Presentation on theme: "1 Where Do We Go From Here? Wrap up and Next Steps Stefan Voigt Philipps-University Marburg World Bank Seminar on “The Empirics of Governance”, May 1 -"— Presentation transcript:
1 Where Do We Go From Here? Wrap up and Next Steps Stefan Voigt Philipps-University Marburg World Bank Seminar on “The Empirics of Governance”, May 1 - 2, 2008
Voigt: How to Measure Institutions2 Alan suggested 7 „discussion-worthy“ questions for day 1. Here are some possible answers that emerged: Question 1: Is the Conceptual Basis of the respective indicator sound? –Creators of WGI: yes, we even document measurement error! –My impression: The creators are not interested in a conceptual basis in the sense of „theoretical“ basis. Their presentations sometimes gave the impression: „We have data – what do we need a theory for?“ The absence of an „Agreed Unified Theory“ does not include the permission to do no matter what As a researcher I start with my own theory – and then I (or better: others) analyze whether it can be refuted by the data. Complete lack of theory makes refutation impossible
Voigt: How to Measure Institutions3 Question 1: Is the Conceptual Basis of the respective indicator sound? (contd.) –Trying to switch the burden of proof: „But do conceptual /definitional differences make a big difference?“ We will never know if we do not invest in more precise, theory-based indicators! Danger (pointed out by Juanita Raiño in different context): focus on indicators that have the virtue of being available or that can be made available at low cost might lead us to analyze wrong aspects (the famous drunk…)
Voigt: How to Measure Institutions4 Question 2: How to Deal with „Institutions“ vs. „Outcomes“? –If we decide to start with a theory (which we should!) and posit that „institutions“ determine „outcomes“, we obviously need both. And the two need to be kept meticulously apart!
Voigt: How to Measure Institutions5 Question 3: Focus on „Formal“ and / or „Actual“ Rules? –Definitely both! –Measuring both allows us to deal with a number of different questions: (1)Are there different solutions for identical problems that work similarly well? (substantial aspect) (2)How important is the (factual) uncertainty reduction brought about by institutions (which can only be calculated as the difference between formal and factual) Steve Knack: Global Integrity Index calculates difference between „in law“ and „in practic“ and regresses on gdp/cap.
Voigt: How to Measure Institutions6 Question 4: Subjective vs. Objective Indicators? –We definitely need both – but we need to be clear about what kind of indicator is best suited for what purpose. –And: New research questions on relationship between subjective and objective indicators deserve to be dealt with E.g.: What are the objective factors that determine the perception of some relevant governance dimension; competing theories can be put to a horse race here (but there might be problems of multicollinearity). Svensson mentioned some: how to explain the differences between self-reported and objective health- data?
Voigt: How to Measure Institutions7 Question 5: How about Robustness? –At least two ways to interpret that question: (1)How robust is the construction of the indicator regarding inclusion/exclusion of specific variables? (2)How robust are the results that we reach by drawing on the respective indicator. –Dorotinsky: PEFA focuses on central government (no sub- national governments, no SOEs) Results might not be robust – in particular in federal countries. –Gomez: complementarity of indicators might also be useful as a check on robustness of results
Voigt: How to Measure Institutions8 Question 6: How about Aggregation? –Marshall: Institutions per se do not matter; it is the combination that does Implication 1: Identify single institutions that are/ could be at work. Impliciation 2: Identify the „institutional environment“ that makes them work / fail. Could be coined the business of „unbundling“ and „rebundling“ Relevance of interaction effects etc.
Voigt: How to Measure Institutions9 Question 7: How to Design Reform? –Turning Marx from his head to his feed: until now, we have always tried to change the world, now it is time to explain it! –We need to have robust cause-effect-relationships; premature attempts at reform can cause huge costs. –Relevance of informal/internal institutions should always be recognized as explicitly as possible (many colonizers and revolutionaries have failed because they did not…) Gelb: actionable vs. action-worthy Levy: sequencing in policy steps.
Voigt: How to Measure Institutions10 Further Questions: –Keefer: Endogenize Institutions! –Murrell: Bank as repository of good questions Bank as facilitator in collecting data that is needed to construct indicators (lots of expertise both in terms of what is factually going on almost everywhere – but also in knowing the relevant data sources).
Voigt: How to Measure Institutions11 Thank you for your attention!