University of Twente Student Theses

Login

Identifying and trading reversals following a downward overreaction : the MOPOI trading algorithm

Odink, K.M. (2016) Identifying and trading reversals following a downward overreaction : the MOPOI trading algorithm.

[img] PDF
9MB
Abstract:The aim of this paper is to develop a trading algorithm that shows potential to disprove the weak form efficiency of the efficient market hypothesis. Such a trading algorithm provides researchers a new tool with which the weak form efficiency can be tested in future research. The trading algorithm that has been developed in this research has been based on the overreaction hypothesis (OH), as formulated by De Bond & Thaler in 1985. The overreaction hypothesis comprises of two sub-hypothesis: OH-I: “Extreme movements in stock prices will be followed by subsequent price movements in the opposite direction” (De Bond & Thaler, 1985, p. 795) OH-II: “The more extreme the initial price movement, the greater will be the subsequent adjustment” (De Bond & Thaler, 1985, p. 795) Based on the overreaction hypothesis, a reversal is expected subsequent to an overreaction. The trading algorithm that has been developed in this paper has been designed to identify overreactions on the underlying assets on which the trading algorithm is employed, and to enter a long position at the end of the identified overreaction, as at this point a reversal is expected to occur on the underlying asset. The trading algorithm that has been developed in this paper has been named the Martin Odink Pessimism Overreaction Identifier trading algorithm, or for short the MOPOI trading algorithm. In short, the MOPOI trading algorithm calculates and assigns a volatility level of downward volatility on the underlying asset to each trading day, based off of a look back period of 22 trading days. From these volatility levels a moving average is calculated, and a standard deviation is plotted from the moving average. Whenever the volatility level of a given trading day exceeds the corresponding level generated by the standard deviation off of the moving average of a series of volatility levels then overreaction is assumed on the underlying asset on that given trading day. This overreaction is assumed to end once the volatility levels no longer exceeds the level generated by the standard deviation off of the moving average. In order to work out the MOPOI trading algorithm, unidentified parameters had to be identified. These parameters included the number of days off of which the moving average is calculated, and the number of standard deviations with which the threshold level is being plotted over the moving average. Aside from these two parameters, more tests were performed as the MOPOI trading algorithm signalled overreactions, and the end of overreactions, at times that were counterintuitive or questionable. These counterintuitive and questionable signals are signalled in the following situations:  In times of long continuous upward movement of the underlying asset, a downward movement of a few tens of a percentage point could cause the MOPOI trading algorithm to signal an overreaction on the underlying market. This may occur as the moving average and the plotted threshold level over the moving average would have reached levels close to zero. A downward movement of tens of a percentage point in a strong upward market, intuitively, is not associated with a downward overreaction on the underlying asset. In order to account for this occurrence, a range of threshold levels have been tested. The threshold level represents a percentage point threshold that the underlying asset needs to decline before a trigger of the MOPOI trading algorithm associated with an overreaction on the underlying asset.  A questionable signal is provided by the MOPOI trading algorithm at times when an asset trades flat, or substantially decreases its steepness of downward movement, subsequent to a decline on the underlying asset, which is being associated with an overreaction on the underlying asset. In such moments, at times, the volatility levels retrace within the levels of the threshold level that has been plotted over the moving average. Therewith, an end is signalled to the overreaction, while no reversal is occurring. To account for these situations, two conditions have been tested that demand a confirmation of a reversal through demanding lower closes of previous trading days.  At times the algorithm signals to enter a long position at times when another long position is still open. The question, however, is whether multiple open long positions should be allowed, as these long positions are based on the same overreaction. A restriction to the number of simultaneously opened long positions have been tested in this paper, in order to investigate whether this is beneficial to the returns generated by the algorithm.
Item Type:Essay (Master)
Faculty:BMS: Behavioural, Management and Social Sciences
Subject:83 economics
Programme:Business Administration MSc (60644)
Link to this item:https://purl.utwente.nl/essays/69695
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page