Taxing Property in Developing Countries

Juan Carlos Suárez Serrato (Duke)

Abstract:

We leverage administrative data, multiple tax changes, and a field experiment to analyze the administration of property taxation in Mexico City – a developing country context with household credit constraints and limited administrative capacity. Using RD and DiD designs that leverage quasi-experimental tax rate increases, we estimate that a 10% increase in the tax rate increases revenue by 3-7%, while also increasing the rate of non-compliance. Using a field experiment, we show that reminder letters to non-compliers increase tax payment significantly.

While it is possible to raise tax revenue through tax increases or enforcement,doing so may not be optimal in a context with liquidity constraints. We provide three pieces of evidence showing that liquidity constraints impact property tax compliance. First, tax rate increases raise the share of taxpayers that pay in installments or pay late, instead of paying the full liability at once. Second, we leverage variation in early payment discounts to estimate a dynamic discrete choice model, which shows that households have a high value for liquidity.Third, we use quasi-random tax increases to instrument for property tax payments and show that consumers experience consumption decreases when they pay property taxes.

Finally,we combine these empirical results with a model of optimal property taxation. Because liquidity constraints increase the welfare costs of taxation, the model identifies the provision of loans to taxpayers with liquidity constraints as an important and hitherto neglected policy instrument.