Research
WORKING PAPERS
We determine if the communication of private banks to their clients with financial interests in Mexico changes or not after Mexico's Central Bank communicates its monetary policy decision (MPD) and also two weeks later, with the publication of the minutes of Mexico's Central Bank monetary policy decision (MMPD) between 2011 and 2019. We use unsupervised Natural Language Processing (NLP) techniques to turn the text that private banks send to their clients about the Mexican economy into vectors of topics. We find that every time, private banks cover a large diversity of topics and words before the MMPD with no evident consensus of topics, and that almost always the quantities of terms and topics are reduced and repeated by almost every bank after the MMPD indicating some surprise (notable exception: the liftoff in December 2015), and that the topics vary depending on the date of the MMPD. The fact that private banks discuss the same topics and write to their clients with sentences that contain the exact same words indicates that the private banks react to the MMPD, independent of their opinion about the Central Bank's statements. We also found weak evidence that a measure of the size of the changes in the private bank's communication with their clients is positively correlated to changes in the long-term yields but negatively correlated to the size of exchange rate movements.
I develop a structural model with nested CES preferences to obtain optimal markups for heterogeneous retailers when the prices of all their inputs are exogenous. The model predicts that if the taste parameters are constant over time, the markups for retailers with higher market share are higher but have more flexibility, implying an incomplete pass-through of retailer input price onto final retailer prices. I then focus on the exchange rate pass-through (ERPT) and use a unique data set of all the price changes of tradeable merchandise in the Mexican Consumer Price Index (CPI) data by store type to test the model. I find, consistent with the model, that (1) ERPT is different by store type; and (2) products sold in stores with negligible market share have the same ERPT regardless of the store type. Both results imply that the ERPT is estimated with bias when the store information is not used.
I use the recently published Muestra Especial del Censo de Población y Vivienda 2010 (2 million homes) and the Encuesta Intercensal 2015 (6 million homes) and slightly modify the model of Monte et al. (2015) to estimate how do workers in Mexican municipalities choose the location of their workplace based on the income gains from commuting. Estimates are very much in line with the intuition: Static estimates for both 2010 and 2015 suggest that those who commute earn an average 30 percent more than their non-commuting counterparts, that the distance traveled is positively correlated with the average income gain, and that commutes tend to be to municipalities located close to the place of residence. Comparing both years suggests that a decrease in local productivity both decreases the number of workers that come from other municipalities with less negative change in productivity and increases the number of local residents that decide to work somewhere else, mitigating the negative effect of the reduction in local wages with higher earnings from the new destinations. I find that some municipalities were not able to mitigate the negative productivity shocks on their income.
What was the impact of railroads in the output of the United States during the 19th century and how can a New Trade model help answer this question? In order to respond I follow three steps. First, I construct a new digital railroad data set and pair it with geographic and topographic features of the U.S. territory to estimate travel times between every pair of U.S. counties for every year between 1840 and 1900. Second, I use these results, together with a Ricardian model of trade and U.S. county output data from the 19th century, to estimate gains from trade using a fixed-point algorithm. Third, I estimate counter-factuals. My estimates suggest that, leaving all factors of production fixed, if the railroads were made suddenly unavailable in 1900 there would have been a 9% reduction in output. Last, I find that removing the most recent observed railroads has a larger effect than adding the most immediate observed railroads. This means that there could have been targeted migration anticipating railroad construction, and not necessarily that railroad construction is the result of observed migration.
PUBLICATIONS
Short-Run Market Access and the Construction of Better Transportation Infrastructure in Mexico (with Aldo Sandoval) Economía 18, October 2017
We calculate the short-run effect that the construction of the 230km-long Durango-Mazatlán highway in the end of 2013 and of the 290km-long Mexico City-Tuxpan highway in the beginning of 2014 produced on welfare in every municipality as well as in market-access in every location of Mexico. Our estimates suggest that the former highway produced benefits not only in the region where the new highway is located, but in vast regions in the north of the country. Analogous estimates show that the latter highway mostly benefited regions near Tuxpan, but these focalized benefits were larger than any of the benefits derived from the construction of the Durango-Mazatlán highway. The municipalities in the south of the country have net short run losses from the infrastructure construction due to losses in competitiveness. Our model is consistent with the observed sectoral growth in Sinaloa, Durango, and Veracruz in the year 2014. Qualitatively, market access and welfare change in the same direction and magnitudes, so this paper recommends to use the market access approach when doing short-run analysis of infrastructure, because it is much less computationally intensive.
Trade and Carbon Taxes (with Joshua Elliott, Ian Foster, Samuel Kortum, Todd Munson, and David Weisbach) American Economic Review Papers and Proceedings 100, May 2010
We study various scenarios for taxing emissions of carbon dioxide (CO2). The question is how carbon tax policies will perform, given international trade, if countries adopt different tax rates. We investigate this question quantitatively using CIM-EARTH, a newly developed open-source computable general equilibrium (CGE) model.
OTHER WORK IN PROGRESS
Estimating the Value of Time in Mexico City (with Óscar Cuéllar)
We use the Encuesta Intercensal 2015 (5.9 million homes representative at the municipal level of the 33 million homes in Mexico) to estimate the average monetary compensation for an extra hour of commuting for the workers in Mexico City, or the value of their time. Mexico City is the home for 9 million workers, but the place of work for 10.9 million workers. Data shows that 42 percent of the residents commute to another municipality for an average gain of 39 percent in wages. Workers from other states that commute to Mexico City gain around 30 percent for doing so. However, we find that salaries are not the only explanation for the commute. There is an amenities component to working, which we find to be strong in Mexico City. This component attracts workers despite the wage. Controlling for amenities and wages, we find that the average compensation for an extra 40 minutes of commute every day, each way (around 8 hours per week) is a 30 percent increase in wages.
A Quantitative Examination of Trade and Carbon Taxes (with Joshua Elliott, Todd Munson, Samuel Kortum, and David Weisbach)
We examine the impact of carbon leakage on a regional carbon tax using a newly developed, open-source computable general equilibrium model, CIM-EARTH. We consider a business as usual scenario with high emissions because of rapid economic growth and slow adoption of clean energy. Relative to our baseline scenario, a production tax in Annex B countries reduces emissions only by 1/3 of the emissions reductions that would be achieved under a global carbon tax. This due to the importance of emissions in developing countries in the future rather than carbon leakage. Carbon leakage - the increase in emissions in non-taxing regions relative to emissions reductions in taxing regions - ranges from 15% to 25%, depending on the tax rate. Border tax adjustments eliminate about half of the leakage.