Steve Lawford

 

IENAC22 (PREV) Forecasting

 

LATEST ANNOUNCEMENTS

  • (22 Apr) I have uploaded the airline cost data that is used in the panel data videos.
  • (22 Apr) I have uploaded a detailed revision video for the Forecasting module, in question-and-answer form #143, to the Drive. This covers both theoretical and empirical questions (discrete choice, panel data, time series) and should be used to complement the theory classes and applied problem sets. The video is ~2h long but it will take you much longer to go through if you attempt all of the questions yourself. Note that video #143 does not cover E2.APS2 or E2.APS3.
  • (22 Apr) I have added the mock exam and a full solution. The mock exam only covers discrete choice and time series ARMA modelling / this year, there could also be questions on panel data, as well as any topic from E2.APS2 (nonlinear hypotheses, structural breaks, nonlinear least squares) or E2.APS3 (autocorrelation).
  • (22 Apr) I have uploaded video solutions to F.APS3 (log GNP and interest rate).
  • (16 Apr) I have added full solutions to the four theory questions that appear in handouts #129 and #137.
  • (14 Apr) I have uploaded F.APS3 (time series ARIMA modelling) and the three datasets.
  • (14 Apr) I have added a handwritten solution to F.APS2.
  • (14 Apr) I have uploaded full video solutions #100 to #115 to F.APS1.
  • (11 Apr) I have uploaded videos #116 to #128 (panel data) to the Drive. This is required material and will be covered during the final exam. The videos include an applied example that you should work through. There are ~3h of videos in total.
  • (11 Apr) I have uploaded F.APS2 (discrete choice models) for revision.
  • (9 Apr) I have uploaded video solutions to E2.APS3 to the Drive.
  • (7 Apr) I have uploaded F.APS1 (discrete choice models).
  • (7 Apr) I have uploaded video solutions to E2.APS2 to the Drive.
  • (27 Mar) Welcome to the course, and good luck!

 

Course and content

Contact details and office hours: It is best to contact me by email with questions - since I receive lots of mail, please mark messages with "IENAC22", somewhere in the subject line, otherwise they may get lost.

In case of problems, you can also contact me on Teams or by (internal) phone on 719532, or on the Google Chat thread.

Please contact me in advance by email to arrange individual meetings, including the following information: (a) a precise (one-line) description of the problem, (b) any supporting data, files or illustrations (e.g. screenshots), (c) your availability and expected time required. Due to very heavy time constraints on my side, I will have to refuse all meetings unless you follow this procedure.

Language: The lectures, classes, all course material, and the assessments, will be in English.

Content: This course develops the material covered in Econometrics 1 and Econometrics 2. The first part of the course introduces autocorrelation and binary choice models (probit and logit). The second part of the course covers panel data models (fixed effects and random effects) and testing. The third part of the course develops some fundamental concepts of time series: difference equations, convergence, linear filters, stationarity of the ARMA(p,q), and the Box-Jenkins methodology for building ARIMA models for forecasting. If we have enough time, we will also cover nonstationarity and unit root testing. The theory is illustrated with detailed empirical examples, using the commercial econometric software package EViews.

Textbook: I will use my own lecture notes on the subject, and further course material as appropriate. A printed version of all material will be given to you towards the end of the course, before the exam.

Please wait until I contact you before going to the Printshop to collect the polycopy! Thank you.

Examinations (in the Forecasting module): The final grade is based on a written examination, in week 21 (100% of grade) - this will cover both applied and theoretical topics. You will not have computer access during the written examination.

I will award a small prize for the final exam performance.

Corrections and errors: Please contact me if you find any errors or typos in the course material, or on this website.

Administration:

Various I.E.s and Séverine C. (BdP).

 

Teaching

~16 hours of lectures and classes (these are interchangeable).

There will also be two 3 hour classes by visiting speakers, that will not be assessed during the exam.

Lectures and classes will be held in weeks 2024 // 14 - 17 (Apr).

Class schedules are sometimes subject to change at short notice.

ENAC Intranet upcoming class schedules.

2024 //

Schedule for week 14

✔ Wed 3 Apr 2024 - Class 1/* - 15:30-17:30 -- room D202 (we worked on E2.APS2)

✔ Fri 5 Apr 2024 - Class 2/* & Class 3/* - 13:15-15:15 & 15:30-16:30 (3h) -- room D204 (we went through full solutions to E2.APS2) #71 #72 #73 #74 #75 #76 #77

To do this week //

(*) 1> Read the Jarque-Bera paper [for E2.APS3 (autocorrelation)] *optional*

2> Complete E2.APS3 before the next class *important*

Schedule for week 15

✔ Mon 8 Apr 2024 - Class 4/* & Class 5/* - 13:15-15:15 & 15:30-17:30 (4h) -- room D206 (we went through full solutions to E2.APS3) #78 #79 #81 #82 #83 #84 #85 (discrete choice theory) #89 #90 #91 #92

✔ Wed 10 Apr 2024 - independent work - 15:30-17:30 -- room D204

✔ Fri 12 Apr 2024 - Class 6/* & Class 7/* - 13:15-15:15 & 15:30-17:30 (4h) -- room D206 #93 #98 (discrete choice theory) (we went through almost complete solutions to F.APS1) #100 #101 #102 #103 #104 #105 #106 #107 #108 #109 #110 #111 #112 #113 #114 #115 // video 102 (F.APS1.Q3) contains optional theory on Huber-White standard errors, video #103 (F.APS1.Q4) contains optional theory on the delta method and the standard errors of estimated probabilities and marginal effects, video #114 (F.APS1.Q15) contains optional theory on the mean and variance calculations for the standard logistic distribution, we did not cover #110 (F.APS1.Q11) #111 (F.APS1.Q12) #112 (F.APS1.Q13) or #114 (F.APS1.Q14) in class as these are less important questions and can be considered optional for the exam

To do this week //

1> Watch videos #94 (probability response curves) #96 (2x2 table of hits-and-misses) #99 (geometric intuition for the logit) *important*

(*) 2> Watch videos #95 #97 *optional*

Schedule for week 16

✔ Mon 15 Apr 2024 - Class */* & Class */* - 13:15-15:15 & 15:30-16:30 (3h) -- room D202 #129 #137 (the material that we saw in class is covered in videos #129 #130 #131 #132 #133 #134 #136 #137 #138 #139 #140 #141)

✔ Wed 17 Apr 2024 - Class */* - 15:30-17:30 -- room D206 #137 (we covered Box-Jenkins and worked on F.APS3; also see video #142 for the Box-Jenkins methodology)

✔ Fri 19 Apr 2024 - Class */* & Class */* - 13:15-15:15 & 15:30-17:30 (2h + 2h independent work) -- room D206 (we went through a complete solution to F.APS3 dataset 1 - log GNP; a less complete solution to this question is covered in video #144 - in class, we went further, and used an ARIMA(3,1,2) for forecasting; F.APS3 dataset 2 - interest rate, is covered in video #145)

To do this week //

1> Watch panel data videos #116 #117 #118 #119 #120 #121 #122 #123 #124 #125 #126 #127 #128 *important / work through the empirical example too* [data for empirical example]

(*) 2> Watch discrete choice videos #102 #103 #114 *optional and more advanced theory (see week 15)*

(*) 3> Watch discrete choice videos #110 #111 #112 #114 *optional missed questions from F.APS1 (see week 15)*

4> Watch time series video #141 *important / partial autocorrelation function*

Schedule for week 17

✔ (Econometrics 1 project presentations - see Econometrics 1 webpage)

Mon 22 Apr 2024 - Visiting Speaker independent work - 14:00-17:00 (3h) -- room Amphi Bréguet D206

Wed 24 Apr 2024 - Visiting Speaker independent work - 15:30-18:30 (3h) -- room Amphi Bréguet D206

Schedule for week 18 new

(Econometrics 2 paper presentations - see Econometrics 2 webpage)

Schedule for week 19

no classes this week

Schedule for week 20

no classes this week

Schedule for week 21

Fri 24 May 2024 - FINAL WRITTEN EXAMINATION - 13:15-15:15 - room Bréguet

 

Resources

Applied Problem Sets new

E2.Applied Problem Set 2 (nonlinear hypotheses, structural breaks, nonlinear least squares) [data for E2.APS2] solutions #71 #72 #73 #74 #75 #76 #77

E2.Applied Problem Set 3 (autocorrelation) [data for E2.APS3] solutions #78 #79 #81 #82 #83 #84 #85

F.Applied Problem Set 1 (discrete choice models) [data for F.APS1] solutions #100 #101 #102 #103 #104 #105 #106 #107 #108 #109 #110 #111 #112 #113 #114 #115

F.Applied Problem Set 2 (discrete choice models) [data for F.APS2] [solution]

F.Applied Problem Set 3 (time series modelling) [data for F.APS3 (U.S. log GDP)] [data for F.APS3 (interest rate)] [data for F.APS3 (passenger numbers)] video solution #144 (log GNP) #145 (interest rate) // note that we went into more detail in class than #144

Airline cost data for panel data videos new

Airline cost data

In-class exercises new

Solutions to time series theory problems / see handouts #129 (Q1) and #137 (Q2, Q3, Q4) or videos #136 (Q1) #137 (Q2) #140 (Q3, Q4)

Other

(1) Statistical tables (pdf)

Past exams new

IENAC17 Forecasting exam ** please note that the mock exam does not contain any coverage of panel data, since this was not covered in class that year; the final exam this year may include discrete choice, panel data, and time series, as well as any topic from E2.APS2 (nonlinear hypotheses, structural breaks, nonlinear least squares) or E2.APS3 (autocorrelation) ** [solution]

Revision video new

Revision video for Forecasting module #143 on Drive (does not include E2.APS2 or E2.APS3)

Reading list (optional)

Of particular use is Greene, W. H. (2000), Econometric Analysis, 4th edition, New Jersey: Prentice-Hall / Chapter 19 binary choice models (probit and logit).

Also helpful is Cameron, A. C. and Trivedi, P. K. (2006), Microeconometrics, Cambridge: Cambridge University Press / Chapter 14 binary choice models (probit and logit).

See Chapter 17 in Hamilton, J. D. (1994), Time Series Analysis, Princeton: Princeton University Press, for excellent technical material on univariate processes with unit roots (asymptotics).

 

Assessment