Ole E. Barndorff-Nielsen

Overview

Ole E. Barndorff-Nielsen (1935–2022) was a statistician and probabilist at the Department of Mathematical Sciences, University of Aarhus (now Aarhus University), Denmark. He is known for foundational contributions to theoretical statistics (exponential families, information geometry, saddlepoint approximations) and to mathematical finance, particularly the development of realized volatility theory and Lévy-based stochastic volatility models in collaboration with Neil Shephard.

Key Contributions

• Barndorff-Nielsen (1998): Introduced the normal inverse Gaussian (NIG) Lévy process as a model for log-price increments; NIG has semi-heavy tails and a closed-form characteristic function, enabling fast option pricing.
• Barndorff-Nielsen (2001): Showed that a superposition of independent OU-type processes driven by subordinators generates long-range dependence in the autocorrelation function of squared observations; established the theoretical basis for long-memory SV via superposition.
• Barndorff-Nielsen and Shephard (2001): Non-Gaussian OU-based SV models driven by Lévy subordinators; proved that realized variance (RV) converges to integrated variance as sampling frequency increases; derived second-order properties of $\tau_t = \int_0^t \sigma_u\,du$. JRSS-B 63: 167–241.
• Barndorff-Nielsen and Shephard (2002): Asymptotic distribution theory for realized volatility; QL estimation of SV model parameters from RV time series; Kalman filter and Durbin algorithm implementations. JRSS-B 64: 253–280.
• Barndorff-Nielsen and Shephard (2004): Power and bipower variation — $BV_t = \tfrac{\pi}{2}\sum_j|r_j||r_{j-1}|$ consistently estimates integrated variance $IV_t$ even with finite-activity price jumps; the difference $RV_t - BV_t$ consistently estimates total jump variation. JFEconomics 2(1): 1–37.
• Barndorff-Nielsen and Shephard (2006): Time-changed Lévy processes; RV is an inconsistent estimator of the time-change whenever $Z$ is non-Brownian ($k_4 > 0$); the irreducible variance component $k_4\Delta\xi$ does not vanish with frequency; RV ACF underestimates true variance ACF; 4th cumulant test for symmetry-conditional jump presence; QL estimation hierarchy (OU, $J$-factor, LNOU, long-memory). See Barndorff-Nielsen and Shephard (2006).

Related

• Neil Shephard
• Realized Volatility
• Time-Changed Lévy Process
• Stochastic Volatility
• Barndorff-Nielsen and Shephard (2006)