Loosely speaking, there are three types of data that economists analyze. The first type is microeconomic data (such as in census and surveys), collected for different units (such as households and firms) at a snapshot in time. The second type is macroeconomics data (such as inflation and unemployment), observations collected on the same variable every month or quarter over a designated time span of, say, 50 years. The third type is financial data such as stock returns, collected at very high frequencies (at every tick of the trading clock) also over a span of, say, 50 years. A generic feature of macroeconomic data is that they are correlated over time, and thus are forecastable. Some macroeconomic time series have trends (never return to the mean) while others are cyclical (recurring movements with similar patterns). Macroeconomists tend to be interested in learning about the series change over time, often with the goal of making predictions. In contrast, financial time series tend to be much less predictable, but they have "fat tails," meaning that big outliers occur quite often. Furthermore, financial time series tends to display "volatility clustering," meaning that the volatility of returns is high for extended periods, and then low for extended periods. Financial economists tend to be more interested in learning about the dynamics of volatility rather than the dynamics of mean returns per se because volatility is often used to measure risk, which is fundamental to portfolio management. Different statistical tools are thus required to analyze macro and financial time series separately.

Intended audience: Upper-level (junior and senior) economics concentrators.

Course Requirements: 5 problem sets (50%); a midterm exam (20%); a final exam (30%).

Class Format: 3 hours per week in lecture format.