Major Topics: Topic 1: Scientific Research Process Empirical Economic Research What is Econometrics? Doing Research Page 1.1 Scientific Research Process Economics, as a science, should follow the Scientific Research Process Problem formulation/definition Theory development Falsification with observational data Problem solution Page 1.2
Scientific Research Process (Continued) See Fig. 1.1 Scheme related to K. Poppers Philosophy Of Science approach Popper was leading philosopher of science around time of Keynes Had great influence on early econometricians Page 1.3 Enhancement Suggestions SRP: Initial Problem Formulation Theory Development Data Falsification Problem Solution Popper s Dialectic: P 1 TT EE P 2 Astrophysics Example: - Radiowave Emissions - Unexpected - Models Developed - Observational Verification - Understanding of Source Economics Example: - Level and Changes in Personal Consumption Expenditures - Keynesian Consumption Function - Time Series and Cross -sectional Studies - Permanent Income and Life Cycle Theories Fig. 1.1 The Scientific Research Process These are the main steps followed in any research study and are the core for this textbook. Notice that there is a feedback step with suggested enhancements to the theory depending on the falsification results. Subsequent chapters will develop each box. Page 1.4
Scientific Research Process (Continued) Other sciences (e.g. physics) more empirically based Rely on combination of theorizing and empirical falsification More emphasis on observational data Economists place more emphasis on theorizing Central methodological imperative of this course is that economic theories must be empirically tested We must place more emphasis on understanding and analyzing observational data Page 1.5 Empirical Economic Research All work in economics starts with theories Econometrics makes economic theories relevant Observational data makes econometrics useful And powerful Empirical Economic Research Triad composed of: Theory Data Econometrics Illustrated By Research Triangle Page 1.6
Theory/Hypothesis What to Test Use to Test Specific Research Question Use to Test Model Specification Data Analysis Input Refinement Econometric Analysis Empirical Stage of Analysis Fig. 1.2 The Research Triangle This triangle depicts the relationship among theory, data and econometrics for empirical economic research. They are all needed for solving a problem. There is not one part of the Triangle that dominates another. The central focus of the Triangle is the specific research question or problem. Page 1.7 Empirical Economic Research (Continued) No one part of triangle dominates Theory highlights key variables or drivers for data collection and empirical modeling Data analysis supports theory development and econometric model specification Econometrics falsifies the theory and directs data collection Page 1.8
Empirical Economic Research (Continued) Role of theory Organizing framework Theories are designed to promote systematic and organized methods of reasoning (Friedman 1953, 7) Create artificial worlds Allows us to test ideas Work does not stop here Page 1.9 Empirical Economic Research (Continued) Role of econometrics Allows for testing theories Stress is on falsification, not verification Verification Can never completely verify anything Can not find all cases so do not know that something will not be overturned in future Example: all swans are white Falsification Just need one occurrence of something happening or not happening Example: find one black swan Page 1.10
Empirical Economic Research (Continued) Role of econometrics (Continued) A theory must be judged by its predictive power... Only factual evidence can show whether it is right or wrong or, better, tentatively accepted as valid or rejected. (Friedman 1953, 8-9) Page 1.11 Empirical Economic Research (Continued) Role of econometrics (Continued) Friedman states (1953, 8-9): the only relevant test of the validity of a hypothesis is comparison of its predictions with experience. The hypothesis is rejected if its predictions are contradicted ("frequently "or more often than predictions from an alternative hypothesis); it is accepted if its predictions are not contradicted; great confidence is attached to it if has survived many opportunities for contradictions. Factual evidence can never "prove" a hypothesis; it can only fail to disprove it, which is what we generally mean when we say, somewhat inexactly, that the hypothesis has been "confirmed" by experience. Page 1.12
Empirical Economic Research (Continued) Role of econometrics (Continued) Testing is difficult Theories are complex and many variables do not have real world counterparts Test implications of theory Testable hypothesis Page 1.13 Empirical Economic Research (Continued) Role of data analysis Data are what we live and die by Must analyze our data to understand and look for Data Scrutiny Scrutinize Summarize Explore Anomolies Outliers Trends Patterns Page 1.14
Empirical Economic Research (Continued) Role of data analysis (Continued) Tools Descriptive statistics Means/variances frequently used Medians/quartiles preferred because of robustness Graphical analysis Simple plots Histograms Scatter Plots Box Plots Page 1.15 What is Econometrics? What is econometrics? [e]conometrics is the application of a specific method in the general field of economic science in an effort to achieve numerical results and to verify economic theorems. G. Tintner (1952) Page 1.16
Doing Research Many questions have to be answered before empirical work can be done What steps do we follow to transform a theory into something to which econometric techniques can be applied? At what point do we introduce data analysis? What data analysis do we do? How do we specify an empirical model? Empirical Economic Research Process (EER) Lays Out Steps To Address Issues Page 1.17 Initial Problem Definition Re-examine the Theory No Theoretical Framework Testable Hypotheses Modeling Stage Problem Solved? Feedback Loop Yes Formulate New Problem Fig. 1.3 Empirical Economic Research Process This flowchart depicts the steps in a research process. The five major steps shown are discussed throughout the textbook. Page 1.18
Theoretical framework Example: Keynes Consumption Function Testable hypotheses Implications OF THEORY THAT CAN BE TESTED A STATEMENT FALSIFIABLE WITH DATA Example: Marginal Propensity to Consume (MPC) Page 1.19 Testable hypotheses (Continue) Example applications Personal exemption and fertility Problem: What is impact of personal exemption on fertility? Theory: Neoclassical Consumer Demand Theory Testable hypothesis: increase in exemption increases fertility Page 1.20
Testable hypotheses (Continue) Example applications (Continue) Money growth and unemployment Problem: explain unemployment and role of money Theory: Rational Expectations Testable hypothesis: there is an inverse relationship between unanticipated money supply growth and unemployment Page 1.21 Testable hypotheses (Continue) Example applications (Continued) Education and convergence Problem: levels of productivity growth in many countries converging. Why? Theory: Growth = f(k, L) where K includes human capital Implication: human capital increases, growth increases Testable hypothesis: education increases, growth increases Page 1.22
Modeling stage Steps Preliminary model specification Data analysis Further model specification Estimation Hypothesis testing Validity checking Page 1.23 Modeling stage (Continued) Preliminary model specification Model form Linear or nonlinear? Additive or multiplicative? Type of data Cross-sectional or time series? Monthly/quarterly/etc.? Primary or secondary? Variables to include? How measured? Specialized variables (e.g., dummies) Page 1.24
Modeling stage (Continued) Data analysis Any measurement errors and if so, how bad? how to correct? Any outliers and if so, impacts? Source and conditions under which data collected? Who collected data? Further model specification Probabilistic assumptions Page 1.25 Modeling stage (Continued) Estimation Major preoccupation of econometrics Hypothesis testing Can be difficult Validity checking Comparing against other results Page 1.26
Other METHODOLOGICAL APPROACHES See Fig. 1.4 This course is under Traditional, Empirical Appraisal Branch Page 1.27 Methodological View of Economics "Traditional" "New View" Objectives of Theory Methods of Theory Appraisals - Theories are Conglomerate Structures - Science is a Social Process - Kuhn Exemplifies - Many Offshoots Exist Realism Instrumentalism Deductiveism Empiricism - Describe - Explain - Causation - "Critcial Realism" - Tool Only - No Relation to Causal Reality - Friedman Exemplifies - Discover Events, Regularities Only - Axiomatic - Logical Deductiveism - Truth not Based on Empirical Reality - Theory Judged on Internal Logic - Consistent with Observed data - Popper and Falsificationism - Theory Judged by External Criteria - Blaug Leading Advocate Fig. 1.4 Relation of Methodological Views This tree diagram relates the different methodological view prevalent in economics according to Gerrard (1995). This textbook falls under the Traditional, Empirical Appraisal of Economic Theories branch of the tree Page 1.28
Summary Econometrics is part of a Scientific Research Process One of three interdependent aspects Theory Econometrics Data Analysis Goal Falsify theories Page 1.29 Knowledge Checks Describe the Research Triangle and how it relates to economic research. What is the role of economic theory in economic research? Explain how econometrics relates to economic theory. Is econometrics the sole means of falsifying theory? What is data analysis and what role does it play in research? What should we look for during the data analysis phase? Briefly define a testable hypothesis and give an example of one other than the one mentioned in the text. Explain how a model specification relates to a testable hypothesis. Describe the steps in the modeling process. Page 1.30
Appendix: Modeling Questions Model form to specify Is it linear or non-linear? Is it multiplicative or additive? Is it deterministic or stochastic? Is there just one equation or several that are interrelated? Page 1.31 Appendix: Modeling Questions (Continued) Type of data to use Should cross sectional data or time series be used? At what frequency should time series data be measured if this is what is used? Months? Quarters? Annual? What time period should the data be measured for if time series data are used? What units should the data be measured in if cross sectional data are used? States? Countries? Industries? Should ratio or categorical data be used? What is the source of the data? Page 1.32
Appendix: Modeling Questions (Continued) Variables to include and measure Should the variables be measured on, say, a per capita basis or in aggregate? Are there any specialized variables such as dummy variables, lagged dependent variables or trend variables that must be added to the model? If so, how are they specified? Page 1.33