Part 1: Econometric strategy
This part is basically the methodology for the extension. We need to present the Econometric approach that we are going to use for the Stata and the extension.
I want to test the effect of Job Polarisation on Wage Inequality. For this, I want to use the following two regressions using Panel Data:
1. The evolution of employment
2. The evolution of wages
Test this in different cycles: (If you think this is the appropriate way of separating times)
2002-2007: The housing cycle and economic growth
2008-2009: Great Recession
2010-2017: After the crisis
Discuss estimation approach, including estimation equations
o Explain all components
o Why is this approach suitable? Why are we using Panel Data for our extension?
What are the empirical challenges to answering the research question?
o Econometric problems; challenges to identification
" Potential endogeneity, omitted variables, causality…
" How will the issues be addressed?
Part 2: Extensions
Main steps per extension (evolution of wages and evolution of employment)
1. Introduce and discuss extension (also formally)
2. Show estimation results, discuss them (also sensitivity analyses, tests, etc.) interpret them
– Data: Introduce DATASET and VARIABLES; definitions and measurement of concepts (like wage etc.)
– Background: institution background (as far as relevant); if possible some stylised facts
– Descriptive Analysis (comparing also full population and estimation sample)
– Use GRAPHS and TABLES to concisely present the analyses (they need to be numbered and be self-explanatory)
o Policy implications- minimum wage, what is the government doing to mitigate wage inequality (see document attached about California)
I have attached a document about writing an economic research paper
o Scientific style
o Do take care with presenting tables and figures
" All tables & figures have a title and numbering
" Clean, condensed presentationílook at journal articles
" Tables & figures should be self-explanatory (thus, care about title, notes etc.)
" Always give standard errors (or t-ratios) and, if appropriate, diagnostic test statistics. Use asterisks for statistical significance
" Don’t just cut and paste Stata outputs for tablesíuse outreg2 or similar
o Do discuss and reflect your results
" Interpretation of coefficient and size of effects; explain tests etc.
" Do you think your results are plausible and robust? Are there any remaining issues of concern?
" Compare different regressions/spec.s (e.g., stepwise inclusion of variables)
" In text, e.g.: Hoynes et al. (2015); usually no direct quotes o See referencing systems in econ journal articles