Uncover Temporal Causal Interrelationships between Key Economic Indicators Through Peter-Clark Algorithm

Abstract This paper proposes a time-variant Peter-Clark (TV-PC) algorithm that extends the Peter-Clark (PC) framework by incorporating time-lagged data to reveal causal interrelationship network between major economic indicators in Australia. By using historical data and without relying on any prior economic knowledge, the developed TV-PC algorithm with time lag input can automatically generate temporal causal interrelationship structures that are difficult to discern qualitatively and quantitatively. Through extracting the generated causal structure, a succinct causal temporal relationship map can be created for each key economic variable. The temporal causal interrelationship maps offer actionable insights into a complex and ever changing economic landscape. The causal relationship revealed about major economic indicators such as CPI, interest rates and housing index can assist in formulating more effective and timely economic and monetary policies and strategies.