Question

Search for this article on Google and write a summary in no more than 300 words

Financial econometrics } A new discipline with new methods

Robert Engle

Department of Economics, University of California at San Diego, 9500 Gilman Drive, La Jolla, CA 92093-0508, USA and Stern School of Business, New York University, 44 West 4th Street, Suite 9-160, New York, NY 10012, USA

Financial econometrics is simply the application of econometric tools to "nancial data. For many years, least-squares techniques provided satisfactory tools. Stock market forecasts, e$cient market tests, and even tests of portfolio models such as the CAPM and APT were essentially implemented with least squares on cleverly manipulated data sets. More recently, however, the "eld has developed its individual character as new statistical tools have been invented to analyze new questions.

Answer 1

• Financial econometric is simply, application of econometric tools to financial data. This has developed its individual character as new statistical tools have been invented to analyze new questions.
• First models were ARCH and GARCH, of Engle(1982) and Bollerslev(1986)and then the stochastic volatility models if Taylor(1986) and Harvey et al(1994)
• This collection of methods has generally been successful in parameterizing and estimating conditional densities. The most significant unsolved problem is the multi variate extension of many of these methods.
• The second important attention gainer is, development of methods to useintra-daily data and ultimately, transaction data called as Engle(2000)ultra-high-frequency data (UHF), to improve estimation of conditional densities.
• Another object ofinterestis the risk neutral conditional density which is the set of probability of returns orprices under which, an agent would value assets by their discounted expected value, with no arbitrage opportunities.
• 2 different strands of financial econometric have approached the option-pricing problem:

1. ​​​​​​option data-infers f*(arbitrage free pricing), m estimates of f and computers option pricing.
2. Closer to conventional econometric research.

• True data generating process is a continuous time diffusion process, hence, both empirical and risk neutral density can be computed. Models that are most convinient for option pricing, incorporate the assumption that options are redundant assets.
• To analyze UHF data, it is necessary to model, not only characteristics of trade, but also the timing.
• Dufour &Engle(2000) followed Hasbrouck's vector auto regression, show that the more frequency transactions, the greater volatility and price response to trades and transaction arrivals are predictable, based on economic variables such as a bid.
• These models may also serve as underlying data generating processes, empirical and risk neutral density fore casts and consequences, for option pricing. This is an important challenge for models of eMperical micro structure.