The Challenge of the Learning Deficit: Presentation to Copenhagen Consensus Expert PanelOutline of the presentationC-B analysis of public sector interventionsPlausible counter-factual: Market outcomesPlausible counter-factual: Sectoral Engineers versus EconomistsPlausible Counter-Factual: Example of MICE (WDR 93 numbers) vs PSCE (WDR 1993 Economics)Plausible counter-factual: MICE (and its ilk) are wrong, wrong, wrongPlausible Counter-factual: Demand for educationThere are substantial (but not spectacular) private returns to educationIs demand for education too low because of externalites?Measuring “schooling capital”, from general to specificKey question: how does return vary with stockThe naively estimated relationship between growth of schooling capital and growth does not suggest “externalities”Two problems with torturing the data to make it confess to output externalities to schoolingBut other assumptions can produce other results—but never strong support for large externalitiesVariation in growth of SK (lower case, red) is small (between 1 and 3 ppa) compared to variation in growth of output…which is not true of “physical capital” (actually, CUDIE)Summary on Economic ExternalitiesAre there non-economic externalities?Plausible counter-factual: Credit constraintsThe gap in attainment by wealth varies widely around the worldC-B of public sector requires…Latest “institutional economics”?Implausible Positive Theory of Public Sector InterventionAlternative positive theory of government production of schooling: Production for ideological control (PIC)NAP explains nothing, PIC everything: OneNAP explains nothing, PIC everything: TwoNAP explains nothing, PIC everything: ThreeNAP explains nothing, PIC everything: FourWhat is an “opportunity”?Slide 31The five opportunitiesSlide 33Adequate resources, autonomy to decision makers, accountability for performanceThis does not mean ‘vouchers’ it means resources, autonomy, and accountability.Not “what” but “how”Schools that serve citizen’s (and particularly the poor’s) interestsAre your interests being served?Experience has led me to focus on improving mechanisms of accountability as the principal “opportunity” in most countriesThe Challenge of the Learning Deficit:Presentation to Copenhagen Consensus Expert PanelLant PritchettKSG and Center for Global DevelopmentMay 28th, 2004Outline of the presentation•What I did not do, and why•C-B analysis without a counter-factual•Assume public sector solutions•Assessing the five opportunities•System expansion•Budget expansion (radial, directed)•Expanding the demand (raising returns)• Financing (lowering costs)•System reformsC-B analysis of public sector interventions•C-B analysis can compare the NPV of incremental costs and benefits of a public sector action (e.g. project, policy, expenditure)This requires, at a minimum:•Plausible positive theory of the counter-factual—and “nothing” is not plausible•Plausible positive theory of public sector intervention—and “perfect” is not plausiblePlausible counter-factual: Market outcomes•Normative rationales for public sector action:–Improve efficiency–Improve equity–Or both•Insights of the two welfare theorems of “market” equilibrium has economists looking for “market failures” as rationales.Plausible counter-factual: Sectoral Engineers versus Economists“Engineers” (doctors, educators)EconomistsAssess the impact of an action on physical outcomes against cost(cost-effectiveness)—no action is the counter-factual (or more particularly not positive theory of existing outcomes)Assess the impact of an action on welfare (well-being, utility) against cost (full on CB)—the counter-factual is a positive theory with optimizing agents.Plausible Counter-Factual: Example of MICE (WDR 93 numbers) vs PSCE (WDR 1993 Economics)MICE (Medical Intervention Cost Effectiveness)—WDR 93 NumbersPSCE (Public Sector Cost Effectiveness)—WDR 93 EconomicsRank set of medical interventions (e.g. taking an aspirin, heart surgery, vector control) in terms of health gain (DALYs) per dollar (cost effectiveness) and define a “basic package”Rank public sector interventions (e.g. subsidize treatment, regulate water supply, anti-smoking campaigns) per DALY to define set of priority actions.Plausible counter-factual: MICE (and its ilk) are wrong, wrong, wrong•MICE is wrong analytically, in every conceivable way–Compares public goods and private goods–C-E ratios normalize out the key feature of health markets•MICE is wrong empirically–Micro: zero displacement effects are wrong–Macro: no evidence at all in the aggregate data•MICE is wrong practically–Assumes the question is “what” not “how”(Aside: The WDR 93 was made incoherent during the review process as a compromise)Plausible Counter-factual: Demand for education•People (and parents acting altruistically for minor children) seek out education purposively to maximize their well-being (across all dimensions) subject to constraints.•There are a variety of potential “market failures”–Externalities–Credit constraintsThere are substantial (but not spectacular) private returns to educationIs demand for education too low because of externalites? •Levels and differences of the Mincer returns completely irrelevant to public policy (directly and indirectly) unless they interact with market failures. (e.g. stocks versus bonds, large cap equity versus small cap equity)• If there are positive external effects to schooling on output then the aggregate effect (macro) should exceed the sum of the micro-economic effects.TadataLtahtH ),(),()()()()(),(sagsfenahtahMeasuring “schooling capital”, from general to specificKey question: how does return vary with stockGeneral form for f() in equation above:Ψ=.58(regression based—see figure above)Returns decline with stockΨ=0(assumption)Returns are unchanged with stock(with other assumptions)(dSK/ds=r)11)( ssfssf /)( The naively estimated relationship between growth of schooling capital and growth does not suggest “externalities”Two problems with torturing the data to make it confess to output externalities to schooling•The problem with most less developed countries is that economic growth has been too low—not “too high”—the growth residual without attributing anything to schooling is very near