Instructions

This is an essay, with arguments supported by empirical evidence.  Hence, like any essay, it should have an ëessay structureíconstituted of Introduction, Methodology, Findings, Ro- bustness Check and Conclusion.

● Introduction:  brief motivation+background/literature.  This part addresses why the topic is worth investigating.

● Methodology:  theory, benchmark model  (regression), data.   This part is mainly to present and explain the chosen regression model.   The choice should be based on reasonable supposed believes, such as good theories or relevant previous Öndings.  It also explains the data used, including their sources, time range, measurement, and mostly importantly in time series analyses, whether they are stationary (or how the original data are processed such that the processed data are stationary).

● Findings:  this part reports and interprets the Öndings of the benchmark regression. This should be the main focus, as it answers the central question of the essay.  You can start with the estimation result, consider the estimated coe¢ cient values including their size and sign; and evaluate their signiÖcance.  Then you can comment on these results; e.g., what they mean, what the policy implications are and, (if your Önding is di§erent from what typical theories or previous author have found) what your Önding is di§erent from expected. (Note that there is only right or wrong method; there is no ërightíor ëwrongíresult if the method is right. It would just be what the data say!)

● Robustness Check:  this part checks ërobustnessíof the Önding established with your benchmark regression.   You can check, e.g.   whether the benchmark regression has problems of multicollinearity, heteroskedasticity or Autocorrelation; or whether some of the RHS variables are endogenous so that an IV estimation would be better; or whether a dummy variable could be added to improve the regressionís Öt  (e.g.   the global Önancial crisis around 2008, the Covid pandemic since 2020Q1).  But it is impossible to consider everything. In practice, people would just consider one or two aspects. See below what are required for our essay.

● Conclusion: the brief round up of what you have done and found.

When reporting your regression result: create your own table and only report the relevant information (Do not copy the Eviews tables directly which would include too much irrelevant

information). Using the Eviews Ögures is Öne but you may like to use your own Ögure/variable names etc.

Note this is an essay, not an Eviews assessment.  You only use Eviews as a ëcalculatorí. Therefore, Eviews screenshots should not be added to your essay.

Topic: What determines ináation in [Country X]?

1.  Choose a country you are interested and investigate what factors a§ect ináation in that country. Length requirement: 2000 words (±10%) excluding tables and references.

2. DeÖnition of ináation:  percentage change in the general price level, where the price level may be measured by the GDP Deáator, CPI, RPI, or HICP in the euro area. Feel free to use any of these. When percentage data is used in Eviews, use Ögures without the % sign; e.g. if the value is 2%, put 0.02 in Eviews.

3. Determine your own benchmark regression.  The LHS will be ináation anyway.  The RHS factors are chosen by yourself. But you should consider a minimum of 3 factors in this benchmark regression.

4.  Collect the data by yourself.  Both annual and quarterly data can be used.  But a minimum of 60 observations is required and, if quarterly data are used, they should not be older than 2000Q1.

5. In your data session:  use the ADF test (with a constant, but no trend) to test data stationarity.  If any time series needed by the regression is found non- stationary at the 10% signiÖcance level, use the ëlinear detrendingímethod taught in L15 to stationarise that time series (Tips: a. if cointegration exists in your benchmark regression, you donít need to detrend the data as that would be the ëspecial caseí; b. if there is no cointegration and if linear detrending doesnít help, consider measuring your variables in growth rate, as growth rates are normally stationary; c.  if no luck with all the above, simply use another variable).

6. In your Robustness Check session:  consider TWO of the following poten- tial  issues (all are to be compared to the benchmark regression):  a) proportional heteroskedasticity (Select only one RHS variable to discuss), b) Autocorrelation (Con- sider up to 4 lags for the error term), c) endogeneity of RHS variable (Select only one variable to discuss). For all these issues, consider how to test their existence and how to deal with the issues (if they exist). In any case, discuss how your benchmark Önding may be a§ected (i.e., whether it is robust to these issues).

7. All equations in your work should be numbered. A list of references (any style) should be provided.

8. Databases: use any free, publicly available sources you can Önd online.  The several free resources used by many in the Öeld of economics are: https://fred.stlouisfed.org/ (Mainly for US but also include data for main economies), https://www.ons.gov.uk/ (Mainly for UK), https://ec.europa.eu/eurostat/data/database (Mainly for European