The CONSULTANCY PROJECT will develop a working knowledge of economic models, provide an appreciation of how to assess economic policy through analysis of economic data

The CONSULTANCY PROJECT will develop a working knowledge of economic models, provide an appreciation of how to assess economic policy through analysis of economic data, and nurture skills that will be helpful for independent research. Students will develop an ability to think analytically and learn how to present written work in a logical and coherent manner 

Assignment

You need to choose a relevant topic/issue, which needs to be approved by the module leader. Once your topic is approved you are required to do the following:

• Formulate a suitable model.
• Describe and discuss data and methodology.
• Estimate the model.
• Analyse the results.

You must choose a topic that is suitable for the project. Essentially, this means you must be able to formulate a model and you must be able to estimate the model with suitable data. It is no good choosing a topic for which there are no suitable data. Getting access to suitable data is key. Below are a few suggestions for topics. One approach would be to explain variations across countries in terms of saving rates or remittances or suicide rates or pollution levels. Another approach would be to focus on a single country and examine sub-national (e.g. regional, county, local authority) variations in crime rates or house prices. Or your approach may be more microeconomic based where you focus on explaining variations across individuals (e.g. salaries of US baseball players) or firms (profits).

Given the time constraint my advice is to make use of secondary data that is readily available online rather than primary or micro-based data that needs to be collected from organisations. For a cross country investigation, World Bank Indicators is a good source as well as Eurostat (if you are interested in European regional data) and OECD. For the UK, the Office for National Statistics (ONS) data is very useful.

Guidelines 

(i) Formulate a suitable model

The model needs to reflect the aim(s) of the project and it should be based on relevant literature. Begin by identifying articles which focus on your topic. For example, if you wish to look at the determinants of savings across countries, do a literature search on savings (using the library search engines as well as Google Scholar). Use the articles to identify a model that you think best captures the determinants of saving. You must carefully justify the inclusion of each variable in the model along with any other details that might be relevant. 

(ii)    Data and methodology

Once you have a model(s), you need to create a data set and decide on the methodology for estimating the model. This component of the project covers all data related issues. This includes, but is not limited to, data sources, definition of variables (including the unit of measurement of each variable), type of data (firm, country, individual etc.) and number of observations. Also, point out any problems encountered, for example if there are missing data. Summary statistics, basic plots of variables (if relevant) and scatter diagrams might be useful. The relevant methodology (estimation method) must be justified. This includes making a case that key ordinary least squares (OLS) assumptions are satisfied (for example, stating that the error terms are uncorrelated with the regressors). 

Note that the empirical economics module has only considered cross-sectional data (data that varies across countries, regions, counties, firms, individuals, and so on) and cross-sectional estimation.

Hence, your data and estimation method should be cross-sectional.

(iii)  Estimation of model

You should consider issues that are relevant in estimating the model, for example, the functional form. This includes (but is not limited to) using logarithms, quadratics and interaction terms. Note that the relationship between economic variables is often non-linear and to this extent the appropriate functional form is especially relevant. Depending on the nature of the topic and the aim of your project, it may be better to rescale variable(s) or use standardized variables. 

You need to be aware of including irrelevant variables or omitting relevant variables in the model(s). To this end you need to consider issues such as multicollinearity and misspecification. Other issues that warrant investigation include heteroskedasticity and structural breaks (stability). Note that the types of functional form mentioned above may help in alleviating “problems” in the model such as heteroskedasticity or misspecification. Investigating and correcting for these problems contributes to ensuring that the model(s) are diagnostically acceptable.

You should also consider whether it is appropriate to use qualitative (dummy) variables in the model or whether you wish to use quantitative proxies for data that is “missing”. Again, the justification must be made clear.

Depending on the topic, you should consider the robustness (sensitivity) of the results. For example, you may wish to use an alternative data series for the same economic variable, so use gross savings as well as net savings or use GDP alongside GNP.

(iv) Analysis of the results

You need to provide a detailed interpretation of the results. For example, are the variable coefficients in line with or in conflict with the a priori assumptions? Are they statistically significant? You should also discuss the outcomes of any additional hypothesis testing that you have undertaken (whether a population parameter equals 1 or a combination of parameters equals 0 or 1). If the results are different to a priori expectations, possible reasons should be suggested. You should also consider the magnitudes of the coefficients (the practical significance). If the t-statistic is small, is it because the coefficients are small or because the standard errors are large? You should be aware of the units of measurement of the variables when considering magnitudes. 

Interpretation is especially important in instances where logarithms, quadratics or interaction terms are used as well as when you have dummy variables or standardised variables.

It is also useful to compare your results to those reported in articles you have read (and which are included in your literature review). 

Other information

Your project needs to include the following sections (and have headings like those below):

1. Introduction

This is where you identify the aim(s) of the project (and specify your research question(s)) and the motivation for choosing the topic (why it is important or interesting).

1. Literature review

This should link to (i) above. Depending on the topic, there may be many relevant articles. You only have time and space to include a few of the main ones. In reporting the literature, you should mention the key features such as region or countries studied, the model specification, the data used, and the key results. You may choose to present the information by authors or themes. By the end of this section, it should be clear which hypotheses have emerged from reviewing the literature.

1. Model, data and methodology

Based on your literature review you should be able to specify an empirical model. Each regressor in the model needs to be justified (use the literature review) along with the relevant apriori assumptions. Pay attention to the form in which the model is presented (e.g. if it represents a population model there are no “hats”). Also, if you write equations with the error terms you will be able to justify the use of OLS. 

The discussion on the data and methodology is link to (ii) above.

You may also wish to discuss the diagnostics that will be carried out (not the actual results). This includes the name/type of test, the null hypothesis and any other details that you think is relevant 

1. Empirical results, interpretation and analysis

This links to (iii) above. You should also consider the best way of presenting your results. If the estimated models are not too cumbersome you can present the results in equation form (like in the lecture notes) with the standard errors below in parenthesis and additional diagnostics in the next lin

below. If, however you have several explanatory variables and a variety of diagnostic results, a table is more appropriate (like what you see in the articles in the literature review). 

The analysis of your results link to (iv) above.

1. Conclusion

In this section you should summarise your results (pointing to any particularly interesting findings) and note any implications (especially for policy if relevant). You could also consider the limitations of the research and how it could be extended. You should be careful to link the discussion in your conclusion to the aims/research question in the introduction: have you done what you intended to do?

Word Limit 

A leeway of 10 per cent is allowed for the length of your assignment; that is, the marker will read up to 4000 words only. The word limit includes footnotes, quotations and in-text references and citations, but excludes the assignment title and other information on the cover page, charts and graphs, tables, references in footnotes, appendices, and a reference list. 

Submission and plagiarism

Information on submitting assignments is given in the “Assessment Handbook” available on myJourney. It also includes advice on avoiding plagiarism and other forms of academic misconduct. Be aware that there is a penalty for late submission.

https://myjourney.hull.ac.uk/learner/course