Anyone who engages in econometrics pays attention to this account
Manuscript: [email protected]
All the code programs, macro and micro databases and various software of the econometric circle methodology are placed in the community. Welcome to the econometric circle community for exchanges and visits.
Today, we would like to introduce the most well-known paper "Missing Women in China and Tea Prices: The Impact of Gender-Specific Income on Gender Inequality" published on QJE by Nancy Qian. Nancy Qian (钱楠魠) is currently a professor at Northwestern University in the United States. He graduated from MIT with a Ph.D. in 2005. Its doctoral supervisors are Esther Duflo, Abhijit Banerjee and Joshua Angrist. What's interesting is that Esther Duflo's doctoral supervisors are Abhijit Banerjee and Joshua Angrist. Therefore, Nancy Qian should also be called Abhijit Banerjee and Joshua Angrist according to seniority. In terms of nationality at birth, Nancy Qian should be called Chinese, Esther Duflo is French, Abhijit Banerjee is Indian, and Joshua Angrist is Israeli. It is not difficult to imagine how strong she needs to train her listening ability when facing three accents of English at the same time to be able to do it with ease!
If you want to know more about Nancy Qian, you can refer to: ①Continuous DID classic literature, Potatoes made the civilization of the old world, ②Material package about instrumental variables, titles, models, endogenous variables, instrumental variables, ③DID use classic literature , Compulsory license: Evidence from the trade law against the enemy, ④All the interpretations, materials, procedures, data, literature and various deformations of DID are here, ⑤Professor Lu Ming’s empirical suggestions for scholars engaged in empirical research, ⑥Measurement A full analysis of economic empirical paper writing, and a highly recommended work. ⑦The top 500 list of top 5 journals is released. These Chinese people deserve their names. ⑧The Nobel Prize couple spent 16 years, and the research on China is finally published in the journal. It's coming out!
This Missing women is also an earlier article using cross-sectional DID for policy evaluation. ①Cross-section data DID describes the paradigm of double-differential policy evaluation for cross-section data. ②Guide to the operation procedure of cross-section data DID, teach you step by step, ③cross-section DID, each A fixed effect, placebo test, displacement test, and other external shock treatments. Of course, the instrumental variable method is also used creatively in the article.
It is precisely because Nancy Qian is good at using the DID method to do influential research, he is called the "DID little princess" in a small circle of academia. After all, Nancy Qian is more familiar with China's national conditions, so after she became famous with the article Missing women, she has launched a lot of academic cooperation with Professor Yao Yang from Peking University, Professor Chen Shuo from Fudan University, and Professor Meng Xin from Australian National University.
Text
on the bottom of the text content, author: Chen Mo Han, the United States Tufts University Department of Economics, communication mail: [email protected]
before authors: the practice of double-difference method DID implicit assumption
MISSING WOMEN AND THE PRICE OF TEA IN CHINA: THE EFFECT OF SEX-SPECIFIC EARNINGS ON SEX IMBALANCE
China’s missing women and the price of tea: the impact of gender-specific income on gender inequality
Main point:
This article uses the exogenous growth of sex-specific incomes in agriculture resulting from China’s post-Mao Zedong reforms to estimate the impact of total income and sex-specific income on the survival rates of children of different genders. Increasing the income of women and keeping the income of men unchanged can increase the survival rate of girls, while increasing the income of men and keeping the income of women unchanged will worsen the survival rate of girls. An increase in women’s income will increase the education level of all children, while an increase in men’s income will reduce the education level of girls, but has no effect on the education level of boys.
1.Introduction P1251-1254
In this part, the author outlines the research motivation and research methods.
Research motivation:
This article first proposes that there is a serious gender imbalance in Asia: 50.1% of the current population of Western European countries are women, while only 48.4% of India and China are women. Amartya Sen (1990, 1992) referred to this observed deficit as "missing women". Most of the missing women in the world are in China and India. It is estimated that between 30 and 70 million women are missing here. However, the author also pointed out that some wealthy Asian regions such as South Korea and Taiwan have the same gender imbalance. Among the population born between 1970 and 2000 in China, with the rapid economic growth, the proportion of men increased from 51% to 57% (see Figure 1 in the original text for details).
Research method:
Next, the author explains how to solve the problem of identifying the impact of income on children's gender (the increase in women's income may come from many aspects): This article uses two post-Mao reforms in China to eliminate the problem of missing variable deviations. In the Mao Zedong era, the centrally planned production target was concentrated on staple corps. In the early post-Mao era reforms (1978-1980), the reforms increased the income of cash corps, including tea and orchard products. Women have a comparative advantage in producing tea, while men have a comparative advantage in producing orchard fruits. Therefore, in areas suitable for tea cultivation, the income generated by women has increased, while in areas suitable for the cultivation of orchard crops, the income generated by men has increased. This allows the author to use the double difference method (DID) strategy to determine the causal relationship of sex-specific income increases to the child’s gender outcome.
To assess the impact of gender-specific income changes, the authors compared the sex ratios between the population cohorts born in counties that planted and did not grow gender-specific crops before and after the reform.
The author first compares the changes in the proportion of boys in the cohort born before and after the price increase in tea-producing and non-tea-producing counties to estimate the impact of increased adult female income on gender imbalance (maintaining adult male income unchanged).
Then, the author compared the impact of the increase in the relative income of men in counties that produce orchard crops and those that do not produce orchard crops (keeping the income of adult women unchanged).
Finally, the author continued to use the same strategy, but took education level as the explained variable, and estimated the impact of increased income by gender on the education level of boys and girls.
2. EMPIRICAL STRATEGY P1254-1266
In this part, the author elaborated and demonstrated the reasons for studying tea picking, outlined the main content and impact of the reform, and constructed four empirical models for different issues .
Tea picking VS orchard?
This article uses the value of tea to replace women’s wages, and the value of orchard crops to replace men’s wages. The author first examined the sex attributes of tea and orchard crops. Its research shows that the number of tea sown per household and the proportion of arable land used for tea production in each household are negatively correlated with the proportion of male workers in the household. (For details, see Table I of the original text, columns (1)–(4)). The number of orchards planted by each household and the proportion of farmland used by the family for orchards are positively correlated with the proportion of male laborers in the family (see original table I, columns (5)–(8) for details). It should be noted that women in the tea industry only have a comparative advantage in picking tea. In the 1982 census, 56% of the labor force in tea production (including picking, pruning, and drying) was male, while 62% of the labor force in orchard production was male. Because women have a comparative advantage in picking, the author interprets this 6% difference as a lower bound estimate of the comparative advantage of women in tea picking.
A simple cross-sectional comparison of the proportion of boys in non-tea-producing and tea-producing counties shows that the proportion of boys in the latter has decreased by one percentage point, while the proportion of girls has increased by one percentage point (see original table II). However, due to the bias of missing variables such as preference changes, this comparison cannot simply be used as an argument to support the conclusion.
Reform content and impact:
During the research period, the central government divided crops into three categories. Category 1 includes crops necessary for national welfare: cereals, all oil crops and cotton. The second category is cash crops, including orchard products and tea (Sicular1988a). Category 3 includes all other agricultural crops (mainly local secondary crops). Only category 1 and 2 crops are subject to quota restrictions. The so-called "two reforms" refer to the gradual reduction of the proportion of planned indicators and the implementation of the household contract responsibility system (HPRS).
Compared with the income from the first type of staple food, the reform has increased the income of tea and orchard products. Note that the income from tea does not exceed the income from orchard products (Figure IIa). Although the yield of category 1 crops has increased, the growth rate has not changed (Figure IIb). After the reform, with the increase in purchase prices, the output of category 2 crops such as melons and orchard fruits has increased at a faster rate (Figure IIc). The tea industry has seen similar growth (Figure IId).
Empirical model
1) Model 1: The impact of category 1 and category 3 crops on the proportion of boys:
As mentioned above, the so-called increase in gender-specific income is based on the two categories of crops relative to category 1 crops (prices continue to be suppressed) and 3. The value of similar crops (never regulated) has increased. Therefore, the impact of changes in the value of the first and third crops on the proportion of boys after the reform should not be changed. Therefore, the author designed the following model:
i represents the county, c represents the population cohort, sex is the proportion of boys, cat1 and cat3 represent the planting amount of the first and third crops, d_L is a dummy variable that represents whether the population cohort was born in the Lth year, and Han represents the Han population proportion. Figure IIIa shows the coefficients of the two interaction terms, which shows that the coefficients of the first and third crops are very close to zero before and after the reform.
2) Model 2: The impact of the tea industry on the proportion of boys after the reform:
The author uses the population cohort born between 1970-1986 as the experimental group and the population cohort born between 1970-1979 as the control group to estimate the following DID model:
Tea is the amount of tea planted, orchard is the amount of orchard crops planted, cashcrop is the planting amount of all cash crops, and postc is the reform variable. If the individual is born after 1979, it is equal to 1. If the increase in the value of tea improves the survival rate of women, then this should be reflected in the decline in the proportion of boys born after the reform, β<0. Conversely, if the increase in the value of orchard crops deteriorates the survival rate of girls, the authors can expect δ>0.
3) Model 3: Is the impact of tea production on gender ratio related to the reform:
The author further pointed out that the DID method has potential flaws, and it may compare the effect of the reform with other changes that may occur before or after the reform (such as changes in gender preferences) ) The effect of confusion. For example, tea-producing regions may experience different gender ratio trends relative to other regions, which may cause DID estimates to not only reflect the impact of tea price increases, but also capture the differences between tea-producing regions and non-tea-producing regions. The author proves the accuracy of the DID estimates by two methods:
First, Figure IIIb plots the proportion of boys in each birth year cohort in tea-producing and non-tea-producing counties. The vertical distance between the two lines shows that before the reform, there were more boys in tea-producing counties, but after the reform, the number of boys in tea-producing counties continued to be lower than that in non-tea-producing counties. The DID estimate is the difference between the average vertical distance before and after the reform. The figure clearly shows that before the reform, there were more boys in tea-producing areas than in non-tea-producing areas, and after the reform, the number of boys in tea-producing areas continued to be lower than in non-tea-producing counties. Therefore, DID estimates will not reflect the pre-reform trend difference in the sex ratio between tea-producing areas and non-tea-producing areas.
Second, the author further returns the interaction terms between the proportion of boys and the number of cash crops grown in the county of birth and the year of birth to more rigorously test whether the impact of tea production on the sex ratio occurs in the year of birth close to the point of reform:
βL is the effect of tea production on the proportion of boys in cohort L. If increasing the price of tea improves the survival rate of girls, βL should remain constant until the time when the reform begins, after which the reform becomes negative. Similarly, δL is the effect of orchard crops on the proportion of boys in the L group. If rising orchard crop prices deteriorate the survival rate of girls, δL should remain constant until the time when the reform begins, after which it should become a positive number.
4) Model 4: Introducing the instrumental variable regression of slope:
There are two problems that need to introduce instrumental variables. First, due to data acquisition problems, the author uses 1997 agricultural conditions instead of previous years, which may cause measurement errors. Secondly, OLS may have omitted variable deviations. For example, after the reform, families who prefer girls are more likely to switch to tea planting. The author introduced slope as an instrumental variable for tea production.
Because tea requires warm and semi-wet mountain tops as growing conditions. If the slope has no direct influence on the differential investment decision and is not correlated with any other covariates in equation (5), then the slope is an effective instrumental variable for the amount of tea planted. The author limited the sample to tea-producing counties and those non-tea-producing counties that are bordered by tea-producing counties.
The first stage of regression:
Return of the second stage:
3. CONCEPTUAL FRAMEWORK P1266-1267
In this section, the author proposes two family models based on the relationship between tea prices and the survival rate of girls, and proposes a joint hypothesis test of family categories and the way parents treat their children.
Since the technology of revealing gender before birth was not available during the most relevant period of this study, the observed gender imbalance was caused by the differential negligence of girls or the killing of female infants in some cases. The possibility of a girl’s survival increases as the needs of girls relative to boys and the cost of gender selection increase.
Increasing the price of tea can improve the relative survival rate of girls in four ways. First, it can increase the relative desire to give birth to a girl by raising the parents’ expectation of their daughter’s future income relative to their son’s. Secondly, if for some reason daughters are luxury goods relative to sons, the increase in total family income will increase the relative demand of girls. Third, increasing women-specific income can increase mothers’ bargaining power. If mothers like girls more than fathers, this will increase the relative survival rate of girls. Finally, because the sex of the child must be revealed at a certain stage of pregnancy, increasing the labor value of adult women may increase the cost of gender selection. The first, second and last explanations are consistent with the unitary and nonunitary models of family decision-making. The third explanation is likely to be consistent with the non-uniform model of the family.
The unified model makes a strong prediction that no matter which family member brings the extra income home, the increase in income should have the same effect on household consumption. Therefore, if the increase in the price of orchard products does not affect the survival rate of girls as much as the increase in the price of tea, then the second explanation can be ruled out. By comparing the impact of rising prices of tea and orchard products on the relative education level of girls, the number of potential explanations can be further streamlined. In this case, the mother’s opportunity cost of time does not apply. In the first explanation, there is a unified family requirement with investment motives. The effect of increasing the relative value of female labor force on the education level of girls is symmetrical with the effect of increasing the relative value of male labor force on the education level of boys.
By examining whether the increase in tea prices and the price of orchard products have the same impact on gender imbalance, the author tested the joint hypothesis that the family is a unified family and parents treat children as a form of consumption. The author examines whether the rise in tea prices has the same effect on the education of girls as the rise in prices of orchard products has the same effect on the education of boys, and tests the joint hypothesis that the family is a unified family and parents treat children as a form of investment.
4. THE DATA P1268-1272
In this part, the author elaborated on the data sources and data processing standards and performed descriptive statistical analysis. For possible confounding variables, the author excluded them based on the data.
The author uses the 1% sample of the 1997 China Agricultural Census and the 1% sample of the 1990 China Census, as well as GIS data from the China Data Center in Michigan. The sample includes all 1,621 counties from 15 provinces in southern China. All provinces producing tea in the 1997 agricultural census were included. The 1990 census contained data on gender, year of birth, education level, type of department and occupation, and relationship with the head of the household. Since cities and rural areas have experienced different family planning policies and market reforms, the author limits the analysis to rural families.
To avoid confusing the estimates with the impact of immigration, the authors further restricted the data to reporting individuals who have lived in the same county for more than five years. This is based on the assumption that the county of birth is the county of residence of those who have reported that they have lived for five years or more.
The author also excluded the interference of the "one-child policy" on the birth rate of girls based on the data. Figure 1 shows that the proportion of males by year of birth is stable over time. In order to maintain rigor, the author only uses data from children over 4 years of age in the DID estimation.
After comparing the descriptive statistics of tea-producing counties and non-tea-producing counties in Table II, the author believes that there are no prominent systematic differences between the two groups.
Figure IVa shows the distribution of tea-producing counties. Darker shades correspond to planting more tea trees. It shows that the tea-producing counties are geographically dispersed, which helps alleviate concerns about their observable characteristics (such as culture) that are systematically different from the control group. Figure IVb shows the slope changes in China, and the shaded areas in the figure are steeper. By comparing the tea-growing counties in Figure IVa with the hilly areas in Figure IVb, we can see the predictive ability of slope for tea planting. The author uses the GIS data shown in Figure IVb to calculate the average slope of each county.
5. EMPIRICAL RESULTS P1272-1279
A.
Column (1) of Girl Survival Rate Table III shows the DID estimate according to formula (2). The results show that for every acre of tea planted, the proportion of boys can be reduced by 1.2 percentage points. Planting an additional mu of orchard crops can increase the proportion of boys by 0.5%; planting cash crops has no effect on gender in general. The coefficients for planting tea and orchard crops are statistically significant at the levels of 10% and 5%, respectively.
The estimated values of β1, δ1, and ρ1 based on equation (3) are plotted in Figure V. They show that for the cohorts born before the reform, the effects of planting tea and orchard crops on the proportion of boys were similar and remained the same in each cohort. However, the people born around the reform period showed different influences. Tea production is associated with a decrease in the proportion of boys, while planting orchard crops is associated with an increase in the proportion of boys. Over time, the difference effect persists. These results lead people to believe that the impact of the production of tea and orchard crops on the proportion of men can be attributed to the agricultural reforms in the post-Mao era, rather than other changes in these areas.
The author also considered the problem of cohort fixed effects (cohort fixed effects control the differences between birth cohorts in the county, and these differences do not vary between counties). The author incorporated the interactive term of the county's dummy variable and the linear time trend to solve this problem. In order to make the estimation results comparable with the 2SLS estimation, the author limited the sample to only counties with geographic data, and regressed the same model as the second stage of 2SLS. Column (2) of Table III shows the basic fixed-effect estimates. Column (3) shows the estimated value after the author controls the county-level cohort trend. The estimated results are similar and significant at the 5% level. This further proves the robustness of OLS estimation.
Column (4) of Table III shows the first stage estimation of 2SLS according to equation (4). The estimated value λ of the correlation between hills and tea production is significant at the 5% level. Column (5) shows the estimated value of 2SLS estimated according to formula (5). This estimate is greater than the OLS estimate and has statistical significance. Column (6) shows the estimated value of 2SLS after controlling for county-level cohort trends. This estimate is similar in magnitude to the OLS estimate, but it is no longer statistically significant. There is no statistical difference between the estimated value of the included trend and the estimated value of the excluded trend. The estimated value that is not included in the trend is larger, but the accuracy of the estimate is lower. The estimated value of 2SLS in column (6) shows that there is no bias in the estimated value of OLS under the condition of considering the time trend of the county-level cohort. In addition, the estimated values of OLS and 2SLS in columns (3) and (6) are almost the same numerically as the initial OLS estimates in column (1). The author once again proves that OLS is robust.
B. education level
This part of the analysis uses county-level-year-of-birth data from a 0.05% sample in the 2000 census. The empirical strategy is the same as before. The author estimates equation (2) with years of education as the dependent variable to test the impact of tea planting, orchard products, and all second-category cash crops on the overall education level. Then, the author repeats the above steps, using the education level of girls, the education level of boys, and the difference in education level between boys and girls as dependent variables to estimate. First, select dummy variables that indicate whether tea, orchards or cash crops are planted in the county, and then use continuous variables to indicate the amount of each crop planted.
Dummy variables: The estimates in Table IVA show that tea production increases the years of education for women, men, and all children by 0.2, 0.25, and 0.15 years, respectively. On the other hand, the cultivation of orchard products reduces the education level of women by 0.23 years, but has no effect on the education level of men. These estimates are statistically significant at the 1% level. The estimates in column (4) show that tea planting reduces the difference in education levels between men and women, while planting orchards increases this difference. The latter is significant at the 1% level. The coefficient of the second category of cash crops is close to zero, which is not statistically significant.
Continuous variables: Columns (5)-(8) of Table IV show that the estimated value has the same sign as the estimated value using dummy variables in columns (1)-(4). Estimates show that for every additional acre of tea garden, women’s education level increases by 0.38 years, men’s education level increases by 0.5 years, and every additional acre of orchard crops reduces women’s education level by 0.12 years. The education level of men has no effect.
In order to observe the time when the influence of tea production on education level appeared, the author examined the influence of tea production according to the year of birth. The author estimates equation (3) with years of education as the dependent variable to test the effect of tea production in different birth years. The author plotted the three-year average of the estimated coefficients for each cohort L in the vectors β1 and δ1 in Figure VI. It shows that before 1976, the educational level of women in the tea and orchard crop areas was similar. After 1976, the tea-producing areas increased, while the orchard crop areas decreased.
C. Robustness test of
family planning policy
If the implementation of family planning policy is systematically different between tea areas and non-tea growing areas, then the empirical strategy will confuse the effect of tea planting with the effect of family planning policy.
The author first points out that there is no evidence that family planning policies are systematically different in tea-producing areas and non-tea-producing areas. Second, the author uses the fact that non-Han ethnic minorities are largely exempt from family planning restrictions and conducted two additional tests of robustness.
First, the author introduces the interaction term between the proportion of Han nationality and the dummy variable of birth year for control. Next, the author re-estimates equation (2) using a sample that contains only minorities. In both cases, the estimates are similar to the main results, indicating that they are not confused with family planning policies. The results were not reported in the paper.
Immigrants
If the migration pattern between tea-producing areas and non-tea-producing areas is significantly different, the OLS estimate may reflect the impact of migration rather than the impact of income changes. In particular, compared to non-tea producing areas, it is assumed that the proportion of women born before 1979 leaving tea producing areas is too high. In this case, empirical strategies will mistakenly attribute changes in gender imbalance to changes in gender-specific survival rates, rather than changes in migration.
In order to solve the immigration problem more directly, the author deliberately overestimated the number of female immigrants from tea-producing areas. The DID estimates previously proposed used a sample of individuals between the ages of 4 and 20 in 1990. The 2000 census reported whether a person currently lives in the county of birth. The author assumes that all individuals under the age of 20 who live in non-born counties are women who were born in tea-producing areas before 1979. Then, the author added these immigration data to the 1990 data and re-estimated equation (2). This seems more conservative than the original model, because the migration rate in 2000 was about an order of magnitude higher than in 1990. Even with this extremely conservative method, the DID estimate will hardly change. The results were not reported in the paper.
6. INTERPRETATION P1279-1280
The empirical results have several theoretical implications. The results of research on survival and education rejected the combined hypothesis that the family is a one-element family and that the parents believe that girls are luxury items relative to boys. The indoor bargaining model provides a simple alternative explanation. If mothers value education more than fathers, and it is more expensive for mothers to ignore children of any gender, then the increase in mothers’ bargaining power will lead to more equal treatment of boys and girls, which is reflected in the data as an increase in the relative survival rate of girls.
7. CONCLUSION P 1281
This article explores the long-standing issue of whether economic conditions affect the outcome of girls relative to boys. The empirical results provide a clear affirmative answer: both gender imbalance and education level respond quickly to changes in gender-specific income. In addition, increasing the total household income without changing the relative income share of men and women has no effect on survival rates or investment in education. The findings of this article show that the worsening of women’s wage disadvantages may be an important source of the increase in the number of missing women in China. Similarly, the increase in the wage gap between men and women may be one of the reasons for the decline in rural school enrollment rates observed in the early 1980s. The policy recommendations for these results are clear. One way to reduce girl mortality and increase investment in children’s overall education is to increase the relative income of adult women. 
Nancy Qian's Publication list:
Regarding the DID double difference method, scholars can refer to the following articles: 1. DID use classic literature, compulsory license: evidence from the enemy's trade law, 2. continuous DID classic literature, potatoes made the old world civilization, 3. cross section Data DID tells the paradigm of double-differential policy evaluation in cross-section, 4. RDD classic literature, RDD model validity and robustness test, 5. Event research method used in DID classic literature "environmental regulation" thesis data and procedures, 6. Generalized The DID method is very classic JHE literature, 7. DID’s classic literature "compulsory license" paper data and do program, 8. MLM activities on economic development, AER cross-sectional data analysis classic text, 9. Multi-period DID classic Literature big bad banks data and do files, 10. Causal inference IV method classic literature, is it system or human capital that promotes economic development? , 11. The establishment of causality on AER, sensitivity testing, heterogeneity analysis and cross-data use classic articles, 12. The second classic of causal inference, the impact of work interruption on the subsequent productivity of workers? 13. Density economics: natural experiments from the Berlin Wall, the best Econometrica paper, 14. Labor and health economics using DID and DDD as identification strategies on AER, 15. A policy evaluation method using cross-sectional data, can also be issued AER, 16. Multi-period DID model classic literature, big bad banks explain ",", 17. Multi-period DID classic literature big bad banks data and do files, 18. Nonlinear DID, double transformation model CIC, quantile DID,
19. What is Fuzzy DID? How to use data to achieve it? 20. Multi-issue DID big bad banks Chinese translation version and detailed explanation, 21. DID industry/region and time trend interaction items, common trend test, dynamic policy effect test, etc. 22. Cross-sectional data DID operating procedure guide, Teach you step by step, 23. DID research dynamics and literature review applied in policy evaluation, 24. Continuous DID classic literature, potatoes made the civilization of the old world, 25. DID double difference method, some error-prone places, 26 Continuous DID, DDD and proportional DID, unobservable selection bias, 27. Weighted DID, an encyclopedia of IPW-DID empirical procedures, 28. DID and DDD, a concise introduction, double and triple difference models, 29. DID process The map display skills summarized in 30. The parallel trend hypothesis test program of DID and other uses of coefplot, 31. Cross-sectional DID, various fixed effects, placebo test, displacement test, other external shock treatment, 32. Double in practice The assumptions implied by the difference method DID, 33. In the past thirty years, the "high light moment" road map of RCT, DID, RDD, LE, ML, DSGE and other methods, 34. The first DID method analysis by an academician of metrology, China closed the city to the new crown The impact of the spread of the virus! and so on.
The following short-linked articles belong to a collection, you can collect them and read them, or you won't find them in the future.
In 2.5 years, nearly 1,000 non-weighted measurement articles in the econometric circle,
You can search for any measurement related issues directly in the official account menu bar,
Econometrics Circle