GMU econ PhD student, liberal, aspie, bi. I post interesting papers. Michael Kremer stan. I ❤️ optimal auction design. Spend more on drugs. Open borders now!
Their empirical strategy is to only compare cousin marriage children to their cousins who were born from out of the family marriages. They check that the marrying cousins aren’t different either by showing that they live just as long as their siblings. 
We can calculate it by figuring out the price of a good which would set demand to 0, and then presume that this shadow price fell when it came into being. To know this, we must estimate the demand curve. 
Tishara Garg is studying the effects of industrial parks in India. These are plots of land set aside for industrialization, with substantial government promotion. These are not, to be clear, special economic zones like Shenzhen, but simply local initiatives. 
Readers will note that the structure of the paper is rather similar to Gentzkow and Shapiro (2010), who studied media bias. In both cases, the first challenge is you need a treatment variable with cardinality.
The previous system worked like this: all the food banks would be in a queue. Whenever Feeding America got a donation of a truckload of food, they would go to the first bank in line. They would then have 4-6 hours to accept or decline the shipment.
Measuring the effects of management is quite difficult, which is why they do this in the lab. Managers aren’t randomly assigned, which may — especially if managers are promoted for reasons only somewhat correlated with skill — flip the sign on the true effect. 
First, though, we should know what GDP is trying to measure. It is not measuring welfare — for that would take knowledge of how pleasant work is — but simply trying to give a sense of the resources available for consumption.
I should be clear about the terms. The interest rate is simply what the rate we need to be paid in order to defer consumption. If interest rates are 6%, then I am indifferent between $100 and $106 in a year. (We will abstract away from risk-aversion here)
The “stationary bandit” view of the state derives from Mancur Olson. In the state of anarchy, some people are bandits who profit by stealing from others. This theft prevents people from increasing output, though. It is profitable for the bandit to settle down and impose taxes.
India’s infrastructure, before the railroad construction campaigns began, was extremely poor. Transport was main through putting packs on the back of cattle, and driving them through fields, which was incredibly expensive. Railroads were both much cheaper, and much faster.
Recall that, in the Solow model, output is a function of labor and capital, with it converging to a steady state. Technological progress is represented by an A multiplying the function. Capital has decreasing marginal returns.
Its predecessor is the Harrod-Domar model, which is essentially a port of short-run Keynesian economics into long run growth. It assumes that factors of production are not substitutable, which gives it an unpleasant “knife-edge” characteristic.
I am not discussing it so much for the empirical results as for the methods, but it follows in the footsteps of Krugman (1981) and Brander and Krugman. If you allow for frictions, trade will create winners and losers. How big are these?
They are able to find the productivity effects easily, because that’s input/output, and they have sufficiently detailed data to measure that. They then turn to whether this is passed through to wages, and estimate a very precise null.
People in the same role differ in productivity. Pay transparency makes it much harder to pay more productive workers more, because you will need to give a formal promotion. 