Template-Type: ReDIF-Article 1.0
Author-Name: Alexandros M Goulielmos
Author-Name: Marcos A Goulielmos
Title: THE PROBLEM OF TIMING IN DECISIONS TO BUY OR TO CHARTER A VESSEL
Abstract: This paper dealt with the concept of timing in reaching decisions, especially in connection with freight markets and sale and purchase of second hand ships. Timing is “the
calendar time we accept as right just before we take a decision”. First, an extensive literature search has been attempted in such disciplines as Management, Decision- making and Decision Support
Systems. Unfortunately, traditional disciplines have frequently offered the excuse for not dealing with timing of decision- making due to ‘lack of information’ and to ‘limited predictive techniques’.
Moreover, modern disciplines (in particular chaos theory) also excuse themselves from addressing the question of timing on the grounds that real phenomena are ‘inherently unpredictable’. The
concept of time in finance has been presented here as applied by Mandelbrot (1997) [and Albert Einstein (in 1905)] and stressed that time is flexible. Moreover found that time series have a speed
and a long-term memory. A non-parametric method called ‘Rescaled Range Analysis’ applied which indeed deals with cycles and long term memories. The research directed also to the discipline
of modeling and forecasting, both classical and chaotic (fractal). This brought us to address the question : should ARIMA (random) or ARFIMA (fractal) models be here and in similar
applications employed ? The final conclusion was that Rescaled Range Analysis was the most appropriate for the analysis of both freight market and second hand ship price market. Additionally,
the normality test by Jarque-Bera test (33>5.99) for freight market has shown the absence of normality and the existence of excess kurtosis (-1.51) as well as excess skewness (0.21). Moreover,
a strong memory in freight rates of H=0.92<1 (H is the Hurst exponent, indicating a persistent time series or black noise) and a fractal dimension of 1.084 as well as alpha=1.092 were found. The
‘Joseph effect’ of Mandelbrot was detected, too. A discussion is then made about shipping cycles. Non-periodic freight cycles were detected from 17 days to 33, and from 80 days to 160. Also, the
second hand 85000 dwt tanker market showed cycles of 4 and 8 years in the past. In this paper second hand price cycles for Aframax were detected of 26 and 78 months. Moreover, the freight
rates were predicted round 5000 units of the BPI index that would prevail in March-April 2007.This prediction proved to be accurate, as the actual level of the index was 5031 in March and 5390 in
April 2007 (as shown by data collected after the bulk of this paper had been written). Finally, we concluded after all that the best timing for decision-making is that which follows a best forecasting
which we have tried finally.
Volume: 36
Issue: 2
Year: 2009
Handle: RePEc:jte:journl:2009:2:36:4