FI4003 Empirical Methods in Finance
Assignment 1: Inflation forecasting
Forecasting methods. Meyer and Passaogullari (2010) suggest two simple
methods to forecast the one-year-ahead inflation rate. Given monthly data,
the first methods uses a regression of the form1
(1) pi12t+12 = α + β(L)pit + εt+12
where pi12t+12 = lnCPIt+12 − lnCPIt is the year-on-year inflation rate, pit =
12 · (lnCPIt− lnCPIt−1) is the monthly inflation rate at an annual rate, and
β(L) = (β0 + β1L + β2L
2 + . . .) is the lag polynomial. The error term εt+12
is assumed to be white noise. The second method uses pi12t , the year-on-year
inflation rate in period t, to forecast the inflation rate over the next twelve
months, pi12t+12.
Task and coverage. The data set series-051021.xls contains the UK
monthly CPI for the period 1988m2-2021m08.2 The assignment requires you
to conduct a small research project on how accurately the above methods can
forecast the annual inflation rate in the UK. You should therefore analyze
the time-series properties of data, implement the forecasting regression, and
produce inflation rate forecasts for each month in the period 2020m9-2021m8
(i.e. the last 12 months in the sample).
Report structure. The report should have the following the structure
1. Introduction
2. Data description
3. Methodology
4. Analysis of results
1See also Stock and Watson (1999).
2The data can also be downloaded here.
1
5. Conclusion
The introduction should comprise a brief motivation for the research. The
data section should provide a description of the data, including an analysis
of the time-series properties of the annual inflation rate. The methodology
section should discuss how you specify and estimate the regression model.
You also need to explain how the model is used for out-of-sample forecasts.
The analysis section should present and interpret your estimation and out-of-
sample forecast results. You should also discuss how one can assess forecast
accuracy and how one might improve upon the two forecasting methods. The
conclusion should not contain any analysis. Generally a report of this type
should have some circularity to it. If you refer to something at the start then
you should refer to this concept at the end, to ensure the work is complete.
General instructions
You should be certain that you fully understand the task. If you want to
check your understanding, please discuss it with me.3
Your report should have a maximum of 1,500 words, the table of contents and
the references do not count towards this word limit. 1,500 words correspond
approximately to for to five normally spaced pages in a Word document with
the font Times Roman at 12pts. References should be in the Harvard style.
The assignment carries 50% of the total marks in FI4003 empirical methods
in Finance. The report will be assessed according to the Common Grading
Scale (CGS). The Feedback Sheet provides further details on the marking
grid.
The paper has to be submitted by 12:00noon on Thursday, 12 November,
2021. Late submissions will be penalised in accordance with the regulations
given in the relevant Business School Policy.
One electronic copy of the coursework has to be submitted to SafeAssign
through MyAberdeen.
3Brooks (2008, Ch. 13) gives an overview how to conduct an empirical research project.
2
References
Brooks, C.: 2008, Introductory Econometrics for Finance, Cambridge Uni-
versity Press, Cambridge, UK.
Meyer, B. H. and Passaogullari, M.: 2010, Simple Ways to Forecast Infla-
tion: What Works Best?, Economic Commentary, 2010-17, Federal Re-
serve Bank of Cleveland.
Stock, J. H. and Watson, M. W.: 1999, Forecasting Inflation, Journal of
Monetary Economics, 44, pp. 293-335.
3



































































学霸联盟