Colombia’s official statistics: Autonomy and accountability at risk?

Colombia’s national bureau of statistics (DANE) has a new director. Besides the challenges natural to directing any statistical agency, Héctor Maldonado will have to face unusually difficult times. Both DANE’s independence and the quality of some of the statistics it produces have been recently put into question.

The last two directors of DANE have resigned reporting inappropriate pressures from the government. These allegations have left the public wondering whether the government has been trying to shape the numbers produced by DANE to improve public assessments of government’s performance.

To add to the turmoil, some of the data produced by DANE’s last administration was subject to strong criticism. The hastiness with which the 2005 Census was designed generated serious concerns among experts. Some surprising results, such as the report that the most remote areas of the country have more PhD graduates than Bogota, added to those concerns. The employment figures have also been strongly debated. The methodology was changed without overlap period, making trends impossible to evaluate; the high reports of unemployment seemed implausible in the context of strong economic growth. In fact, the bitter debate between the former DANE director and government officials was partly stirred by a government’s request to access micro-data that would allow double-checking employment statistics.

The debate may remind many of the well-known trade off between autonomy and accountability. But this is one case where accountability, not only to the government but to society as a whole, would probably help protect autonomy.

DANE’s tradition of independence in the production of economic statistics stems from a history of great respect from academics and the public in general, perhaps even more than from formal arrangements. Perceived attempts of interference by other government agencies would, in normal times, have stirred an outcry from experts. But the recent government’s request to release micro data has only echoed what independent experts had been demanding for months, after DANE added to the change in methodology a decision to stop releasing some information it traditionally made available (monthly figures, absolute levels of unemployment and employment, and micro data from the households surveys). Maintaining openness about the production of these statistics is fundamental to guaranteeing continued support from experts. This support is in turn an important force for countervailing potential government pressures. Of course, a high degree of openness cannot come at the expense of violating the confidentiality of the micro data, without which individuals and firms would have incentives to misreport their behavior. But the traditional arrangements for the release of the data did not attempt against confidentiality.

Formal arrangements could also be shaped to guarantee a greater degree of autonomy. There have been proposals to place DANE’s direction in the hands of a committee rather than a single director, to give it financial and administrative independence, to dictate a formal four-year period for the director. However helpful these changes could be, nothing will replace a strong support from the public and experts. In turn, support can only be garnered if DANE renews and maintains its vows of transparency. This is certainly one of the great challenges Héctor Maldonado will face.

4 Responses to "Colombia’s official statistics: Autonomy and accountability at risk?"

  1. Vitoria Saddi   October 23, 2007 at 8:42 am

    Marcela, Is DANE facing a political battle similar to the INDEC in Argentina? What are the benefits that politicians get from reporting incorrect data?Thanks, Vitoria

  2. Anonymous   October 25, 2007 at 8:37 am

    Colombia = Argentina when it comes to statistics reliability? That is sad to hear!

  3. Guest   October 25, 2007 at 8:39 am

    Lets hope the new director of DANE is independent and honest !

  4. Marcela Eslava   October 29, 2007 at 10:29 am

    Interesting comments about apparent closeness to Argentina’s episode of confrontation between INDEC and the government. The two episodes are different in nature, but in terms of public perception they unfortunately boil down to the same: lack of trust in offical statistics, and the impression that they are contaminated by the government’s wishes. The users of official statistics fear, in both cases, that the government may be trying to manipulate the figures to show a good economic performance. What are the differences? In the Argentina case, it has been suggested that the government directly tried to manipulate the numbers, by removing an official in charge of inflation statistics when those statistics showed a growing inflation trend. My understanding is that in that case the whole academic community repudiated the government’s decision. In Colombia, meanwhile, the frictions have been more related to the government expressing concern or disagreement about methodologies. In some cases the academic community has validated those concerns (the case of the household surveys), while in others it has opposed them (the government’s pressure to include in the census questions that allowed pinpointing specific individuals for what was perceived as purposes relating national security).