Flexible learning
Online postgraduate courses
Advanced Forecasting for Time Series
Please note to take this course you must first have completed Time Series Econometrics and Forecasting
The course aims to develop students' econometric skills and provide practical guidance on how to forecast. Economics needs to forecast a non-stationary and evolving world, using a forecasting model that differs from the economic mechanism. The resulting framework, its basic concepts and main implications are sketched. Many famous theorems of economic forecasting no longer hold—rather, their converses often do.
We will examine how standard macroeconometric models fail in many forecasting scenarios, which provides guidance on how to correct such forecast failure. We shall also look at how the forecast theory developed can be applied to other disciplines. The course will be practical. All empirical examples will be worked through using the econometric software package OxMetrics.
This module can be taken alone or as part of a PG Certificate, PG Diploma or Full Masters Program.
-203.jpg)
- 10 Credits
- 100 hours of study
- 15 contact hours
- 85 hours for private study
Faculty
Structure
Structure
Advanced Forecasting for Time Series is an elective 10 credit course and therefore students are expected to input approximately 100 hours of study into the course.
The total number of contact hours is 15 hours. This leaves 85 hours for private study.
There will be an assessment at the end of the course.
Lectures
This module consists of 2 - hour lectures per day for 5 days, plus a 1 - hour tutorial per day.
There will be optional clinics on the last day of the course.
The dates of each lecture are confirmed closer to the start of each term. If you have any questions about dates, please contact edu@timberlake.co.uk.
OxMetrics
Module Programme
Introduction to Economic Modelling and Forecasting
- The framework for economic modelling and forecasting, its basic concepts and main implications willbe sketched. The theory of reduction underpins economic modelling: Models with no losses onreduction are congruent; those that explain rival models are encompassing. The main reductionscorrespond to key econometrics concepts (causality, exogeneity, invariance, etc.), and are the nullhypotheses of model- evaluation tests, sustained by a taxonomy of evaluation information.Congruent and encompassing sub- models can, therefore, be justified, motivating the question ‘howshould they be selected’? The key problems in forecasting are also highlighted, emphasising thedistinction between determining economic relationships or testing theories and forecasting. Goodmodels may forecast badly and bad models can forecast well – a concept that will be exploredthroughout the course.
Introduction to OxMetrics
- In this applied session we will introduce OxMetrics (data input, transformation, graphics, modulesand recording results) and PcGive, the basic modelling tool, including model formulation, selection,and evaluation. This session will also explore the forecasting tools available in the software, includinggraphical and statistical output. Various applications will illustrate the software.
Selecting Forecasting Models
- Model selection theory poses great difficulties: all statistics for selecting models and evaluating theirspecifications have distributions, usually interdependent, and possibly altered by every modellingdecision. General-to-specific (Gets) modelling will be described, emphasising automatic procedures.Gets mimics reduction by simplifying a congruent general unrestricted model (GUM) to a dominantminimal representation. Autometrics will be explained and its properties discussed. The properties ofmodel selection will be discussed by way of a class Monte Carlo experiment, in which eachparticipant generates a draw of data from a DGP using PcNaive, a software package withinOxMetrics, and we compare the retention of relevant and irrelevant variables to the theorypredictions, contrasting the results with the notion of ‘size’. Methods for handling more candidatevariables than observations are shown, leading to empirical model discovery
Forecasting Problems
- We examine the main sources of forecast failure using artificial data in a series of examples tohighlight the results. PcNaive will be used to generate the forecasting examples explored. A range ofparameter changes in integrated-cointegrated, I(1), time series are hardly reflected in econometricmodels thereof: zero-mean shifts are not easily detected by conventional constancy tests. The breaksin question are changes that leave the unconditional expectations of the I(0) components unaltered.Thus, dynamics, adjustment speeds etc. may alter with a low chance of detection. However, shifts inlong-run means are generally noticeable. We’ll draw important implications for the choice offorecasting device.
Foundations of Unpredictability
- Six aspects of the role of unpredictability in forecasting are distinguished, compounding fouradditional mistakes likely when estimating forecasting models. Many of the famous theorems ofeconomic fore- casting do not hold in a non-stationary and evolving world, when the model andmechanism differ; rather their converses often do. Equilibrium-correction models are shown to be arisky device from which to forecast. Potential explanations for the intermittent occurrence of forecastfailure include poor models, inaccurate data, inadequate methodology, mis-calculation ofuncertainty, structural change, over-parameterization, incorrect estimators, and inappropriatevariables. In fact, using a simplified taxonomy of forecast errors, most of these can be shown not toexplain forecast failure, and the forecast-error taxonomy shows that forecast failure dependsprimarily on forecast-period events, particularly location shifts.
Robustifying Forecasts
- In this practical session we shall explore one method of robustifying forecasts to location shifts.Differencing lowers the polynomial degree of deterministic terms: double differencing usually leadsto a mean-zero, trend-free series, as continuous acceleration is rare in economics (except perhapsduring hyperinflations or major technological shifts). The impact on forecast performance is traced. Anew explanation for the empirical success of second differencing is proposed. Forecasting will beconducted for several model variants, with and without forecast failure. The role of parameterestimation uncertainty is considered. The practical role of forecast-error corrections will beinvestigated, and many theoretical issues illustrated through both successful and unsuccessfulforecasting, including how to cope with location shifts. Examples will include Japanese Exportforecasts and UK GDP forecasts.
Robust Systems
- In this practical session we shall work with the Cointegrated Vector Autoregressive model (CVAR) toshow how to implement robustification in the system. We shall use the well known example offorecasting UK M1 money demand when there was a legislative change leading to large forecastfailure. This practical session will demonstrate the use of multivariate modelling within PcGive andresults in the very general class of robust forecasting devices that uses local averages to estimate thechanging location and growth rate parameters.
Forecasting Breaks
- We note research on forecasting breaks, and the demanding conditions under which that might bepossible, as well as learning about breaks during transitions. The possible roles of parsimony andcollinearity in forecasting highlight the potential importance of excluding irrelevant, but changing,effects. While intercept corrections help robustify forecasts against biases due to location shifts,they are ineffective for measurement errors: conversely EWMA corrections are excellent formeasurement errors, but not breaks. Rapid updating is related both to moving windows and toforecasting breaks, with some properties that can help alleviate failure. Forecast pooling can alsosometimes help, but needs to be combined with model selection to exclude really bad forecastingdevices. Then, pooling can lead to improved forecasts over the best of a set of devices in a world ofmis-specified models and location shifts. However, care is needed in selecting what enters the pool,and indiscriminate pooling (as in Bayesian model averaging) can be counter-productive.
Forecasting with Factors
- Dynamic Factor Models are a common approach to forecasting. In this practical session we discussthe computation of Principal Components and use these to build forecasting models. We willcompare direct and iterated forecasts, as well as a comparison with unobserved component models.We’ll explore the generic approach of including both factors and variables, made feasible withAutometrics by applying selection to ensure sparsity. We will conclude with a forecasting game: Whocan obtain the lowest RMSFE for three target variables?
Conclusions and Discussion
- In this session we shall draw together various aspects that we have discussed on the course. Someextensions will be discussed, including extending the robust forecasting device to longer horizons andapplications to other fields including climate. We will end with a guide to modelling and forecasting,illustrated by an example using UK CO2 emissions. First, impulse and step indicators are selected atvery tight significance levels holding all other variables fixed. Next, the regressors are selected over atlooser significance levels. The selected model is solved for the cointegrating, or long-run, relation andthe non-deterministic terms are reparameterized to differences, with step indicators included in thecointegrating relation. The non-integrated formulation is re-estimated and used to produceconditional forecasts. The VAR is then constructed to obtain unconditional system forecasts and theimportance of step indicator saturation is observed.
Prerequisites
English Language Requirements
Both Programmes are open to applicants anywhere in the world. We may ask applicants to provide a recognised English language qualification, dependent upon their nationality and where they have studied/worked previously.
The requirement is an IELTS (Academic) Test with an overall score of at least 6.5, and a minimum of 6.0 in each element of the test. We will also consider other English language qualifications. If their score is below our requirements, they may be eligible for one of Lancaster University's pre-sessional English language programmes.
Academic Requirements
Applicants to the Postgraduate Certificate of Achievement, Postgraduate Certificate, Postgraduate Diploma or full MSc in either programme require either an upper second-class degree in economics, econometrics or related subjects.
Learning Outcomes
- Use the model selection software, Autometrics, to select macroeconometric models used forforecasting
- Compute point and interval forecasts using standard and robust forecasting devices
- Evaluate forecasts of macroeconomic variables
- Develop an understanding of when forecasts may or may not perform well.
- Engage in abstract thinking by extracting the essential features of complex systems tofacilitate problem solving and decision-making
- Communicate and present complex arguments in oral and written form with clarity andsuccinctness
- Present, interpret and analyse information in numerical form
- Utilise effectively statistical and other packages
- Apply basic statistical techniques to analyse economic and financial datasets
- Work effectively both individually and within a team environment
Frequently Asked Questions
All sessions are conducted live and online at a scheduled time, but are also recorded. Students may attend live and watch the recordings back to recap the material or watch the recordings only if unable to attend live. We always advise students to attend live where possible as this will allow them the best opportunity to engage with the content and ask the lecturer's questions.
All modules are examined through online coursework submissions, you will have the support of your module lecturer/tutor in this poccess.
No, all necessary software is provided to students.
All courses are recorded and available on the LUMS internet platform throughout the current academic year. They can therefore be viewed 24 hours a day.
Timberlake Consultants and Lancaster University Management School (LUMS) Economics department have a longstanding partnership; combining 40+ years of industry expertise with over 50 years of academic excellence. We are delighted to build on this with our micro-credential postgraduate courses.
Flexible learning
Online postgraduate courses
Specialist courses
Best-in-field experts
Tailored learning
Career-oriented education
Subscribe To Our Newsletter
Flexible learning
Online postgraduate courses
Specialist courses
Best-in-field experts
Tailored learning
Career-oriented education