Oaxaca-Blinder Decomposition Method Research Articles

Effects of Business Age and Size on Freight Demand: Decomposition Analysis of Indian Establishments

Freight studies have traditionally used business size measures such as employment and gross floor area as predictors of freight generation, giving very limited attention to the effects of an establishment’s age on freight demand. This study uses establishment-based freight survey data collected in seven cities of Kerala, India, to analyze the impacts of the age of an establishment on its freight demand. This is achieved by grouping the establishments into pre-specified classes with relatively homogeneous freight demand pattern. This classification is based on a data-driven a posteriori segmentation of industrial classes. These groups are further divided into sub-groups based on founding year of an establishment. As the business age increases, establishments grow in business size indicators and hence the changes in productivity can be either because of age or because of size or both. It is very important to separate the effects of these two indicators. The Blinder–Oaxaca decomposition method is used in this study to disentangle the effects of size and age on freight production and freight trip production. Age is found to be a fundamental driver of freight demand in younger establishments while the explanatory power of business size variables in explaining freight demand diminishes with age. The study findings illustrate the potential omitted variable bias that can occur when freight demand is estimated using business size indicators, without controlling for the business age differences.

Labour Market Discrimination in India

Discrimination along lines of caste, religion, gender, region, language, etc. is a reality in India today as it has been for most of its history. The fault lines in the approach to understanding wage disparity are exposed in its once-too-often overlooking of the sociopolitical realities of their subjects. The paper aims to bring out the wage discrimination in Indian labour market in both rural and urban areas belonging to different occupational activities using the simple ordinary least square regression analysis and Blinder–Oaxaca decomposition method to show that the wage disparity among different social groups can be attributed to discrimination.

What drives the gap? Applying the Blinder\u2013Oaxaca decomposition method to examine generational differences in transportation-related attitudes

Considerable recent work suggests that Millennials’ behaviors may be converging with those of Generation X as they enter later life stages, but few have investigated whether attitudes, which are often strong predictors of behavior, are undergoing the same convergence. In this study, we analyze the existing generational gap in four transportation-related attitudes (currently pro-urban, long-term pro-urban, pro-car ownership, and pro-environment), and examine the differential effects of other characteristics, including life-stage variables, on these attitudinal gaps. We apply the threefold Blinder–Oaxaca decomposition method to a statewide (weighted) sample of 1029 Millennials and 946 Generation Xers from California to unravel these effects. The method distinguishes among: (1) effects due to the cohorts having different characteristics (endowments); (2) effects due to those characteristics having different influences on attitudes (coefficients); and (3) the interaction of those two effects. We observe that Millennials’ attitudes: (1) differ from those of Generation X only by small, albeit statistically significant, amounts on average; and (2) are closer to those of Generation X as they gain on a host of life-stage variables such as marital status, income, and education. For example, if Millennials were married, employed, and earning higher incomes at the same rates as Generation X (but retaining their own model coefficients), the generational gap in the currently pro-urban attitude would be reduced by 24%. This study brings an econometric approach to the study of generational divides in transportation-related attitudes, with findings suggesting that Millennials might be leaving part of their uniqueness behind as they enter later life stages.

Assessing food and nutrition security in Nepal: evidence from diet diversity and food expenditure patterns

The widespread prevalence of nutrient deficiency and poor nutritional outcomes has been a significant concern in developing and emerging countries like Nepal. The level of diversity in household diets is an indirect measure of diet quality and the extent to which nutritional needs of households are being met. However, there is a limited understanding on the trends, patterns and determinants of the dietary diversity in Nepal. This study is an attempt to enrich the literature on this issue. Based on unit level data from three rounds (1995, 2004, and 2011) of the Nepal Living Standard Survey (NLSS), we used multilevel modeling, quantile regression, and the Blinder-Oaxaca decomposition method to decipher the trends, determinants and drivers of dietary diversity in Nepal. Our study finds that the changes in household socio-demographic and agricultural characteristics are very important in explaining the improvement in dietary quality. Changes in household characteristics account for at least 37% of the observed improvement and agriculture-related changes explain at least 16% of the observed improvement. Our findings are highly robust across the different model specifications and concludes by bringing a focus on multi-sectoral approach to tackle nutrition issues in Nepal.

Economic Inequality in Unmet Refractive Error Need in Deprived Rural Population of Iran.

Purpose:To determine economic inequality in unmet refractive error (RE) need and its determinants in deprived rural population of Iran.Methods:In this population-based study, two villages were randomly selected from among underserved villages of Iran. After selecting the participants, optometric examinations, including uncorrected and corrected visual acuity and subjective and manifest refraction, were done for all the participants. Then, unmet need for glasses was determined. Concentration index (C) was used to assess inequality, and Oaxaca–Blinder decomposition method was applied to decompose the gap between the two groups based on the determinants.Results:Of 3851 samples, 3314 participated in the study (response rate = 86.05%). The data of 3255 participants were used for analysis. The value of C and 95% confidence interval (CI) was −0.088 (−0.157 to −0.020), indicating a pro-poor inequality in unmet need. The prevalence (95% CI) of unmet need was 11.74% (9.25–14.22) in the poor and 6.51% (4.96–8.06) in the rich, with a gap of about 5% in favor of the rich (P < 0.001). A marked percentage of the gap was due to the explained portion (b = 5.73; P = 0.031). In the explained portion, the variable of economic status (b = 3.48; P = 0.004) and myopia (b = 0.88; P = 0.031) caused inequality in favor of the rich and against the poor, respectively. In the unexplained portion (b = −0.51; P = 0.372), the variables of education (P = 0.002) and place (P = 0.001) had statistically significant effects on inequality.Conclusions:There is a significant pro-poor economic inequality in the prevalence of unmet need in rural areas of Iran. Although part of this inequality is related to variables such as education and myopia, a major portion (two thirds) of this inequality may be due to the direct effect of economic inequality.

Socioeconomic inequality in smoking and its determinants in the Islamic Republic of Iran.

The role of socioeconomic inequality and related factors has not been well reported in tobacco consumption. To investigate the socioeconomic inequality in smoking and its associated factors in the Islamic Republic of Iran. Data were collected from surveillance for noncommunicable diseases in 2005, which included 89 404 people aged 15-65 years. Economic status was defined by principal component analysis on variables related to socioeconomic status. Concentration index and slope index of inequality were used to determine the inequality value. The gap between the high and low economic status groups was decomposed using the Oaxaca-Blinder decomposition method for explained and unexplained components. The total prevalence of smoking was 17.0%; 28.0% in males, and 5.8% in females, 15.8% in urban and 19.1% in rural areas. The concentration index was -0.032 in the whole of country; -0.098, in males, -0.246 in females, 0.014 in urban and -0.059 in rural areas and varied in different provinces of country. The smoking rate was 18.0% for the first quintile and 13.5% for the fifth quintile, a gap of 4.5%. The major part of this gap was related to differences in education level, sex, marital status and age in economic groups. There was a pro-rich socioeconomic inequality in smoking, especially in females and in the southern provinces. Increase in education level and empowering females of low socioeconomic status are sound interventions for alleviating inequality and for tobacco control.

Partisan bias in inflation expectations

We examine partisan bias in inflation expectations. Our dataset includes inflation expectations of the New York Fed’s Survey of Consumer Expectations over the period June 2013 to June 2018. The results show that inflation expectations were 0.46% points higher in Republican-dominated than in Democratic-dominated US states when Barack Obama was US president. Compared to inflation expectations in Democratic-dominated states, inflation expectations in Republican-dominated states declined by 0.73% points when Donald Trump became president. We employ the Blinder–Oaxaca decomposition method to disentangle the extent to which political ideology and other individual characteristics predict inflation expectations: around 25% of the total difference between inflation expectations in Democratic-dominated versus Republican-dominated states is based on how partisans respond to changes in the White House’s occupant (partisan bias). The results also corroborate the belief that voters’ misperceptions of economic conditions decline when the president belongs to the party that voters support.

Understanding Changes in Socioeconomic Inequality of Unintended Pregnancy Among Iranian Married Women: A Blinder-Oaxaca Decomposition Analysis

Background: Policymakers are interested in investigating effects of governments' policies on socioeconomic inequality in public health. Objectives: This study aimed to analyze levels of and changes in socioeconomic inequality of unintended pregnancy after the changes in family planning policies and to investigate determinants of its changes in Iran. Methods: Required data were extracted from Iran's Multiple Indicator Demographic and Health Surveys conducted in 2010 and 2015. We used data from 1123 and 900 married pregnant women aged 15 - 49 years in 2010 and 2015, respectively. Wagstaff normalized concentration index was used to measure unintended pregnancy inequality. The contribution of various factors to the measured inequality in 2010 and 2015 was investigated by decomposing concentration index. Changes in the unintended pregnancy inequality in 2010 - 2015 and its determinants were assessed using Blinder-Oaxaca decomposition method. Results: Pro-rich unintended pregnancy inequality declined by 120% from -0.145 to 0.030 in 2010 - 2015. However, the pro-poor unintended pregnancy inequality in 2015 was not statistically significant. Households’ economic status and women’s age at pregnancy were the two leading factors with positive contributions while contraceptive non-use before pregnancy and women’s education level had the most negative contributions to the reduced pro-rich inequality of unintended pregnancy in 2010 - 2015. Conclusions: Pro-rich unintended pregnancy inequality not only did not increase, but also declined to zero after the changes in family planning policies. Providing sustainable livelihood for disadvantaged households with women at reproductive ages can maintain this favorable condition in the future.

The Association of Unhealthy Diet with Socioeconomic Inequality in Children: A Study in Kurdistan, West of Iran

Background: The prevalence of obesity is an increasing public health problem. Objectives: The present study aimed at estimating the prevalence of unhealthy diet among children aged 10 - 12 years in Kurdistan, west of Iran and assessing the association of unhealthy diet with socioeconomic inequality. Methods: The present cross-sectional study was conducted on 2506 children living in Sanandaj city, west of Iran in 2015. The subjects were selected from the schools by multistage sampling method. The required data were collected using the food frequency questionnaire. The socioeconomic status (SES) was calculated by measuring household asset by using principal component analysis technique. The inequality was measured using concentration index, and Blinder-Oaxaca decomposition method was utilized to determine the proportion of different determinants causing inequality. Results: The prevalence of unhealthy diet was 50.75% (95% confidence interval (CI): 48.79 - 52.71). The concentration index for unhealthy diet was -0.14 (95% CI: -0.18, -0.09), indicative of concentration of unhealthy diet in the group with low SES (P < 0.001). The Blinder-Oaxaca decomposition analysis showed that the poor-rich gap in the prevalence of unhealthy diet was 27.77%, while 59.41% of the observed gap attributed to the explained component. The major causes of inequality were mother’s level of education and place of residence, respectively. Conclusions: The obtained results showed that the prevalence of unhealthy diet was higher in children with low SES. Increasing mothers’ awareness of health literacy and performing place-based interventions can be effective in decreasing socioeconomic inequalities in unhealthy diet and prevention of chronic diseases in children.

Changes in Perceived Accessibility to Healthcare from the Elderly between 2005 and 2014 in China: An Oaxaca-Blinder Decomposition Analysis.

Elderly people are characterized with high needs for healthcare, accompanied by high barriers in access to healthcare. This study aimed to identify temporal changes in access to healthcare and determinants of such changes from the elderly in China, over the period between 2005 and 2014. Two waves (2005 and 2014) of data were extracted from the Chinese Longitudinal Healthy Longevity Survey (CLHLS), measuring changes in perceived accessibility to healthcare when needed by the elderly (≥65 years). The effects of the explanatory variables (need, predisposing and enabling factors) on the changes were divided into two components using the Oaxaca–Blinder decomposition method: (1) the endowment portion as a result of distribution differences of the explanatory variables and (2) the coefficient portion as a result of differential responses of the dependent variable to the explanatory variables. Perceived accessibility to healthcare from the elderly increased from 89.6% in 2005 to 96.7% in 2014. The coefficient portion (82%) contributed more to the change than the endowment portion (63%) after adjustments for a negative interaction effect (−45%) between the two. Lower perceived accessibility was associated with older age, lower income, lower affordability of daily expenses and lower insurance coverage. But the coefficient effects suggested that their impacts on perceived accessibility to healthcare declined over time. By contrast, the impacts of gender and out-of-pocket payment ratio for medical care on perceived accessibility to healthcare increased over time. Perceived accessibility to healthcare from the elderly improved between 2005 and 2014. Gender gaps are closing. But the increased effect of out-of-pocket medical payments on perceived accessibility to healthcare deserves further investigation and policy interventions.

Educational expansion and health disparities in Ethiopia, 2005–2016

Research shows that basic education improves population health, yet it remains unclear whether the expansion of primary education decreases health disparities. In this paper, I assess whether disparities in healthcare utilization decreased in conjunction with educational expansion among women of reproductive age in Ethiopia. Healthcare utilization rates in low-resource countries are often confounded with simultaneous developments in education and access to basic healthcare. Using decomposition of rates, I first disentangle the changes in health disparities induced by educational expansion from the overall increase in healthcare utilization. Then, I use the Blinder-Oaxaca decomposition method to investigate the determinants of disparities in healthcare utilization and how these determinants changed over a 10-year period as primary education became more prevalent. Overall, disparities in healthcare utilization in Ethiopia decreased over time, yet the association between educational expansion and health disparities varies by region. Literacy explains much of the disparities in healthcare utilization, yet it loses significance over time as primary education becomes widespread. Economic factors remain persistent sources of disparities, and non-financial barriers such as the distance to travel and women's ability to travel alone become more significant. Heterogeneity in healthcare utilization across regions has distinct implications for how educational expansion may shift health disparities.

PMU104 EXPLAINING DIFFERENCES IN HEALTH-RELATED QUALITY OF LIFE (HRQOL) BETWEEN PERSISTENT AND NON-OPIOID USERS WITH CHRONIC NON-CANCER PAIN CONDITIONS (CNCP): A POST-REGRESSION DECOMPOSITION ANALYSIS.

There is some evidence that persistent opioid users (POU) have poor HRQoL as compared to non-opioid users (NOU). However, how much the patient-level characteristics contribute to the differences in HRQoL between POU and NOU is not known. We used Blinder-Oaxaca decomposition method to identify factors that contributed to the differences in HRQoL between POU and NOU among adults with CNCP. We used longitudinal data from multiple panels (2012 – 2015) of the Medical Expenditure Panel Survey. Adults older than 18 years, without cancer diagnosis, complete information for Short-Form 12 (SF12) questionnaire, and with self-reported CNCPs (arthritis, back/neck pain conditions, musculoskeletal conditions, or headache or migraine) were included in the study (N = 6,038). POU was defined as opioid use in two consecutive rounds during the baseline year. HRQoL was measured using Short Form-12’s physical health component (PCS) and mental health component (MCS) summary scores in the follow-up year. Oaxaca-Blinder linear regression decomposition technique was used to identify factors that contributed to differences in HRQoL between POU and NOU. 14% of our sample reported POU. POU had significantly lower PCS (33.7 vs 45.5; difference:11.8, p<0.001) and MCS scores (45.9 vs 50.0; difference:4.1, p<0.001) compared to NOU. Decomposition analysis revealed that a difference of 5.91 points (50.3%) in PCS was explained by differences in pain severity, multimorbidity, and employment status between POU and NOU. When baseline PCS was added to the model, it explained 61.1% of the difference in PCS between the groups. Since difference in MCS between the groups was not clinically meaningful, no decomposition analysis was performed. Among adults with CNCP conditions POU had significant HRQoL burden. Poor HRQoL among POU can be mostly explained by higher rates of depression, anxiety, multimorbidity, and lower rates of employment in this group compared to NOU.

Socioeconomic inequality in cardio-metabolic risk factors in a nationally representative sample of Iranian adolescents using an Oaxaca-Blinder decomposition method: the CASPIAN-III study.

The present research was conducted aiming at assessing the association of socioeconomic inequality in the prevalence of risk factors associated with cardio-metabolic disorders in a sample population of nationally representative Iranian adolescents and to identify its influencing factors. This study was conducted as part of a national-based surveillance program performed on 5625 individuals aged 10-18years in 27 provinces in Iran. To determine the socioeconomic status (SES) of participants, we defined a new variable by applying the principal component analysis. Doing so, the socioeconomic inequality in cardio-metabolic risk factors was examined over the tertiles of SES using concentration index (C). Then, Oaxaca-Blinder decomposition analysis was carried out in order to decide upon the roots of inequality in the health system. The mean (standard deviation) age of participants was 14.73 (2.41) years. The prevalence of cardio-metabolic parameters had considerable difference across SES tertiles. Elevated fasting blood glucose (FBG), elevated triglycerides (TG), abdominal obesity, elevated total cholesterol (TC), and metabolic syndrome (MetS) increased linearly by increasing SES tertiles. C index for depressed high density lipoprotein- cholesterol (HDL-C) was negative, which was suggestive of inequality in favor of high SES groups and for other cardio-metabolic parameters, it was positive, which indicate inequality was in favor of the lowest SES groups. The highest gap between the first and third tertiles of socioeconomic was for frequency of abdominal obesity; 13.18% of the lowest SES groups and 20.11% of the highest SES groups had abdominal obesity which accounts 6.93% gap in favor of the highest SES groups. The living area could be named as the main variables standing for the inequality of elevated FBS, elevated LDL-c, low HDL-c and abdominal obesity frequency between the first and the last SES group. In addition, BMI could stand as the main independent variable explaining the gap in elevated TG, elevated TC, elevated BP and MetS prevalence across the lowest and the highest SES group. The study revealed the considerable inequality in the prevalence of cardio-metabolic risk factors between the highest and the lowest SES groups of Iranian adolescents. Living area and BMI are the two main factors which explained inequality in prevalence of cardio metabolic risk factors between SES groups. These estimations could provide health policy markers with practical information for future complementary analyses.

Factors contributing to racial disparities in influenza vaccinations.

BackgroundRacial/ethnic disparities in rates of influenza vaccinations in the US remain an issue even among those with access, no out-of-pocket costs, and after adjusting for confounders. We used an approach called the Oaxaca-Blinder (OB) decomposition method to ascertain the contribution of covariates individually and in aggregate to the racial disparity in influenza vaccination.MethodsWe included members > = 18 years of age as of 05/01/2014 with continuous enrollment through 04/30/2015. Influenza vaccination was defined by diagnosis, procedure, or medication codes, or documentation in the immunization table. Characteristics were reported by race. Logistic regression models estimated the odds of vaccination associated with: (1) race; and (2) covariates stratified by race. The Oaxaca-Blinder (OB) method calculated the contribution of covariates to the difference or disparity in vaccination between Blacks and Whites.ResultsWe found that among adults, 44% were vaccinated; 55% were Black; and 45% were White. Black members have 42% lower odds of vaccination than White members. The contribution of the differences in the average value of the study covariates between Black and White members (the OB covariate effect) accounted for 29% of the racial disparity. The contributions to the total White-Black disparity in vaccination included: age (16%), neighborhood median income (11%), and registration on the online patient portal (13%). The contribution of the differences in how the covariates impact vaccination (OB coefficient effect) accounted for 71% of the disparity in vaccination between Blacks and Whites.ConclusionIn conclusion, equalizing average covariate values in Blacks and Whites could reduce the racial disparity in influenza vaccination by 29%. For health system vaccine campaigns, improving registration on the patient portal may be a target component of an effective system-level strategy to reduce racial disparities in vaccination. Additional information on patient-centered factors could further improve the value of the OB approach.

Access to Formal Credit in the Indian Agriculture: Does Caste matter?

Access to resources and opportunities can be a critical factor in improving outcomes for disadvantaged groups. Improving access to financial resources, in particular, is widely acknowledged to facilitate upward economic and social mobility. Conversely, lack of access to resources for certain groups based on caste, class, gender and ethno-social identities can perpetuate inequalities. In this context, this paper attempts to analyse the access to credit by social groups and decomposes the gross credit differentials using Oaxaca-blinder decomposition method using unit-level data from the All India Debt and Investment Survey, NSSO, 2013. The descriptive analysis clearly shows that there is a significant credit differential between forward caste (FC) and other social groups (SC, ST and OBC). Access to credit varies across social groups based on many factors. The decomposition result indicates that the discrimination coefficient against SC is 49per cent which explains that SCs are being discriminated by 49 per cent compared to FCs in the formal credit market. In case of ST, the discrimination coefficient against is 61per cent and for OBC it is 48per cent. Interestingly, the endowment difference is less among ST (38per cent) compared to SC and OBC (around 51 per cent). Also, the FC treatment advantage (benefit of being a FC in the credit market) is 5.7 per cent whereas the cost of being an SC in the credit market (treatment disadvantage) is 35.1 per cent. As expected, the disadvantage component for ST and OBC is 33.1 per cent and 17.8 per cent respectively. Thus, we see that although programmes, schemes and policies to promote the economic empowerment of lower castes through finance have been implemented on a large scale since the 1990s, they have not been very effective.

Differences in entrepreneurial intentions of Chinese college students: evidence from an Oaxaca–Blinder decomposition

China has been characterized by sharp regional differences in quality of higher education. Using the data from a field survey on the entrepreneurial intentions of Chinese college students, this work adopts the Oaxaca–Blinder decomposition method to assess the extent and the determinants of the disparity of entrepreneurial intentions of college students from rural and urban areas. The analysis shows that both endowment and coefficient effects contribute to the rural–urban gap in entrepreneurial intentions of college students. This suggests that, though the different entrepreneurial intentions are mainly the result of the different personal qualities, more profound works are required to encourage the entrepreneurial intentions of college students from rural and urban areas.

Is Cooking Still a Part of Our Eating Practices? Analysing the Decline of a Practice with Time-Use Surveys

People now spend less time doing housework in general, and cooking in particular. So is cooking still a central feature of our daily eating practices? This article compares trends in household cooking durations in France and the USA in the period 1985–2010 using time-use surveys and practice theory. We ask how the association between cooking and eating at home has changed over time, and how it has contributed to the decline in the time spent on household cooking. Descriptive statistics show that US households spent 20 minutes less time per day cooking in 2010 than in 1985 (15 minutes less time per day in France). Linear regressions indicate that the association between cooking duration and the number of eating events at home has declined in the USA but not in France. The Blinder-Oaxaca decomposition method points to this fact as the primary reason for the change in cooking time in the USA; in France, decreased cooking time is accounted for primarily by changes in population characteristics. French and American food practices have followed gradually diverging trajectories, with cooking less a feature of eating practices – even at home – in the USA, whereas the association between eating and preparing food at home remains stable in France.

Gender Inequality and Access to Microfinance: Evidence from Ghana

AbstractThis paper draws on the qualitative and quantitative evidence from a dataset generated from a survey of 499 households in Ghana to explore the implications of access to microfinance for the gender asset gaps. Two sets of statistical analysis are undertaken. The first is a cross-sectional regression analysis that examines the importance of microfinance for intra-household gender wealth inequality. The second uses the Oaxaca-Blinder decomposition method to examine the gaps in wealth between female-headed and male-headed households. The results show that higher access to microfinance is associated with lower gender asset gaps within and across households. The evidence suggests that targeting credit to financially constrained households, and in particular to women, can contribute to reducing poverty and gender inequality.

Determinants of under-5 mortality in Burkina Faso.

The aim of this article is to determine the factors associated with under-5 mortality and their evolution from 1993 to 2010 and to analyse the contributors of socioeconomic inequalities in mortality of children under-5 years during the same period. The data used in this study were derived from the four rounds of Demographic and Health Survey (DHS) conducted in Burkina Faso in 1993, 1998 and 2010. Concentration measurement, logistics regression and Oaxaca-Blinder decomposition method were used to analyse data. Multivariate analysis revealed that being the first child (odds ratio = 1.8 for 1993, 1.7 for 1998, 1.2 for 2003 and 1.3 for 2010) or a twin (odds ratio = 4.5 for 1993, 2.8 for 1998, 2.7 for 2003 and 4.8 for 2010) were also significantly associated with the probability of dying. The variable (parity) was the main contributor to the part of the inequality due to differences in group characteristics and that would be due to the fact that women from poor households have greater parity compared to those from rich households. For a reduction in mortality and inequalities related to mortality, the implementation of actions in favour of poor households and promotion of family planning programmes for birth spacing will be required.

Decomposing the Labour Productivity Gap between Migrant-Owned and Native-Owned Firms in Sub-Saharan Africa

Migration studies have been primarily based on the movement of individuals from developing to developed economies, with a focus on the impact of migrants on host country wages. In this study we take a different angle by exploring the labour productivity of migrant-owned firms versus native-owned firms in 20 African economies using firm-level data. We find that labour productivity is 78 per cent higher in migrant-owned firms than native-owned firms. Using the Oaxaca-Blinder decomposition method we find that structural effects account for 80 per cent of the labour productivity gap. Returns to manager education largely explain the productivity advantage of migrant-owned firms over native-owned firms. Interactions with the government, access to finance, informality, and power outages are also considerable contributors to the labour productivity gap.