Experiments and Natural Experiments

Advantage of tax reform policies can affect some

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: n the US and WTFC in the UK) have received lots of attention (Blundell and Hoynes, 2004). Figure 1a: 1987 Tax Holiday in Iceland Employment Rate 95% 90% 85% 80% 75% 1982 1984 1986 1988 Year Empirical (Bianchi et al. 2001) 1990 1992 Fortin – Econ 560 Lecture 1C o A difficulty here is that single mothers may benefits from a combination of transfers and in-kind programs. [see Blundell and Hoynes’ diagram] o More generally, it is unclear that participants are always informed about the changes in the features of the benefit schedule (Chetty and Saez, 2009) 1) A first approach is a difference-in-difference methodology that compares the average behaviour before and after the reform for the eligible group with the before and after contrast for the comparison group. This approach removes unobservable individual effects and common macro effects to recover the average effect of the reform, under the assumptions of common time effects across groups and no composition changes with the group. But this is not without problems. For example, Eissa (1995) uses the Tax Reform Act (TRA) of 1986 to identify the effect of MTRs on labor force participation and hours of married women. This tax reform TRA 1986 cut top income MTR from 50% to 28% from1986 to 1988 but did not significantly change tax rates for the middle class. Substantially increased incentives to work of wives of high income husbands relative wives of middle income husbands. Source: Federal Govt Fortin – Econ 560 Lecture 1C How is this potentially related to the proposal of “income splitting” in Canada? Difference-in-Difference Methodology in a Nutshell Step 1: Simple Difference Outcome: LFP (labor force participation) Two groups: Treatment group (T) which faces a change [women married to high income husbands] and control group (C) which does not [women married to middle income husband] Simple Difference estimate: D = LFPT - LFPC captures treatment effect, if in the absence of treatment, LFP equal across 2 groups Note: this assumption always holds when T and C status is randomly assigned. To test for this assumption, we can compare LFP before the treatment: = − . Fortin – Econ 560 Step 2: Difference-in-Difference (DD) If ≠ 0, we can estimate DD: = − =[ − where A = after reform, B = before reform Lecture 1C −[ − DD is unbiased if the parallel trend assumption holds: absent the treatment, the difference across T and C would have stayed the same before and after DD can be estimated by OLS to control for additional covariates = After + Treat+γAfter*Treat+ + It is easy to show that γ = . DD most convincing when groups are very similar to start with [closer to randomized experiment] Should always test DD using data from more periods and plot the two time series to check parallel trend assumption [e.g. Imbens et al.] Use alternative control groups [not as convincing as potential control groups are many] In principle, can create a DDD as the difference between actual DD and DDPlacebo (DD between 2 control groups) Fortin – Econ 560 Lecture 1C Results in Eissa (1995): 1) Participation elasticity around 0.4 but larg...
View Full Document

Ask a homework question - tutors are online