Estimating Volatility in the Merton Model: The KMV Estimate is not Maximum Likelihood

Benjamin Christoffersen, David Lando*, Søren Feodor Nielsen

*Corresponding author af dette arbejde

Publikation: Bidrag til tidsskriftTidsskriftartikelForskningpeer review

21 Downloads (Pure)

Abstrakt

We compare two methods for estimating the asset volatility in the Merton model using observed equity prices: maximum likelihood and an iterative method commonly referred to as the KMV method. The two methods often yield extremely similar estimates, which has led to the conjecture that the two methods are equivalent. We show that this is not true and we provide a necessary and sufficient condition that the inverse of the equity pricing function would have to satisfy for the two methods to be equivalent. Moreover, we show numerically that this condition is very close to being true for in-the-money options.
OriginalsprogEngelsk
TidsskriftMathematical Finance
Vol/bind32
Udgave nummer4
Sider (fra-til)1214-1230
Antal sider17
ISSN0960-1627
DOI
StatusUdgivet - okt. 2022

Bibliografisk note

Published online: 30 August 2022.

Emneord

  • Distance-to-default
  • EM-algorithm
  • KMV method
  • Merton model

Citationsformater