This paper presents the results of the functional data analysis of the prices of the Ibex 35, Dow Jones Industrial Average, DAX and Nasdaq since their inception to December 2016, aiming to identify temporal patterns that can be incorporated into the investment decision-making process. It has been have identified Kondratieff cycles (44-60 years), Juglar cycles (c. 9 years), Kitchin cycles (c. 3.4 years), 2-year cycles, 22-26-week cycles, 30-calendar-day cycles (also known as Lunar cycle) and 3-calendar-day cycle (all found in the four markets except for the Kondratieff cycle in the Ibex-35, due to the lack of sufficient data). In addition, this research shows that the model results are improved if the possibility of variation of +/-10-30% in the period of the cyclical movement is incorporated. Finally, this research explores the application of phase plane plots to the seasonal analysis, detecting as months of change April to June and October to November. This conclusion is consistent with the seasonal pattern "Sell-in-May-and-Go-Away".
Descriptors: Temporal patterns, stock index, Kondratieff, Kuznets, Juglar, Kitchin, Principle of Variation, Lunar Cycle, Functional Data Analysis and Phase Plane Plot.
Universidad Pontificia Comillas. Madrid (España)
19 June 2017