2-1Tool of Public Financeopositive economicsoHow the economy worksoWhatoWhenoWhyonormative economicsoWhether or not it is producing good resultoGood/ Better off oBad/ Worse offoWhat ought to be2-2提问，请准备：o请用预算图表示substitution effect和 income effect，并说明normal/inferior /Giffen good。o请用收入图示意normal/inferior good。o请用收入图示意necessaries 和luxury good.o请用数学表达式说明上述概念。o完全竞争市场（perfect competition）应该具备哪些条件？McGraw-Hill/IrwinCopyright 2008 by The McGraw-Hill Companies, Inc. All rights reserved.CHAPTER 2Tools of Positive Analysis2-4The Role of TheoryoEconomic theory Economic theory helps specify the factors that might affect a given kind of behavior. But, theory alone cannot say how important any particular factor is.oThe assumption that maximization of utility is taken throughout the book.2-5The Role of TheoryoEconomic modelsnvirtue of simplicitynjudging a modelnlimitations of modelsoEmpirical analysisoTo review: substitution effect income effect normal/inferior/Giffen good2-6Causation vs. CorrelationoStatistical analysisnCorrelationnControl groupnTreatment groupoConditions required for government action X to cause societal effect YnX must precede YnX and Y must be correlatednOther explanations for any observed correlation must be eliminated2-7Experimental StudiesoBiased estimatesoCounterfactualoExperimental (or randomized) study2-8Conducting an Experimental StudyoRandom assignment to control and treatment groups2-9Pitfalls of Experimental StudiesoEthical issuesoTechnical problemsoResponse biasoImpact of limited duration of experimentoGeneralization of results to other populations, settings, and related treatmentsoBlack box aspect of experiments2-10Observational StudiesoObservational study empirical study relying on observed data not obtained from experimental studyoSources of observational datanSurveysnAdministrative recordsnGovernmental dataoEconometricsnRegression analysis2-11Conducting an Observational StudyoL = 0 + 1wn + 2X1 + + nXn + nDependent variablenIndependent variablesnParametersnStochastic error termoRegression analysisnRegression linenStandard errorwnL0Interceptis 0Slopeis 12-12Types of DataoCross-sectional dataoTime-series dataoPanel data2-13Pitfalls of Observational StudiesoData collected in non-experimental settingoSpecification issuesogarbage in . . . garbage out2-14Quasi-Experimental StudiesoQuasi-experimental study (= natural experiment) observational study relying on circumstances outside researchers control to mimic random assignment2-15Conducting a Quasi-Experimental StudyoDifference-in-difference quasi-experimentsoInstrumental Variables quasi-experimentsoRegression-Discontinuity Quasi-Experiments2-16Pitfalls of Quasi-Experimental StudiesoAssignment to control and treatment groups may not be randomoNot applicable to all research questionsoGeneralization of results to other settings and treatments2-17oIt is possible that two different economists can examine the same situation but reach entirely different conclusions. Why is this so?oReasoning resulting in differing results includes the time period under examination, the data sources and proxies used, the econometric tools employed, and many other reasons. This is not to say that because different researchers come to different conclusions the analysis is confused. It just means further investigation is needed.