Ole E. Barndorff-Nielsen: Sand, Wind and Inference

This paper reviews Ole Eiler Barndorff-Nielsen’s research in the first decades of his career. The focus is on topics that he kept returning to throughout his scientific life, and on papers that he built on in later important contributions. First his early contributions to the foundations of statistical inference are reviewed with focus on conditional inference and exponential families, two topics in which he had a lifelong interest. The second half of the paper reviews his research on wind blown sand and hyperbolic distributions and processes, including his early contributions to modelling of turbulent wind fields. This research laid the foundations for his later work on financial econometrics and ambit processes.

💡 Research Summary

This paper offers a comprehensive review of Ole E. Barndorff‑Nielsen’s early scientific contributions, organized around two enduring themes: (1) the foundations of statistical inference, especially conditional inference and exponential families, and (2) the modelling of wind‑blown sand and hyperbolic distributions and processes. In the first half, the author traces Barndorff‑Nielsen’s pioneering work on the theory of ancillary statistics and sufficiency. Building on the Rasch measurement model, he introduced a hierarchy of concepts—B‑sufficiency/B‑ancillarity, S‑sufficiency/S‑ancillarity, G‑sufficiency/G‑ancillarity, and later L‑sufficiency/L‑ancillarity—each designed to make the conditionality principle operational in the presence of nuisance parameters. The paper illustrates these ideas with the classic Neyman‑Scott problem, showing how the statistic SSD is L‑sufficient for the variance parameter while the sample means are L‑ancillary, thereby yielding an unbiased, consistent estimator that overcomes the bias of the naïve maximum‑likelihood estimator.

Barndorff‑Nielsen’s exact theory for exponential families is also summarised. Assuming a minimal canonical representation (f(x;\theta)=b(x)\exp{\theta\cdot t(x)-\kappa(\theta)}), he distinguished full, regular, and steep families, proved that the mapping (\tau(\theta)=E_\theta