The present paper discusses some important issues involved in estimation of state income originating in the registered manufacturing sector with illustrations drawn from Gujarat. It highlights the present practices in preparing quick estimates for latest years when results from the regular data source of Annual Survey of Industries is not finalized. It also argues that the current practice of following single deflation method is inadequate and misleading for deriving estimates of value addition at constant prices in this sector. When the industrial structure and technology are rapidly changing with substantial differences in price trends of inputs and outputs, double deflation is the right method to serve the purpose. The paper also draws attention to micro-level inconsistencies and unexplained jumps in annual estimates of input-output ratios obtained from the ASI results. Finally, the paper points out the serious flaw of gathering information in ASI without updating regularly the Census sector frame at the state level. It results in ignoring new large and medium manufacturing units from the ASI results. It is shown for Gujarat state that more than 1000 such units remain uncovered by ASI resulting in the serious underestimation of vale addition of about 23%. This is likely to be a widespread phenomenon not confined only to Gujarat. Urgent action to address these problems will go a long way to improve quality and credibility of the estimates.