Inference in differences-in-differences with few treated groups and heteroskedasticity

Differences-in-Differences (DID) is one of the most widely used identification strategies in applied economics. However, how to draw inferences in DID models when there are few treated groups remains an open question. We show that the usual inference methods used in DID models might not perform well when there are few treated groups and errors are heteroskedastic. In particular, we show that when there is variation in the number of observations per group, inference methods designed to work when there are few treated groups tend to (under-) over-reject the null hypothesis when the treated groups are (large) small relative to the control groups. This happens because larger groups tend to have lower variance, generating heteroskedasticity in the group x time aggregate DID model. We provide evidence from Monte Carlo simulations and from placebo DID regressions with the American Community Survey (ACS) and the Current Population Survey (CPS) datasets to show that this problem is relevant even in datasets with large numbers of observations per group. We then derive an alternative inference method that provides accurate hypothesis testing in situations where there are few treated groups (or even just one) and many control groups in the presence of heteroskedasticity. Our method assumes that we can model the heteroskedasticity of a linear combination of the errors. We show that this assumption can be satisfied without imposing strong assumptions on the errors in common DID applications. With many pre-treatment periods, we show that this assumption can be relaxed. Instead, we provide an alternative inference method that relies on strict stationarity and ergodicity of the time series. Finally, we consider two recent alternatives to DID when there are many pre-treatment periods. We extend our inference methods to linear factor models when there are few treated groups. We also derive conditions under which a permutation test for the synthetic control estimator proposed by Abadie et al. (2010) is robust to heteroskedasticity and propose a modification on the test statistic that provided a better heteroskedasticity correction in our simulations.

Inference in differences-in-differences with few treated groups and heteroskedasticity

Differences-in-Differences (DID) is one of the most widely used identification strategies in applied economics. However, how to draw inferences in DID models when there are few treated groups remains an open question. We show that the usual inference methods used in DID models might not perform well when there are few treated groups and errors are heteroskedastic. In particular, we show that when there is variation in the number of observations per group, inference methods designed to work when there are few treated groups tend to (under-) over-reject the null hypothesis when the treated groups are (large) small relative to the control groups. This happens because larger groups tend to have lower variance, generating heteroskedasticity in the group x time aggregate DID model. We provide evidence from Monte Carlo simulations and from placebo DID regressions with the American Community Survey (ACS) and the Current Population Survey (CPS) datasets to show that this problem is relevant even in datasets with large numbers of observations per group. We then derive an alternative inference method that provides accurate hypothesis testing in situations where there are few treated groups (or even just one) and many control groups in the presence of heteroskedasticity. Our method assumes that we know how the heteroskedasticity is generated, which is the case when it is generated by variation in the number of observations per group. With many pre-treatment periods, we show that this assumption can be relaxed. Instead, we provide an alternative application of our method that relies on assumptions about stationarity and convergence of the moments of the time series. Finally, we consider two recent alternatives to DID when there are many pre-treatment groups. We extend our inference method to linear factor models when there are few treated groups. We also propose a permutation test for the synthetic control estimator that provided a better heteroskedasticity correction in our simulations than the test suggested by Abadie et al. (2010).

The Taylor rule under inquiry: Hidden states

This work empirically evaluates the Taylor rule for the US and Brazil using Markov-Switching Regimes. I find that the inflation parameter of the US Taylor rule is less than one in many periods, contrasting heavily with Clarida, Gal´ı and Gertler (2000), and the same happens with Brazilian data. When the inflation parameter is greater than one, it encompasses periods that these authors considered they should be less than one. Brazil is used for comparative purposes because it experienced a high level inflation until 1994 and then a major stabilization plan reduced the growth in prices to civilized levels. Thus, it is a natural laboratory to test theories designed to work in any environment. The findings point to a theoretical gap that deserves further investigation and show that monetary policy in Brazil has been ineffective, which is coherent with the general attitude of population in relation to this measure.

Taylor rule with hidden states

This work evaluates empirically the Taylor rule for the US and Brazil using Kalman Filter and Markov-Switching Regimes. We show that the parameters of the rule change significantly with variations in both output and output gap proxies, considering hidden variables and states. Such conclusions call naturally for robust optimal monetary rules. We also show that Brazil and US have very contrasting parameters, first because Brazil presents time-varying intercept, second because of the rigidity in the parameters of the Brazilian Taylor rule, regardless the output gap proxy, data frequency or sample data. Finally, we show that the long-run inflation parameter of the US Taylor rule is less than one in many periods, contrasting strongly with Orphanides (forthcoming) and Clarida, Gal´i and Gertler (2000), and the same happens with Brazilian monthly data.

Election over, battle for Brazil rages

Last week I sat down with a Brazilian acquaintance who was shaking his head over the state of national politics. A graduate of a military high school, he'd been getting e-mails from former classmates, many of them now retired army officers, who were irate over the recent presidential elections. "We need to kick these no-good Petistas out of office," one bristled, using the derogatory shorthand for members of the ruling Workers Party, or PT in Portuguese.