Foundation course in quantitative foundations
Financial Econometrics
Financial econometrics for people who will test trading and investment ideas: regression done properly on returns, the standard errors that survive real data, panels, event studies, and the statistics of searching through many strategies.
Financial econometrics is how a researcher tells a real effect from a lucky one. It takes regression and hypothesis testing and adapts them to returns, which are noisy, fat-tailed, dependent over time and searched through by thousands of people for the same patterns. This course teaches the methods and the discipline.
Why do most backtests fail out of sample?
Mostly because many ideas were tried and the best one was kept. A strategy chosen from a hundred candidates will look good by chance. The final unit shows how to count the trials, adjust the statistics for them and estimate how much of a backtested Sharpe ratio is likely to survive. Employers ask about this in interviews because it costs them money when it is ignored.
How is the course taught?
Audio lessons carry the reasoning, explainer films show the estimators at work on real data, and the workbooks hold worked examples, Python notebooks and cheat sheets. The empirical project is marked by the faculty with written feedback.
Who is this course for?
Anyone who will test trading or investment ideas on data. It is part of the Foundation Certificate in Computational Finance, which opens the Professional Diploma in Algorithmic Trading and the Fellowship in Computational Finance.
What you will be able to do
- Estimate and interpret regressions of returns, and choose standard errors that are robust to the features of market data.
- Run an event study and a panel regression and report them correctly.
- Test for stationarity and explain why regressing one trending series on another misleads.
- Correct for multiple testing and judge whether a backtested result is likely to be luck.
Syllabus, module by module
Regression on returns
Ordinary least squares, what the coefficients mean, and the assumptions market data breaks.
Standard errors that survive
Heteroskedasticity, autocorrelation, Newey-West and clustered errors.
Cross-sections and panels
Fama-MacBeth regressions, fixed effects and pooled panels of stocks over time.
Event studies
Abnormal returns around earnings, index changes and announcements.
Stationarity and spurious regression
Unit roots, differencing, and why levels of prices are dangerous to regress.
Many tests, few discoveries
Data mining, p-hacking, the deflated Sharpe ratio and false discovery control.
How the certificate is earned
Quizzes at the end of each unit and a graded empirical project: an event study or a cross-sectional return test written up as a short research note. On a pass you receive the Course certificate in Financial Econometrics, issued by the Institute with a verification code. Read how certification works.
Before you start
Probability and statistics at the level of the Institute's foundation courses, and basic Python.
This course counts toward
- Foundation Certificate in Computational FinanceThe entry quantitative finance certificate and the first semester of both four-semester credentials. Nine courses that settle the mathematics, statistics, data and market knowledge every later course assumes, closed by a capstone.
- Professional Diploma in Algorithmic TradingThe top credential on the practitioner path. Four semesters, 41 courses and a 160-hour capstone covering everything between the mathematics and a book you could defend to a risk committee.
- Fellowship in Computational FinanceThe top credential on the researcher path. Four semesters, 38 courses, and a dissertation supervised in writing and examined by two faculty members.
Questions applicants ask
What is financial econometrics?
It is statistics applied to financial data: estimating relations between returns and their drivers, and testing whether those relations are real. It differs from textbook statistics because market data is noisy, fat-tailed and dependent over time.
How is this different from Time Series Analysis for Finance?
This course covers regression, panels, event studies and testing. Time Series Analysis for Finance, at practitioner level, goes deeper into forecasting models such as ARIMA and volatility models.