2.A second source of topics is the academic literature. Google Scholar is the best place to search the academic literature. Type a keyword and it will return hundreds of articles. You may read an article  arguing  that  the  urban  land  use  is  determined  by  income,  population,  and  urban transportation conditions. Following this article, you can collect data from China City Statistical Yearbook 2018 on (1) urban population, (2) per capita income, (3) transport infrastructure, and (4) urban land use and analyze how the first three factors may affect urban land use.

3.A third source is textbooks in econometrics. Most econometric textbooks emphasize empirical examples or exercises. Thus, they provide a large pool of potential topics. The easiest approach is to take one of the problems and apply the empirical model to your own data.

4. Of course, your topics are not restricted to the sources mentioned above. I also encourage you to find your own topics through deep thinking. Deep thinking produces interesting research questions. To give an example, you may model housing price to be jointly determined by demand and supply factors. However, there are many of them. It is then your job to narrow down to a few major factors and collect data accordingly. These cannot be done without deep thinking. Even if you adopt a research question raised by others, deep thinking will help you refine the question and generate new insights. For instance, in the model of Chinese housing price, you may want to consider factors overlooked by others but may be important in the Chinese context, such as administrative hierarchy and geographical location. These factors may bring further insights into your results.

Below are a few exemplary topics:

.    Estimate aggregate production function using regional (province- or city-level) data.

.    Estimate determinants of pollutants emission using regional data.

.    Estimate determinants of housing price using regional data.

.     Estimate  β -convergence using national data.

.    Estimate the environmental Kuznets curve using national data.

Although there is no restriction to the scope of topics you may try, to ensure that you obtain meaningful results from the analysis, please adhere to the following principles.

1. Please make sure you test an economic model, rather than an accounting identity. An economic model is a hypothetical functional form (according to some theory) that describes how one variable is determined by other variables. The exact form of this function is unknown and must be estimated using real-world data. For instance, economists often view the entire economy as a factory, where  inputs  (capital  and  labor)  are  converted  into  outputs  (GDP)  using  a  certain technology. A commonly adopted functional form is the Cobb-Douglas one, i.e.,  Y  = AKaLβ , where  Y  stands for GDP,  K  for capital stock,  L  for labor input, and  A  is called the total factor productivity (TFP). In this formulation, the parameters  a  and  β  are unknown, which can be estimated using real-world data. Within the regression model, we can test whether the technology exhibits constant returns to  scale (a + β = 1), increasing resturns to  scale (a + β > 1),  or decreasing returns to scale (a + β  < 1).

Accounting identities, on the other hand, are known formulas that must be universally true. This statement has two implications. One, the parameters of the formula are all known, which means there is no need to estimate them. Second, the relationship must be always true for any data set.

Guidelines on designing the analysis

This module covers quite a few important methods, including the OLS regression model, test of a single parameter, test of joint hypothesis, test for heteroskedasticity and WLS, nonlinear model, instrumental variable, etc. You are expected to employ appropriate methods (potentially statistical methods not covered by this module) in your empirical analysis. Although there is no fixed rule for good research design, quality researches share these common features:

1.   The analytical framework is carefully chosen to answer the research question and to analyze the data.

2.   Alternative model specifications or extensions of the model are explored to extract further

information from the data, to address data problems, and to consolidate the main findings. 3.   The results are interpreted and analyzed in detail.

Please avoid these common mistakes among past students:

1. Trying all the regression models or analytical methods learned in this module. Please bear in mind that your ultimate objective is to answer research questions. The coursework is not supposed to be an  exercise on  everything you learn. Contents that are unrelated to the research question damage the quality of your work.

2.   Presenting the analytical results without much interpretation. It is the interpretation, not the numerical results generated by software that answers the research question. Without proper interpretation, the results make little sense.

3.   Copying the analytical framework of a past  student work that earned a high mark. Their analysis serves their research question and their data, which are different from yours. Blindly copying other students’ analytical framework often results in a poor report.

Guidelines on format

1.   I suggest no more than 2,000 words for the report (excluding frontpage, appendix, references, tables, and footnotes or endnotes). The mark is not explicitly linked to the word count.  However, if your report is very short, it is unlikely to meet the marking criteria (see marking  scheme).

2.   I recommend the following structure for the final report:

a.    Title;

b.   Motivation and research question

c.    Description of data sources, variable measurement, and empirical model (why the regressors are important determinants of the dependent variable and what are their expected signs)

d.   Presentation of analytical results, interpretations, and statistical inferences e.    Discussion of results and conclusion

f.    References (if any);

g.   Appendix (see below).

3.   All Stata code and regression output must be reported in the appendix, placed at the end of the report. You should also include figures and tables in the main text and tables should be formatted as those in the textbook (for example, Table 8.3, though you can skip the 95% confidence intervals). Please do not use screenshots of any kind in your report.

4.   Please use the accompanying MS Word template to prepare your final report. Please insert your digital signature as a picture in the cover page. Please do not alter the format (font, line spacing, page margin, etc.) of the first two pages of the document. Please submit your final report as a MS Word document. PDF files are not accepted.

Important Dates

Marking components and weight

Peer assessment and individual marks

Your individual mark will be jointly determined by the base mark (the mark awarded to the report), and peer assessment. In principle, the individual mark can be higher or lower than the base mark,  depending on your relative contribution. The weight (20%) determines the magnitude of these  deviations. The algorithm used by the Learning Mall peer assessment function is described here:

http://webpaproject.lboro.ac.uk/academic-guidance/a-worked-example-of-the-scoring- algorithm/.

Please place your evaluations truthfully, with full respect for your teammates' effort. I will not override peer assessment results even if disputes arise after the release of component marks.