Estrategia de inversión de portafolios de TES basada en estimaciones de la tasa de redescuento del Banco de la República

Abstract

This work aims to develop an active investment strategy to be applied to a portfolio compound of TES due July 24 2.020 (that implies to buy or sell them continuously) based on estimations of the discount rate of the “Banco de la República”. This strategy will be compared with a passive one, in which no assets rotate, and will be defined which one has the greatest yield. In the construction of the tools that will be used, monthly figures of Colombia's CPI and money were taken, from January 1996 to June 2011, and their trend, cycle, seasonality and random changes were calculated, using the classical model of time series decomposition. Models between IPC cycle and the cycle of M1 were ran, and the inverse relationship between the cycle of CPI and the cycle of M1 wasdetermined, a conclusion already validated by other authors for Colombia. Later models between the discount rate of the “Banco de la República” and the cycle of IPC were ran, where a direct relationship between them was demonstrated. Given this result, the model for the relationship discount rate of the “Banco de la Republica” = f (CPI cycle) were ran, to infer changes in the discount rate for the next month; the graph of the cycle of the PCI was also used with equal claim. Based on these forecasts, decisions about buying or selling a portfolio of TES due July 24, 2020, with original face value of $ 1,000 million, were made. The final values at market prices obtained during the period using both methods were compared with that of a portfolio with the same TES reference which title did not rotate during this period. This methodology was the "back testing" of the tools. As a main conclusion it can be emphasized that in Colombia the developed tools are useful to manage structural portfolios, as its application produces significantly higher revenues than those obtained when the portfolio is not rotated.