The Topalova paper's estimates are based on a sample survey that was not designed to generate estimates at the disaggregated level she uses in her analysis. People need to stop citing it.
Her paper is completely dependent on district level poverty estimates, which the NSSO survey that she uses is just not set up to produce. If she really believes in the methodology, there are two large follow up rounds available for analysis - 2004-05 and 2011-12, which are the rounds that show large scale poverty reductions in India. It would take a week or two of work at most, less if you have a competent RA. All the same code will work. She has not bothered.
See here for instance for a citation that NSSO surveys are not set up for district level estimates of poverty rate -
‘But, in NSSO surveys, sample sizes are fxed in such a way that, reliable direct survey estimates (i.e. estimates obtained using domain-specifc data and through traditional survey estimation method) can be obtained for planned domains (e.g., state) only, whereas sample sizes for unplanned domains (e.g., districts, taluk, block, municipalities, gram panchayats) either going to be very small or even zero for some domains….Chaudhuri and Gupta (2009) have rendered the district level poverty estimates of India using HCES 2004–05 data of NSSO. However, the study acknowledged the limitation of large standard error observed for some districts which being the fundamental problem in determining local level estimates due to sample size problem’
Nice piece Nicholas Decker. This reminds us that economics is complicated, and the effects of tariffs or trade liberalization are difficult to foresee.
As you note, however, these words of caution are not a call to throw up trade barriers.
The data shows that, on balance, tariffs, and trade restrictions are net negatives for everyone.
Many don’t realize, for example, that tariffs on imports function as tariffs on exports; depressing both at the same time. You cannot “fix” a trade deficit (no should we really want to) with tariffs.
The Topalova paper's estimates are based on a sample survey that was not designed to generate estimates at the disaggregated level she uses in her analysis. People need to stop citing it.
I’ll reread it. May I have more detail please?
Her paper is completely dependent on district level poverty estimates, which the NSSO survey that she uses is just not set up to produce. If she really believes in the methodology, there are two large follow up rounds available for analysis - 2004-05 and 2011-12, which are the rounds that show large scale poverty reductions in India. It would take a week or two of work at most, less if you have a competent RA. All the same code will work. She has not bothered.
See here for instance for a citation that NSSO surveys are not set up for district level estimates of poverty rate -
https://www.mospi.gov.in/sites/default/files/publication_reports/Estimation%20of%20Disaggregate%E2%80%91Level%20Poverty%20Incidence.pdf
‘But, in NSSO surveys, sample sizes are fxed in such a way that, reliable direct survey estimates (i.e. estimates obtained using domain-specifc data and through traditional survey estimation method) can be obtained for planned domains (e.g., state) only, whereas sample sizes for unplanned domains (e.g., districts, taluk, block, municipalities, gram panchayats) either going to be very small or even zero for some domains….Chaudhuri and Gupta (2009) have rendered the district level poverty estimates of India using HCES 2004–05 data of NSSO. However, the study acknowledged the limitation of large standard error observed for some districts which being the fundamental problem in determining local level estimates due to sample size problem’
Nice piece Nicholas Decker. This reminds us that economics is complicated, and the effects of tariffs or trade liberalization are difficult to foresee.
As you note, however, these words of caution are not a call to throw up trade barriers.
The data shows that, on balance, tariffs, and trade restrictions are net negatives for everyone.
Many don’t realize, for example, that tariffs on imports function as tariffs on exports; depressing both at the same time. You cannot “fix” a trade deficit (no should we really want to) with tariffs.
Furthermore, as I discussed here, trade appears, more than anything, to reduce the risk of war: https://www.lianeon.org/p/in-defense-of-free-trade