Crop output estimates are always fraught with uncertainty, as production happens not in factories, but over millions of hectares of open fields exposed to the elements. Their numbers are, therefore, subject to margins of error and upward\/downward revisions much more than in steel or cement. But when the gap between initial and final estimates or that of the government vis-\u00e0-vis private trade are of the order of 10 million tonnes (mt), there\u2019s a serious problem. The Union agriculture ministry has pegged this year\u2019s wheat crop at 94 mt-plus, a jump of 7.5 mt over 2014-15. This, when both market arrivals and government procurement have fallen by about 5.1 mt, even as wholesale wheat prices are ruling roughly 15 per cent higher than last year. The fact that even for 2014-15, the ministry\u2019s final figure of wheat production turned out be about 9.2 mt lower than the first advance estimate of 95.76 mt, raises further doubts about the quality of official data. And that isn\u2019t restricted to wheat. There are two broad concerns. Does Krishi Bhawan have a robust mechanism to track the progress of sowing in various states, the status of crops at different stages of growth, arrivals in mandis and stocks with the trade, apart from imports, exports and prices both in domestic and international markets? The importance of such information is obvious for policy purposes, whether it relates to fixing of minimum support prices for promoting\/dissuading cultivation of particular crops or supply-side management with a view to control inflation. Recent experience \u2014 at least with regard to onions, pulses, sugar and now wheat \u2014 points to the government not being particularly possessed of timely market intelligence. Being slow in reacting to both shortages and gluts has, in turn, produced knee-jerk responses \u2014 typically in the form of imposition of export curbs or stock-holding limits without a thought to what these would do to producer incentives. The second concern is larger, having to do with credibility. Liberalised market economies don\u2019t need the government to directly engage in production and trade, but its role in statistical data collection and dissemination \u2014 which is a public good much like primary education, healthcare or law and order \u2014 is well recognised. Trust in government economic data is premised on its not being in business and, hence, unlike private industry, not financially standing to lose from giving out the right information. It is this credibility that becomes a casualty when official information on farm production or GDP growth is seen to be providing an exaggerated picture of buoyancy, which does not appear to square up with the ground reality. The agriculture ministry should explain the basis for its optimistic output estimates.