1 Why a separate discipline?

  • Millions of data emerge in business and finance every day, and econometrics, along with statistics, helps us organize, summarize, and analyze this data properly

  • The appropriateness of econometric analysis depends on data type, model specification, model assumptions, goodness of fit, etc.

Financial econometrics is designed for quantitative finance users, providing specialized analytical skills required for:

  1. asset pricing,
  2. risk managing,
  3. portfolio optimization,
  4. volatility and co-volatility forecasting,
  5. developing hedging strategies,
  6. options pricing and derivatives,
  7. monetary policy governance, etc.

  • Financial econometrics is particularly beneficial to financial markets participants, including banks, insurance companies, investment funds and regulatory agencies, as well as monetary policy makers

  • Econometric analysis of financial time–series data such as stock returns, cryptocurrency prices, exchange rates, interest rates or inflation, is complex due to their unique empirical properties, and hence requires a special attention

  • For instance, financial time–series data typically exhibit non-normality (heavy tails) and conditional heteroscedasticity (time-varying variance)

  • Unlike traditional econometrics, the modern approach does not treat non-normality and heteroscedasticity as problems, but instead focuses on developing new methods and models that can accommodate or adjust to these issues

The majority of financial econometrics techniques focus on measuring, modelling, and forecasting volatility, a crucial risk parameter that is unobserved in practice (but can be estimated)

What is definition of volatility?
Volatility is the standard deviation of returns, representing the percentage of price fluctuations for a given asset, portfolio, or market over a specific time period

  • Different types of volatility are used in practice according to their role and purpose
TABLE 1.1: Volatility types
Type Description Measure
Historical volatility Point estimate of volatility using past returns
(sample standard deviation or sample variance)		 ex-post
Implied volatility Reflects market expectations of future volatility,
derived from option prices (e.g. BSM, MLN, etc.)		 ex-ante
Conditional volatility Dynamic volatility conditioned on recent information
and often used for forecasting (GARCH type models)		 ex-ante
Realized volatility Similar to historical volatility but uses high frequency
returns to estimate volatility at lower frequency		 mostly ex-post

  • Realized volatility is estimated ex-post but can be predicted out-of-sample

  • Besides volatility, co-volatility also draws the attention of researchers and practitioners, as it captures the relationship between multiple assets, commonly measured by the covariance of returns, which can be estimated using various methods

  • It is typical for volatility from one asset to spillover to another asset

  • When spillovers are quick (sudden) and occur during crisis periods, it is evidence of contagion, indicating that markets share common negative trends