City Research Online

Abstract

Technical analysis is the study of price movements in traded markets so as to forecast future movements or identify trading opportunities. Following a review of the history and research of technical analysis, three empirical chapters evaluate a number of propositions popular among technical analysts.

One approach used widely over the last century assumes that support and resistance levels can be predicted by projecting the ratios between the length and duration of successive trends, in particular using Fibonacci ratios like 1.618. This proposition is rejected for the Dow Jones Industrial Average by identifying turning points and testing for clustering by developing a block bootstrap procedure. A few significant ratios appear to support such anchoring by the market, but no more than would be expected by chance.

The thesis then reports a survey based experiment that tests whether individuals themselves do have an in-built tendency to anchor forecasts of future trends on previous trends. The significance of the survey results are tested using a novel kernel density estimator based bootstrap methodology. Respondents' forecasts do bear some relationship to the size of the most recent trend by certain whole-number ratios by more often than would be expected by chance.

The third experiment addresses the criticism that academic studies do not use a rich enough characterisation of technical analysis. 120 active market-timing strategies are tested using a regression based framework of equity fundamentals, macroeconomic fundamentals, behavioural variables and a diverse set of mainstream statistical indicators from technical analysis. Our recursive approach uses time-invariant rolling and expanding estimation windows as well as conditional windows based on the presence of structural breaks, identified using the conditional reverse ordered cusum method (ROC), of Pesaran and Timmermann (2002). Models that include both fundamental and technical indicators perform well, even allowing for realistic levels of transactions costs. And accounting for structural instability via the ROC method also improves performance.

Publication Type:	Thesis (Doctoral)
Subjects:	H Social Sciences > HG Finance
Departments:	Bayes Business School Bayes Business School > Finance Doctoral Theses Bayes Business School > Bayes Business School Doctoral Theses

Export

Downloads