On Forecasting Rate of Decline in Infant Mortality in South Asia Using Random Walk Approximation

United Nations Millennium Development Goals(MDGs) has set the target of reducing high rates of infant (and child) mortality (IMR) by two thirds to be reached by 2015 using 1990 as the benchmark year. The availability of time series data on Infant Mortality rate from Inter- Agency Group for Child Mortality Estimation(IGME) led by UNICEF and WHO, it has become possible to track the rate of progress towards this goal. Using the IGME 2012 data for all the South Asian Countries, I have considered three specific issues in this article. (1). How does the South Asian Countries fair in reducing the IMR towards this MDG target? Although the time series data exhibit declining trends for all the countries in South Asia, to what extent such trends are attributed by their average annual progress trajectory over the period for which data are available? (2). Whether deterministic or stochastic trend can attribute the IMR decline in South Asian countries and what alternative time series models be used to forecast the decline in Infant Mortality? Can we find a serviceable representative model for the entire region? (3) In case, a satisfactory representative model for the entire region exists, how do we assess the forecast accuracy using the model and quantify the propagation of forecast error?