank p = 0.009) and large-sized coronary arteries (HR 3.26; 95%CI [1.35; 7.86], log-rank p = 0.02). DES and DCB strategies are comparable methods of treating ISR regardless of the diameter of the treated vessel in a long-term follow-up.

Do relationships matter?—Director networks and corporate open innovation

This paper examines the impact of a firm's important social relationship, director network, on its open innovation. By matching the director network data of Chinese A-share listed companies from 2008 to 2018 with manually collected data on joint patent applications, we conduct an empirical study and explore the possible influence mechanisms. The results show that the director network has a significant positive impact on corporate open innovation. The conclusion still holds after endogeneity treatment, propensity score matching estimation, and other robustness tests. Heterogeneity analysis reveals that director networks promote more R&D cooperation among firms than firm-university and firm-research institution R&D cooperation; director networks have a more positive impact on the R&D cooperation activities of non-state-owned firms compared to state-owned firms; moreover, in terms of the type of director networks, independent director networks have a significantly higher positive impact on firms' open innovation than non-independent director networks. It is further found that director networks can facilitate firms' participation in R&D collaborative activities by alleviating firms' financing constraints and reducing transaction costs. This paper enriches the research related to social relations expanding firm boundaries. The findings provide some references for firms to further develop open innovation activities and the construction of innovation systems.

The Impact of Multimorbidity on Labor Force Participation Among the Middle-Aged and Older Working Population in the United States.

Multimorbidity, known as multiple chronic conditions (MCC), is the coexistence of two or more chronic health conditions (CHC). The near-retirement-age population with MCC is more likely to experience discontinued labor force participation (LFP). Our objective was to evaluate the impact of MCC on LFP among adults aged 50-64 and to explore heterogeneous effects between self-employed and non-self-employed workers. We constructed our sample using the Health and Retirement Study (HRS) from 1996 to 2018. We adopted an individual fixed-effect (F.E.) model and propensity score matching (PSM) to measure the impact of MCC on the probability of being employed and changes in annual work hours. 50.5% of respondents have MCC. Individuals with MCC exhibit a predicted probability of being employed that is 9.3 percentage points (p < .01, 95% confidence interval [95% CI]: -0.109, -0.078) lower than those without MCC. Compared with non-CHC, MCC significantly reduced annual working hours by 6.1% (p < .01, 95% CI: -0.091, -0.036) in the F.E. model and by 4.9% (p < .01, 95% CI: -0.064, -0.033) in PSM estimation. The effect is more pronounced for the self-employed with MCC, who have 13.0% (p < .05, 95% CI: -0.233, -0.026) fewer annual work hours than non-CHC based on the FE model and 13.4% (p < .01, 95% CI: -0.197, -0.070) in PSM estimation. MCC significantly reduces LFP compared with non-MCC. MCC has a heterogeneous impact across occupational types. It is important to support the near-retirement-age working population with multimorbidity through effective clinical interventions and workplace wellness policies to help manage health conditions and remain active in the labor market.

Propensity score matching for estimating a marginal hazard ratio.

Propensity score matching is commonly used to draw causal inference from observational survival data. However, its asymptotic properties have yet to be established, and variance estimation is still open to debate. We derive the statistical properties of the propensity score matching estimator of the marginal causal hazard ratio based on matching with replacement and a fixed number of matches. We also propose a double-resampling technique for variance estimation that takes into account the uncertainty due to propensity score estimation prior to matching.

The Effects of Spousal Caregiving on Middle-Age and Older Caregivers' Health and Well-Being: Evidence From Vietnam.

Background: Population aging is escalating globally, intensifying the demand for long-term care (LTC), primarily met by informal caregivers, notably spouses. Evidence from developed countries suggests potential adverse effects on caregivers' well-being. Yet, research on this topic is scarce in developing nations. We investigate the effect of informal caregiving on older spousal caregivers' health and well-being in Vietnam, a rapidly aging country with an early stage of LTC system development. Methods: Utilizing the national survey on aging in Vietnam with propensity score matching estimations to mitigate potential endogenous problems of the decision to provide care between caregivers and non-caregivers. Results: Findings showed caregiving increased poor psychological well-being, life dissatisfaction, and functional limitations by 7.3%, 9.7%, and 8.6%, respectively. The caregiving effects are heterogenous by demographic characteristics. Conclusions: We are the first to examine spousal caregiving in Vietnam, highlighting the urgency of addressing its negative impacts and suggesting several potential policy interventions.

Performance Gap between Best-bet System of Rice Intensification and Farmers’ Practices in Zamfara State, Nigeria: An Additive-Multiplicative Translog Stochastic Frontier Profit Function

Inadequate information about the performance of a technology can hinder its uptake. The aim of this paper was to provide a thorough examination of the performance gap between the best bet system of rice intensification (BB-SRI) and farmers’ practices (FP) in Zamfara State, Nigeria. Using a multi-stage sampling technique, a sample of 300 rice farmers were selected from the Bakolori Irrigation Scheme, Zamfara State, among which 40 farmers (13%) used the BB-SRI while the remaining 260 farmers (87%) used FP. The data were analyzed using descriptive statistics, farm budget technique, additive-multiplicative stochastic frontier translog profit function and propensity score matching (PSM) estimator. The findings revealed that yield increased from 5,285 kg/ha with FP to 12,735 kg/ha with BB-SRI, an increase of 141%. Rice production was profitable under both FP and BB-SRI with return on investment (ROI) increasing from 2.64 using FP to 6.66 under BB-SRI, an increase of 152%. There was evidence of profit inefficiency in rice farming since the profit could still be raised by 7% and 21% using FP and BB-SRI, respectively. The yield, ROI and profit efficiency gaps between the BB-SRI and FP ranged from 7,452 kg/ha to 7,510 kg/ha, 4.02 to 4.05, and from 0.08 to 0.09, respectively. We recommend therefore that important investments and training in SRI should be considered as a top priority for the transformation and sustainability of the rice sector in order to ensure its successful promotion, uptake and diffusion across other rice producing States in Nigeria.

Effect of chronic lung diseases on angina pectoris among Indian adults: longitudinal ageing study in India

The study aimed to evaluate the effect of chronic lung diseases, namely chronic obstructive pulmonary diseases (COPD) and asthma, on angina pectoris in individuals aged 45 years and above. Identifying vulnerable subpopulations suffering from COPD and asthma at higher risk of future cardiovascular events using the rose angina questionnaire is imperative for tailored primary and secondary prevention approaches. The present study utilizes the data from the Longitudinal Ageing Study in India, wave 1, conducted during 2017–2018. The sample size included 58,830 individuals aged 45 years and above. Angina was measured based on seven questions from Rose's questionnaires. Descriptive statistics and bivariate analysis were employed to examine the prevalence of angina among individuals with COPD and asthma. Further, multivariable logistic regression and propensity score matching (PSM) methods were used to assess the independent effect of COPD and asthma on angina after controlling the selected background characteristics. We employed PSM in two different models and included various additional factors in model 2, such as smoking, chewing tobacco, alcohol use, ADL, IADL, body mass index, physical activity, high cholesterol, hypertension, diabetes, and chronic heart disease. The current study shows that the prevalence of angina pectoris, COPD and asthma was 6.0%, 2.3% and 4.7%, respectively, among individuals aged 45 years and above in India. The prevalence of angina pectoris was higher among individuals with COPD (9.6% vs. 5.8%) and asthma (9.9% vs. 5.7%) than those without COPD and asthma, respectively. Additionally, angina pectoris was more prevalent among females and rural respondents with COPD (10.8% and 11.0%) and asthma (10.3% and 10.3%) compared to males and urban respondents with COPD (8.0% and 5.7%) and asthma (8.9% and 7.9%). Moreover, in the adjusted model, individuals with COPD (AOR 1.43, 95% CI 1.03 1.98) and asthma (AOR 1.44, 95% CI 1.17 1.77) had nearly 1.5-fold higher odds of having angina pectoris than those without COPD and asthma. The PSM estimates showed that individuals with COPD had 8.4% and 5.0% increased risk of angina pectoris compared to those without COPD in model 1 and model 2, respectively. We observed that, after adjusting to lifestyle, health-related and morbidity factors in model 2, both average treatment effect on untreated (ATU) and average treatment effect (ATE) values decreased by nearly 3.5%. Additionally, the PSM estimates demonstrated that individuals with asthma had a 3.4% and 2.9% increased risk of angina pectoris compared to those without asthma in model 1 and model 2, respectively. The study suggests that COPD and asthma are significantly associated with angina pectoris, and individuals with COPD and asthma have a higher risk of developing angina pectoris. Additionally, angina pectoris was more prevalent among females, rural respondents and adults aged 45–54 with COPD compared to males, urban respondents and those aged 65 and above, respectively, with COPD. Moreover, the findings of our study underscore the targeted primary and secondary interventions and team-based care approach among individuals with COPD and asthma to reduce the risk of CVD events in future.

Climate smart agriculture impact on food and nutrition security in Ethiopia

IntroductionThe objective of this study was to analyze how innovations in Climate-smart agriculture (CSA) contribute to improving the food and nutrition security of smallholder households, thereby promoting sustainable food systems.MethodsA cross-sectional household survey was conducted among a multi-stage sample of 424 smallholder farmers drawn from five different agroecosystems. To examine households’ food and nutrition security, we used food consumption score (FCS) and modified household dietary diversity score (HDDS) in propensity score matching (PSM) and endogenous switching regression (ESR) estimation models.ResultsPSM results showed that crop residue management, compost, and agroforestry have a significant effect on improving households’ food and nutrition security by 21.3, 13.6, and 16.6%, respectively, whereas Soil and water conservation (SWC) has reduced adopters’ food security by 12.9%. However, the conditional average treatment effect, or ESR result, reveals that households’ food and nutrition security has improved as a result of the adoption of crop residue management, compost, SWC, and agroforestry.DiscussionAlthough the impact of crop residue management, compost, and agroforestry is positive, the effect of SWC on household food security has been inconclusive. Hence, it is important to upscale the adoption of multiple CSA innovations to improve smallholder household’s food security in the face of climate change.

Evaluating disparity of subjective cognitive decline between male veterans and non-veterans in the United States using propensity score matching estimation: A behavioral risk factor surveillance system survey cross-sectional study.

Subjective cognitive decline (SCD) is one of the most important early onset symptoms of Alzheimer's Disease. Previous studies consistently reported that SCD is associated with quality of life, socio-economic factors, and related health comorbidities. However, the impact of veteran status on SCD has been little investigated. This study conducted a cross-sectional study to address disproportionate effects in subjective cognitive decline between veterans and non-veterans in the United States. Propensity score matching (PSM) was applied in this observational study to achieve covariate balancing and reduce selection bias, providing a more accurate estimate of the isolated effect of veteran status on SCD. Our study utilized 32,431 forty-five years or older non-institutionalized White, Black or African-American, and Hispanic or Latin-American male population from the 2019 Behavioral Risk Factor Surveillance System data. We first identified 10,685 paired PSM samples for the binary veteran status using the preselected covariates. Next, we performed a logistic regression for modeling the relationship between the veteran status and the SCD status using the PSM samples along with the covariates selected by a BIC-based stepwise selection. Our analyses revealed a statistically significant causal association between veteran status and SCD after PSM (odds ratio (OR): 1.16 and 95% confidence interval (CI): 1.06-1.27). We obtained a similar result before PSM with an OR of 1.20 and 95% CI of 1.10-1.31. When we focused on a minority group (Black or African-American males), we found a significantly increased veterans' risk of SCD, especially after propensity score matching (OR: 1.69, 95% CI: 1.16-2.45). We also found several factors such as employment status, difficulty dressing/walking/running errands, general health status, physical health status, unaffordability of medical costs, mental health status, and comorbid conditions including stroke, blindness, high cholesterol, and arthritis as statistically significantly associated with SCD (P-value < 0.05). Similar to post-traumatic stress disorder and traumatic brain injury, our study demonstrated a causal association between SCD and military-related activities in the United States, which has a disproportionate impact on the minority population. This study sets the groundwork to further research in this domain to diagnose neurological diseases early among veterans.

The effect of weather index insurance on social capital: Evidence from rural Ethiopia

AbstractWe study the effect of weather index insurance (WII) uptake on social capital. We measure individual social capital using experimental and survey‐based measures and relate it to the actual purchase of WII. We use propensity score matching (PSM) and an instrumental variable (IV) to address endogeneity concerns. Our descriptive and PSM estimates show that WII uptake negatively and significantly affects social capital. We find that insured households contribute less to the public good than uninsured households. We also find qualitatively and quantitatively similar results using our IV approach, yet insufficient power renders most of our IV estimates insignificant. We however report robust significant effects of instrumented WII uptake on sociopsychological outcomes—WII uptake increases perceptions of self‐sufficiency and free‐riding behavior: these are potential channels through which a negative effect on social capital comes about. Although far from conclusive, our paper provides several pieces of evidence that suggest WII uptake negatively affects social capital.

International labour migration, farmland fallowing, livelihood diversification and technology adoption in Nepal

This article investigates the effect of temporary international labour migration on farmland fallowing, adoption of agricultural intensification technologies and livelihood diversification. Using nationally representative data, combined with empirical methods that allow causal inference, the authors find that households with international migrants are over 50 per cent more likely (based on propensity score matching estimates) to have fallow land than those without. Temporary international migration promotes the adoption of some agricultural intensification technologies and causes rural households to diversify their livelihoods. Land fallowing may increase food insecurity, while agricultural intensification may improve it, for an uncertain net effect.

Goodhart’s law in China: Bank branching regulation and window dressing

After the removal of geographic restrictions on branching in 2006, China’s city commercial banks (CCBs) can apply for permission to branch outside their province. This paper shows that CCBs report a higher provision coverage ratio (PCR) before filing an application, thereby making the bank look safer to regulators. Our finding is robust to controlling for possible endogeneity of the branching application decision by employing propensity score matching estimators, and it is confirmed when we consider a quasi-natural experiment of deregulation reversal. Tests of the dynamic effects show evidence of reversals in PCR adjustment after applications. Higher PCR before branching applications cannot be explained by alternative rationales for manipulating loan loss reserves such as fundamental provisions, earnings management, capital management, and market signaling. Window dressers receive more supervisory penalties after filing applications relative to other branching banks. Our finding of window dressing in response to bank branching regulation confirms Goodhart’s insight that when a regulatory metric becomes a target, it ceases to be a good measure.

The shift in ideal family size: Examining the impact of marriage and parenthood

Motivated by role theory and the concept of the “self-fulfilling prophecy”, this study examines how entries into marriage and parenthood affect the fertility ideals of men and women in China. To address this question, we apply propensity score matching estimators to data from the China Family Panel Studies (CFPS) within a counterfactual framework. We find that marriage and parenthood have a positive effect on fertility ideals, and this result remains robust when we apply difference-in-differences matching estimates. In terms of gender asymmetry in marriage, men and women respond differently to marriage and parenthood. Entry into marriage has a positive effect on men's fertility ideals, while parenthood influences women's ideals. Beyond the average treatment effect, we further examine heterogeneous treatment effects as a function of estimated propensity scores. The study reveals that the life transition treatment effect is the highest among those with a moderate propensity for entering into marriage and parenthood.

Can internet adoption promote the productivity of firms in China: evidence from a longitudinal study

PurposeThis paper aims to conduct research to examine the impact of Internet adoption on the productivity of firms in non-urban China.Design/methodology/approachThe study investigates the impact of Internet adoption on firms' productivity in non-urban China. More specifically, the authors conduct a comprehensive and rigorous study while addressing concerns related to firm-level endogeneity by utilizing firm-level panel data. Information on firms in non-urban areas is collected from China's Annual Surveys of Industrial Firm data. For robustness, the authors implement the instrumental variables approach and propensity score matching estimations to strengthen the evidence for suggestive causal inference. Furthermore, the authors also examine the mechanisms and group heterogeneity.FindingsEvidence indicate that the adoption of Internet technology positively impacts the total factor productivity (TFP) of firms in non-urban areas. According to the heterogeneity analysis, the marginal effect of Internet adoption is more significant and pronounced for labor-intensive, private and small-scale manufacturing firms. Moreover, additional evidence suggest that Internet adoption is beneficial for non-urban firms in expanding their business and enlarging their market. It has also been found that the positive effect of Internet adoption on firms' TFP is amplified by expanding public infrastructure.Originality/valueThe current study supports that the informatization strategy benefits non-urban firms and promotes rural revitalization. The findings suggest the possibility of firms borrowing market size from the closest cities and supporting the ongoing policies of investing in broadband infrastructure to narrow the urban-rural digital gap in China.

Firm investment steering through state-level policy uncertainty

Does state-level policy uncertainty (SPU) stifle firm investment? We document fresh evidence of a strong negative association between SPU and firms’ capital spending. This finding remains robust after excluding states with overrepresentation, through alternative measures of investment and SPU, as well as during low- and high-SPU periods. Furthermore, this effect is stronger for firms facing low market competition. Interestingly, however, firms intensifying lobbying efforts can maintain higher investments during policy turmoil. Additional analyses propose that the mitigating role of lobbying may be explained by the lobbying firms’ greater access to debt financing. Using matched samples based on propensity score matching (PSM) estimation ensures the empirical resilience of these findings.

How sensitive are matching estimates of active labor market policy effects to typically unobserved confounders?

Using a rich and unique combined administrative-survey dataset, this paper explores how sensitive propensity score (PS) matching estimates of Active Labor Market Policies (ALMPs) based on the selection-on-observables assumption are to typically unobserved covariates. Using a sample of German unemployed welfare recipients, the analysis shows that typically unobserved factors such job search behavior, concessions willing to make for a job as well as (mental) health are in fact relevant confounders. However, results also show that matching on the PS using only typically observed covariates reduces imbalance in terms of typically unobserved covariates by about 46 percent in this setting. In line with this finding, the inclusion of typically unobserved covariates yields very similar estimates to estimates based on a standard specification. Hence, a standard matching approach based on rich and high quality administrative data appears to be sufficient to obtain estimates that are rather robust to unobserved confounding.