Dama International Journal of Researchers (DIJR), ISSN: 2343-6743, ISI Impact Factor: 0.878 Vol 1, Issue 6, June 2016, Page 68-75, Available @ www.damaacademia.com Dama International Journal of Researchers, www.damaacademia.com, editor@damaacademia.com 67 Time Series Analysis of Road Accidents in Ghana Salifu Nanga, (M.Phil. MSc. BSc.) Department of Statistics, School of Graduate Studies, University of Ghana, Abstract The mortality and morbidity trends from road accidents are rising annually in almost all the developed countries. The total number of deaths and injuries continues to rise despite the decrease in road casualties in relation to the total number of vehicles. Road traffic accident in Ghana is also increasing at an alarming rate and has raised major concerns. In this study, univariate time series was used to model road accidents. The Box – Jenkins method was applied for a 20 year period from 1991-2010.A model was subsequently developed to fit the time series data. SARIMA (1,1,0) × (0,1,1)12 was found to be the best model for road accident cases with a maximum log likelihood value of 245.48, and least AIC value of 5.892, RMSE value of 17.930 and MAPE value value of 1.688. An ARCH-LM test and Ljung-Box test on the residuals of the models revealed that they are free from heteroscedasticity and serial correlation respectively. Keywords: ARIMA model, SARIMA model, Bank of Ghana, Box Jenkins, Forecasting, Road Accidents I. INTRODUCTION Road accident is the eleventh most common cause of death in the world and accounts for 2.1% of all deaths globally (World Report on Road Traffic Injury Prevention, 2004).Ghana loses 1.6 % of Gross Domestic Product (GDP) through road accidents (NRSC, 2007). Between the short period from January to August 2011, 1,431 people have died as a result of road accidents (Ghana web, 2011). This is a very alarming phenomenon which has to be tackled tactfully. Some of the socio-economic effects include disability and therefore a high dependency burden, and for some victims the seriousness of their disability could result in being jobless. With men representing 70% of national casualties, it has serious implications. Monthly data on road accidents were collected in Ogun State. The Box-Jenkins approach of model identification, parameter estimation and diagnostic check was adopted. The result revealed that the monthly road accidents is basically an ARIMA (2,1,1) model, ARIMA (2,1,0) model, ARIMA (2,1,1) model (Olobatuyi ,2012). A study carried out in Lagos state on road traffic accidents for the period 2005 – 2010 revealed that the ARIMA model of order (2, 1, 0) was appropriate (Fadugba ,2012). A study also carried out in Egypt revealed that the monthly series of the number of road traffic accidents from 1990 to 2008 was used, where results showed that the adequate model is SARIMA (1,1,1)*(0,0,1)12 (Abass, 2004). Another study by CAPMAS for monthly data on road accidents Egypt revealed that SARIMA(1,1,1)*(0,1,1)12 was appropriate for road accidents, SARIMA (0,1,1)*(0,0,2)12 for fatalities and SARIMA (1,1,1)*(1,1,1)12 (Ho et al , 2002). To analyze annual road accident data between 1980 and 2010, a Box-Jenkins approach was used to determine the patterns of road traffic accident cases, injuries and deaths along the Accra-Tema motorway (Okutu, 2011). ARIMA (1, 1, 1) was used to model injury and death as a result of accidents while ARIMA (0, 1, 2) was used to model accident cases. Forecasts revealed that there will be a steady increase in road accidents in the next five years. A statistical analysis of the systematic yearly increase in the number of accidents was carried out by Boakye et al in 2013. Data was collected on yearly road traffic accidents and population values of Ghana for the