# How to write an economics research paper

These are general guidelines for writing a research paper with international economics data. They help you structure your research work and avoid common pitfalls so that you can progress rapidly and write a solid paper.

The first step in the process is to read the literature and to summarize:

1. The main findings.

2. The data and empirical methodology used.

Write a literature review that is about 3-4 pages long. Organize it around ideas, not around individual papers. For example, if the question is about the determinants of banking crises describe what the literature has found about the effect of credit booms, then about prudential supervision, then about the link between crises and economic growth, then about the role of currency depreciation, etc. See how various papers link to these topics. By focusing on key ideas you see more clearly what is missing in the literature and can find out how to contribute.

Based on the literature review, spell out in 2-3 paragraphs what you will do and with what data. Explain why this is novel and how it extends the literature. Convince the readers that what you do is important.

Make two tables:

1. One table with a list of the variables used in the analysis, their definition, coverage in terms of countries and years, and the sources.

2. A second table with summary statistics: mean, standard deviation, maximum, and minimum for each variable. Also include variable correlations with an indicator for the statistical significance of each pair-wise correlation coefficient.

Look at your data and these two tables and describe how many observations you can use across how many countries and years; whether there are any outliers (i.e. extreme values); whether the correlations make sense, etc. Your objective is to start the empirical analysis with data you understand. Don’t run regressions without first making sure your data are clean and make sense.

Stay very close to the empirical models used in the literature. Your contribution could be that you use the same models but with new data. Or it could be that you introduce one or more new explanatory variables that have not been used in the literature. You may be looking at a region or a country that has not been investigated recently or at all. The important thing is to base your analysis on the best approaches in the literature and to make one step forward.

Think whether you have endogeneity issues, i.e. whether the dependent and independent variables are jointly determined. If so, how do you plan to address that problem – lagging variables, instrumental variable estimations? Do you have omitted variable problems, i.e. are you including all variables that should be included? Any other econometric issues?

What robustness checks do you plan to carry out? You can estimate empirical models with a few key explanatory variables and then include more variables in a bigger model. You may consider re- estimating the models without outliers. You could estimate the specifications with subsamples based on different income groups or time periods. You could allow for autocorrelation, heteroscedasticity, random or fixed effects, etc. The point is to estimate enough different models so you are convinced that your key results hold true.

When you are satisfied that you have interesting and robust empirical results, start putting together the paper. Begin with an introduction that explains what you do and how it contributes to the literature. Your contribution should be spelled out directly on the first page of the paper. Then, in section 2 describe your data, in section 3 explain the empirical approach, in section 4 describe your results, and conclude in section 5. Next, add the list of references ensuring consistent citation style. Then come the appendix tables and figures if you have any, then the tables with results.

To describe the results, look at the tables and draw the most important conclusions. Explain them. Note if you find something peculiar or particularly interesting. Don’t describe all the results in great detail as readers get lost. If they are interested they can look at the tables.

The conclusion should be 2-3 paragraphs where you summarize your main results, explain what could be done better in future research, and what are the implications of your findings for policy.

Then, get feedback and keep revising until the paper is polished.

Let us know if you need assistance with a research project. We can help with constructive and specific feedback and suggestions on data analysis, presentation, style, and language. Contact us using the online form.

**Literature review focused on ideas, not on papers**The first step in the process is to read the literature and to summarize:

1. The main findings.

2. The data and empirical methodology used.

Write a literature review that is about 3-4 pages long. Organize it around ideas, not around individual papers. For example, if the question is about the determinants of banking crises describe what the literature has found about the effect of credit booms, then about prudential supervision, then about the link between crises and economic growth, then about the role of currency depreciation, etc. See how various papers link to these topics. By focusing on key ideas you see more clearly what is missing in the literature and can find out how to contribute.

**Explain your contribution**Based on the literature review, spell out in 2-3 paragraphs what you will do and with what data. Explain why this is novel and how it extends the literature. Convince the readers that what you do is important.

**Know your data**Make two tables:

1. One table with a list of the variables used in the analysis, their definition, coverage in terms of countries and years, and the sources.

2. A second table with summary statistics: mean, standard deviation, maximum, and minimum for each variable. Also include variable correlations with an indicator for the statistical significance of each pair-wise correlation coefficient.

Look at your data and these two tables and describe how many observations you can use across how many countries and years; whether there are any outliers (i.e. extreme values); whether the correlations make sense, etc. Your objective is to start the empirical analysis with data you understand. Don’t run regressions without first making sure your data are clean and make sense.

**Empirical model grounded in the literature**Stay very close to the empirical models used in the literature. Your contribution could be that you use the same models but with new data. Or it could be that you introduce one or more new explanatory variables that have not been used in the literature. You may be looking at a region or a country that has not been investigated recently or at all. The important thing is to base your analysis on the best approaches in the literature and to make one step forward.

Think whether you have endogeneity issues, i.e. whether the dependent and independent variables are jointly determined. If so, how do you plan to address that problem – lagging variables, instrumental variable estimations? Do you have omitted variable problems, i.e. are you including all variables that should be included? Any other econometric issues?

What robustness checks do you plan to carry out? You can estimate empirical models with a few key explanatory variables and then include more variables in a bigger model. You may consider re- estimating the models without outliers. You could estimate the specifications with subsamples based on different income groups or time periods. You could allow for autocorrelation, heteroscedasticity, random or fixed effects, etc. The point is to estimate enough different models so you are convinced that your key results hold true.

**First draft**When you are satisfied that you have interesting and robust empirical results, start putting together the paper. Begin with an introduction that explains what you do and how it contributes to the literature. Your contribution should be spelled out directly on the first page of the paper. Then, in section 2 describe your data, in section 3 explain the empirical approach, in section 4 describe your results, and conclude in section 5. Next, add the list of references ensuring consistent citation style. Then come the appendix tables and figures if you have any, then the tables with results.

To describe the results, look at the tables and draw the most important conclusions. Explain them. Note if you find something peculiar or particularly interesting. Don’t describe all the results in great detail as readers get lost. If they are interested they can look at the tables.

The conclusion should be 2-3 paragraphs where you summarize your main results, explain what could be done better in future research, and what are the implications of your findings for policy.

Then, get feedback and keep revising until the paper is polished.

**Need assistance with a research project?**Let us know if you need assistance with a research project. We can help with constructive and specific feedback and suggestions on data analysis, presentation, style, and language. Contact us using the online form.