This study tries to find out the determinants of the aggregate price and fit a parsimonious statistical model for aggregate price level in India. Two measures, namely, CPI and WPI are taken for aggregate price level. Several statistical models has been used including cointegration method. When regressions are performed with due care to non-stationarity, regression results are very poor. But this does not indicate that the variables are not related, but there is much stronger evidence on their relationship when cointegration techniques are used. This result is robust to both the measures of aggregate price levels and to all five different specifications of cointegration test. The statistical a-theoretic ARIMA models for aggregate price level are good enough. These two helps us to forecast future values of aggregate price level.