1 / 10

Impact Evaluations in Good Times and Bad

Impact Evaluations in Good Times and Bad. Forum Kajian Pembangunan March 22, 2011 Firman Witoelar , DECHD, Discussant. Issues. Selection biases Spillovers and hidden/unintended outcomes Timing of impacts Data requirements. Selection biases. - Non-random program placement

huong
Download Presentation

Impact Evaluations in Good Times and Bad

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Impact Evaluations in Good Times and Bad Forum Kajian Pembangunan March 22, 2011 FirmanWitoelar, DECHD, Discussant

  2. Issues • Selection biases • Spillovers and hidden/unintended outcomes • Timing of impacts • Data requirements

  3. Selection biases - Non-random program placement - Selection bias: reasons to participate in a program are correlate with outcomes of interest • Researchers may not know ‘reasons’ but have data on observables • Is there selection on observables? • What is the direction of the bias? • Choose comparison group carefully based on observables • Can also be selection based on unobservables… • DID may help if unobservables are time invariant • The two biases can work in different directions

  4. Spillovers and “hidden outcomes” • Spillovers • may underestimate program impacts if comparison group is contaminated • hard to deal with due to • market responses • government responses (e.g . local government)

  5. Spillovers and unintended outcomes • Unintended outcomes: • Examples • Employment Guarantee Scheme (Maharashtra, India) • Work is guaranteed at low wage rate: thought to be self-targeted • However, likely to spill to private labor market • No one want to work below EGS wage: wages will be the same between participants and non-participants • Social insurance (e.g. Jamkesmas) • Outcome of interest: program take-up /coverage • But…w ill a universal social insurance lower the take up of employee-provided insurance?

  6. Timing of impacts • When are the programs expected to have impacts? • Short-term or long term impact? • Lasting or dissipating impacts? • Exit strategies: • When programs are phased out, will behavior change?

  7. Data requirements • Data collection should be built-in in the project design and evaluation design (e.g. PKH/CCT) • Same survey instruments administered for program participants and non-participants • Collect well defined outcome measures: self-reported, official records, physical measures • Collect enough information (individuals, household, communities) to deal with heterogeneity • Cover the time period over which the projects are expected to have impacts

  8. Data requirements (continued) • Detailed information about the programs: • institutional background • timing of the programs • program eligibility • other programs that are operating in the communities • Panel data may be desired: • Comparability of survey instruments • Attrition is important: absence of patterns in observables no guarantee (Witoelaret al, forthcoming)

  9. Other examples • Frankenberg , Suriastini, Thomas (2005) – BidanDesa program • 1989 , placement of 50,000 “BidanDesa” • non-random placement • Study exploits: • timing of placement (similar to the Posyandu paper) • anthropometric measures • rich socio-economic panel data • Giles, Satriawan (2011) - post-crisis food supplementation program (PMT) • Study exploits: • communities’ exposure to the program • variation in child age and program eligibility • anthropometric measures • rich socio-economic panel data

  10. On RCT: …also check out current edition Boston Review (March/April 2011) “Small Changes, Big Results” - Glennerster and Kremer (JPAL) arguing for applying experiments and behavior economics in global development) PranabBardhan, Jishnu Das, and others discuss http://www.bostonreview.net/

More Related