Events

Jose Menchero, Bloomberg: Solving the "Curse of Dimensionality" Problem in Multi-Asset-Class Risk Models

Estimating a robust risk model risk for a portfolio that spans multiple asset classes is a challenging task due to the “curse of dimensionality” (i.e., the problem of estimating too many relationships from too few observations). While the sample covariance matrix is easily computed, it is susceptible to capturing spurious relationships that make it unsuitable for portfolio construction purposes. In this talk, we present a new approach for constructing risk models that span multiple asset classes. We also discuss the implications for portfolio risk management and portfolio construction....

Markus Pelger, Stanford: Interpretable proximate factors for large dimensions

This papers deals with the approximation of latent statistical factors with sparse and easy-to-interpret proximate factors. Latent factors in a large-dimensional factor model can be estimated by principal component analysis, but are usually hard to interpret. By shrinking the factor weights, we obtain proximate factors that are easier to interpret. We show that proximate factors consisting of 5-10% of the cross-sectional observations with the largest exposure are usually sufficient to almost perfectly replicate the population factors, even if these do not have a sparse structure. We derive...

Matthias Weber, Swiss Re: Concrete examples of trend analyses and forward-looking modelling in Swiss Re's underwriting

Abstract:

In insurance, underwriting performance is a function of exposures, losses relative to exposures and premiums relative to exposures. Getting losses and loss trends right (--> cost of goods sold) is critically important. A small estimation mistake typically has a large impact on the bottom line. Swiss Re is determining loss relevant trends using advanced analytics, often in collaboration with universities, government organizations, NGOs, rating agencies, consultants, investment management firms, lawyers, and others. Findings are used for both capital allocation and...

Mariana Olvera-Cravioto, UC Berkeley: PageRank on directed complex networks

Abstract: The talk will center around a set of recent results on the analysis of Google’s PageRank algorithm on directed complex networks. In particular, it will focus on the so-called power-law hypothesis, which states that the distribution of the ranks produced by PageRank on a scale-free graph (whose in-degree distribution follows a power-law) also follows a power-law with the same tail-index as the in-degree. We show that the distribution of PageRank on both the directed configuration model and the inhomogeneous random digraph does indeed follow a power-law whenever the in-degree does...

Lisa Goldberg, John Arabadjis, and Jeff Bohn to speak at Stanford's AI in Fintech Forum

Read the agenda here.

Start date: 2018-02-08 08:30:00 End date: 2018-02-08 17:30:00 Venue: Arrillaga Alumni Center, Mccaw Hall Address: 326 Galvez Street, Stanford, CA, 94305

Lisa Goldberg presents research paper "The Dispersion Bias" at UC Santa Barbara

Abstract: Estimation error has plagued quantitative finance since Markowitz launched modern portfolio theory in 1952. Using random matrix theory, we characterize a source of bias in the sample eigenvectors of financial covariance matrices. Unchecked, the bias distorts weights of minimum variance portfolios and leads to risk forecasts that are severely biased downward. To address these issues, we develop an eigenvector bias correction. Our approach is distinct from the regularization and eigenvalue shrinkage methods found in the literature. We provide theoretical guarantees on the...

Lisa Goldberg and Jeffrey Bohn to speak at Swissquote Conference 2017 on FinTech

From the event website: Innovation in financial technology (FinTech) has transformed the financial services industry over the past decade and the technological changes are ongoing. Continuous pressure to innovate will shape customer behaviours, business models, and the long-term structure of the financial services industry. This unprecedented interplay between finance and technology offers great potential for developing new financial services business models and products. The 8th annual...

Kellie Ottoboni, UC Berkeley: Nonparametric Risk Attribution for Factor Models of Portfolios

Factor models are used to predict the future returns of a portfolio with known positions in many assets. These simulations yield a distribution of future returns and various measures of the risk of the portfolio. Clients would often like to identify sources of risk in their portfolios, but this is difficult when factors influence the portfolio in nonlinear ways, such as when returns are measured on a log scale and when the portfolio contains nonlinear instruments. We develop a two-step method to partition risk among factors in a portfolio which accounts for these nonlinearities: first,...