CONSTRUCTING AN ARMA MODEL
PRESENTATTION OF THE DATA
The object of the hereby analysis are two time series. The first one consists of the daily closing prices of the BET index - the first index developed by BVB (Bucharest Stock Exchange) which acts as the reference index for this market. BET is a free float weighted capitalization index of the most liquid 10 companies listed on the BVB regulated market and started with a value of 1000 points. The data covers the period from its start, September 22, 1997, to May 24, 2011.
The second series consists of the closing prices of the SNP stock (expressed in RON – the Romanian currency) and it stands for the OVM Petrom company. Since the company was listed on September 3, 2001, the data covers the period form that day to May 24, 2011. It is worth mentioning that SNP is one of the ten stocks based on which the BET is computed, with a weight of approximately 19.29%. The evolution of both series is illustrated in Figure 1.
Figure 1. Evolution of the BET index and of the SNP stock closing prices
The figure shows how the prices of the BET index and of the SNP stock have evolved from September 1997 to May 2011 (for SNP the data are from March 2001, when the company was listed on the stock exchange). A general positive trend can be observed till 2007, with a historical maximum in that specific year, as Romania entered the EU. In the following period, the market registered a steep decline as a result of the international financial crisis. The economy started to recover only after February 2009 and kept a general positive trend since then.
ANALYSIS OF THE DATA
Before proceeding to finding an appropriate model that describes each of the two series, these should be checked for stationarity. If a time series contains a unit-root, the coefficients obtained through the model would allow the errors to have an increasing effect on the current variable, which is not consistent with the rational economic theory. A unit-root test...