All over the world, the standard practice in economic data gathering allows for considerable revisions from time to time. For example, preliminary estimates of GDP are produced periodically (either quarterly or annually), and are then updated regularly. Typically, it takes about four years to arrive at the final data. One reason for this lag is that initial estimates are based on projections from previous data plus few data available for the current period, thus excluding relevant inputs. Another reason is the need to periodically update the underlining data generating methodology.

In the case of GDP, this takes place typically at five-year intervals, and is driven primarily by structural shifts in the economy over time. These shifts could affect the base year, the weight on output and prices, and the definition of variables, which often create divergences between changes in the estimates and actual structural changes in the economy. Essentially, the process of moving from preliminary estimates to final data yields better understanding and more accurate description of the economy.

However, it becomes worrisome when the divergence between the final data and the preliminary estimates becomes so considerable to the extent that they paint radically different pictures of the economy. This is where the events of the past six months in Nigeria come to fore. The first is the GDP rebasing exercise. After about 19 years of delay, the new estimates show that the value of economic activities had been underestimated by 89 percent, while the economy had transformed from agrarian-based to service-based. The second is the revised poverty estimates for the country released by the World Bank, which reveals that the poverty rate had been overestimated. The new estimate puts the poverty rate in 2010 at 35.2 per cent against the initial estimate of 62.6 per cent.

While these estimates places the economy in a better position than initially thought, enormous discrepancy between the estimate and the final data poses concern. To this end, government policies are formulated as well as evaluated on the basis of available data as policymakers are likely to adopt different measures based on the different estimates. For example, based on the initial estimate, poverty rate increased by 11 percentage points between 2004 and 2010 despite the impressive 6.6 percent economic growth rate over the same period, suggesting that growth has not been pro-poor, and necessitating policy revision in order to position the economy on a more inclusive growth path. In contrast, new data shows that poverty actually reduced by 16.4 percentage points, implying that growth was pro-poor, on account of which current policies should have been sustained. Evidently, data inconsistencies could lead to undesirable policy reversal or, simply put, policy somersault.

The quality of data is undoubtedly crucial for economic development and policymaking. There is an urgent need to improve ‘knowledge production’ in Nigeria as pre-condition to improving the efficacy of public policy. This will require developing human and technical capacity across the data generation value-chain. Presently, the country relies heavily on multinational agencies for various statistics. While this is not bad in itself, reliance on external capacity might reduce the incentive to develop local capacity in these areas. Also, there is a need for more investment in data production. One of the major challenges to quality research in Africa is the absence of data, which often leads to redirection of research from the problems facing the economy to the problems that are researchable, given available data. As the saying goes “a stitch in time saves nine”, although in this case the whole nation could be saved from the deleterious impact of implementing policies based on error-prone data.