Project description
Question
The archive data2014.zip contains data on key economic indicators and house-
hold expenditure for the UK. Your task is to write a report where you produce
three (3) year forecasts of the following:
1. The UK’s Public Sector Net Debt (RUTN), using a multiple regression
model.
2. The UK Households’ expenditure on clothing (ADFQ), using a Time Series
Decomposition model.
3. The UK’s Employment of over 16s (MGRZ) using an ARIMA/ARMAX
model.
Note: you are neither required nor expected to obtain additional data for any
of above tasks.

Specic Instructions
The report should be 4,000 to 5,000 words in length including the executive
summary, gure and table captions and notes, and foot/endnotes if any of the
latter are used (a word count should be included).
Appendices should include one (1) stata .do le, NOT a log le, for repro-
ducibility of your results. (If a stata .do le is not included or if it does not
accurately reproduce all the analysis presented in the report, inclusive of any
in the Appendices, including, but not limited to, graphs, gures, tables and
results, the report will be referred to Academic Misconduct.)
Remember: you are asked to provide a report NOT AN ESSAY, addressed
to a knowledgeable audience.

I strongly urge you to consider the following:
 Think carefully what is asked of you. Do not rush into producing results
that may not withhold statistical, or other, scrutiny.
 Explore the available data rst, and think carefully about the series you
want to forecast.
 Pay particular attention to the time dimensions of the series.
 Think about any economic and/or business theory that could guide your
analysis.
 Consider the series you are trying to predict and how they are expected
to behave in the wider prevailing economic conditions.
 Your modelling approach should be guided and supported by robust sta-
tistical checks.
 Remember that forecasts should always be evaluated using appropriate
methods (e.g. in- and out-of-sample forecasts, calculation of RMSE or
MSE for each model you compare, Akaike and Bayesian information cri-
teria etc).
 Also, remember, that simple presentation of results is not sucient, and
description of a gure does not constitute discussion”. You need to pro-
vide a proper discussion linking empirical observations and results to both
theory and actuality.
 Equally important to the content, is the presentation of the report i.e.
language used (under no circumstances should you use colloquialisms),
format of gures and tables (raw stata output is unacceptable – DO
NOT copy + paste from stata’s results window into your report),
structure of the report etc.