Empirical analysis of daily cash flow time series and its implications for forecasting

TL;DR

This paper analyzes real-world daily cash flow data from small and medium companies, revealing that common assumptions like normality and stationarity are often invalid, and proposes a new forecasting approach that outperforms classical methods.

Contribution

It introduces a new cross-validated test for non-linearity in cash flow time series and demonstrates the importance of data-driven forecasting strategies for cash management.

Findings

Normality and stationarity assumptions rarely hold in real cash flow data.

Non-linearity is often significant and relevant for forecasting accuracy.

Proposed forecasting method outperforms classical approaches.

Abstract

Cash managers make daily decisions based on predicted monetary inflows from debtors and outflows to creditors. Usual assumptions on the statistical properties of daily net cash flow include normality, absence of correlation and stationarity. We provide a comprehensive study based on a real-world cash flow data set from small and medium companies, which is the most common type of companies in Europe. We also propose a new cross-validated test for time-series non-linearity showing that: (i) the usual assumption of normality, absence of correlation and stationarity hardly appear; (ii) non-linearity is often relevant for forecasting; and (iii) typical data transformations have little impact on linearity and normality. Our results provide a forecasting strategy for cash flow management which performs better than classical methods. This evidence may lead to consider a more data-driven approach such as time-series forecasting in an attempt to provide cash managers with expert systems in cash management.