Let me sketch a couple of thoughts.
1. Is Amazon a villain? I don't think so. Professor Krugman knows the Dixit and Stiglitz taste for variety model. Amazon has sharply increased all of our access to variety. This increase in consumer surplus should not be ignored.
2. The weakness in Dr. K's piece is that he ignores basic ideas in labor and urban economics. The land used by an urban mall has an opportunity cost. Twitter can buy it and convert it to a higher productivity use. But, if a coal mine shuts because of Federal regulation; that land won't be used for any other purpose. Department stores and suburban malls have a much higher opportunity cost than rural mining sites. Krugman ignores this basic point.
3. Dr. K ignores local labor market competition. In cities (relative to rural areas), workers have a much thicker local labor market. In English, if USC fired me tomorrow --- I can work as an academic economist in LA working for other universities or private employers. They same holds for an urban retailer. While she may not find another job in retail, there are always other opportunities to earn in a thick urban labor market. By "thick", I mean perfectly competitive market. In rural areas where the mining takes place, there is more likely to be monopsony in the labor market with a single buyer of labor such as a coal mine. In this case, the newly unemployed miners have nowhere to turn if they want to remain (and they are stuck there because they are home owners). The retail workers in the city can be renters in multi-family buildings. This last point is a key point of my new paper with Jonathan Eyer.
Prolonging Coal’s Sunset
In recent years, the share of U.S electricity generated by coal has fallen from nearly 50% to 33%. The costs of this transition are spatially concentrated, and mining states have already lost income due to the reduced demand for coal. Coal states have enacted policies to encourage local power plants to purchase from within state mines. We document that power plants in states and counties with substantial mining activity are more likely to be coal fired and to purchase more within political boundary coal. These results are robust to including flexible controls for the distance from power plants to mines. While coal states benefits from local protectionism, these efforts impose social costs because coal mining and coal burning creates significant environmental consequences. We quantify these effects and find that a one-percentage point increase in the proportion of coal plants in a NERC region with an in-state coal mine results in approximately 2.3 million additional annual tons of CO2 emissions.