Topologically Mapping the Macroeconomy

Abstract

An understanding of the economic landscape in a world of ever increasing data necessitates representations of data that can inform policy, deepen understanding and guide future research. Topological Data Analysis offers a set of tools which deliver on all three calls. Abstract two-dimensional snapshots of multi-dimensional space readily capture non-monotonic relationships, inform of similarity between points of interest in parameter space, mapping such to outcomes. Specific examples show how some, but not all, countries have returned to Great Depression levels, and reappraise the links between real private capital growth and the performance of the economy. Theoretical and empirical expositions alike remind on the dangers of assuming monotonic relationships and discounting combinations of factors as determinants of outcomes; both dangers Topological Data Analysis addresses. Policy-makers can look at outcomes and target areas of the input space where such are not satisfactory, academics may additionally find evidence to motivate theoretical development, and practitioners can gain a rapid and robust base for decision making.

Preliminary Version - 24 pages

References in corpus (5)

• Topological Data Analysis of Financial Time Series: Landscapes of Crashes
• Pore-geometry recognition: on the importance of quantifying similarity in nanoporous materials
• Ball mapper: a shape summary for topological data analysis
• Cyclicality, Periodicity and the Topology of Time Series
• Fibres of Failure: Classifying errors in predictive processes