Insegnamento a.a. 2018-2019

20594 - ECONOMETRICS FOR BIG DATA

Department of Economics

Course taught in English
Go to class group/s: 23
DSBA (8 credits - I sem. - OB  |  SECS-P/05)
Course Director:
LUCA SALA

Classes: 23 (I sem.)
Instructors:
Class 23: LUCA SALA


Mission & Content Summary

MISSION

The main goal of this course is to give students a working knowledge of the most used econometric techniques. The key concepts of statistical theory underlying each method will be covered, but special emphasis will be placed on implementation of each method in actual applications. Students will also receive an introduction on how to conduct practical econometric analysis of actual data using software and programming languages such as Stata and Matlab. The course will be divided in two parts. The first, taught by Prof. Giustinelli, deals with regression and causal inference methods used in the analysis of cross-sectional and longitudinal data, typically used in micro-econometrics (where the focus is on the individual behavior of individuals, households, firms and so on). The second, taught by Prof. Sala, deals with time series data and methods, typically used in macro-economic applications (where the focus is on the interaction of macroeconomic variables). As observational data, most commonly used in non-experimental sciences such as economics, hardly tell the researcher what is the effect of a certain treatment variable on a given outcome variable of interest, economists have devised a variety of approaches to address questions of cause-and-effect among economic variables both in microeconomics and macroeconomics. The unifying theme of the two parts of the course will be a focus on understanding causality.

CONTENT SUMMARY

Part I (Microeconometrics):

- Introduction to Microeconometrics: The experimental ideal. 

- Core Regression: Fundamentals of linear regression.

- Core Regression: Regression and causality.

- Core Regression: Heterogeneity and nonlinearity.

- Core Regression: Some Extensions to weighting and limited dependent variables.

- Instrumental Variables (IV): Classical IV and Two Stages Least Squares.

- Instrumental Variables (IV): Heterogeneous Potential Outcomes and Local Average Treatment Effect.

- Parallel Worlds: Individual fixed effects; Differences-in-differences; Panel data methods.

- Regression Discontinuity Designs (RDD): Sharp and Fuzzy RDD; relationship with IV.

- More on Regressions: Types and choice of loss functions; Quantile regression.

- More on Regressions: Parametric vs. non-parametric specification and estimation.

- Impact of Machine Learning and Big Data on Microeconometric Analysis.

 

Part II (Macroeconometrics):

- Why multivariate models.

- Simultaneous equation bias. The problem of identification.

- The Sims critique to old Keynesian macroeconometrics.

- VAR models.

- Granger causality (application: Sims, 1972).

- Structural VAR and identification (applications: a) Sims, 1980, b) Blanchard-Quah, 1989 and Gali, 1999. c) news shocks).

- External instruments.

- A first pass on Big Data: principal components and factor models.

- Dealing with causality. A microfounded model: the Real Business Cycle model.

- Estimating a microfounded model: using the Kalman filter to build the likelihood function.


Intended Learning Outcomes (ILO)

KNOWLEDGE AND UNDERSTANDING

At the end of the course student will be able to...

Part I (Microeconometrics):

  • Define key concepts in econometrics, for instance “econometric causality;” “fundamental problem of causal inference;” “average treatment effect;” “instrumental variable;” “loss function.”
  • Explain key differences and links between distinct but related econometric concepts, for instance “identification and statistical inference;” “potential, realized, and counterfactual outcomes;” “average treatment effect, average treatment effect on the treated, and average treatment effect on the untreated;” “classical instrumental variables estimator and local average treatment effect estimator;” “sharp and fuzzy regression discontinuity designs;” “parametric, semiparametric, and nonparametric specification of an econometric model.”  
  • For each causal method covered during the course, describe the roles of the key assumptions underlying the method in yielding identification of the causal effect parameter of interest. Distinguish untestable assumptions from testable ones, and describe how to test the latter. Discuss the main consequence/s of the failure of each assumption, illustrating with specific examples or applications.
  • For each causal method covered during the course, describe the basic data requirements for application of each method and discuss advantages and disadvantages of each method.

 

Part II (Macroeconometrics):

  • Being familiar with the main concepts and tools of time series analysis and being able to use them in other contexts.
  • Understand a vast majority of the scientific literature on time-series and macroeconometrics.
  • Identify what are the modelling assumptions underlying any structural macroeconometric model.
  • Translate the main assumptions in economic theories into restrictions on the empirical statistical model.

APPLYING KNOWLEDGE AND UNDERSTANDING

At the end of the course student will be able to...

Part I (Microeconometrics):

  • Design a simple randomized experiment to quantify the average causal effect of a treatment variable on an outcome variable of interest.
  • Given a specified microeconomic application, a data set, and a causal parameter of interest, select the most appropriate microeconometric method among those covered in class.
  • Implement the chosen method to quantify the causal parameter of interest and test hypotheses about the parameter’s true value, using actual data and a statistical software covered in class.
  • Evaluate the causal effect(s) of a policy intervention or program, using given microdata and the most appropriate econometric method among those covered in class.

 

Part II (Macroeconometrics):

  • Perform empirical analysis to uncover the effects of shocks in the economy.
  • Design a well-functioning empirical macroeconometric model.
  • Communicate effectively the empirical results of his/her analysis.
  • Use programming software, Matlab, to perform different kind of time-series analyses.
  • Do empirical analysis in a constructive way and think critically.

Teaching methods

  • Face-to-face lectures
  • Exercises (exercises, database, software etc.)

DETAILS

The learning experience of the course includes: (i) face-to-face lectures, introducing and illustrating the main topics of the course; (ii) interactive in-class discussions around stylized microeconomic and macroeconomic applications, focusing on specific aspects of their implementation and interpretation; (iii) Students will also be exposed to statistical/econometric software, such as Stata and Matlab. This will allow them to get a hands-on introduction to the programming techniques that are a necessary part of any econometric study.


Assessment methods

  Continuous assessment Partial exams General exam
  • Written individual exam (traditional/online)
  x x

ATTENDING AND NOT ATTENDING STUDENTS

To the end of measuring the acquisition of the above-mentioned learning outcomes, the students’ assessment will be based on a written final exam accounting for 100% of the final grade. The exam will consist of exercises, open-ended and/or closed-ended questions. The exam will assess students’ knowledge and understanding of the topics and students’ ability to carry out and interpret results from empirical econometric work.


Teaching materials


ATTENDING AND NOT ATTENDING STUDENTS

The main course material for both attending and non-attending students is:

Main textbook for Part I (Microeconometrics):

  • Angrist, J.D. and J.-S. Pischke (2009), Mostly Harmless Econometrics, Princeton University Press.

 

Main textbook for Part II (Macroeconometrics):

  • L. Sala , “Lecture Notes on Time Series Analysis” (available on Blackboard).
  • W. Enders, “Applied Econometric Time Series”, last edition, selected chapters.
  • J. D. Hamilton, “Time Series Analysis”, Princeton University Press, 1994, selected chapters.

 

The slides of the course, additional readings and support material will be uploaded to the Blackboard platform of the course.

Last change 04/07/2018 10:07