Opinion: Estonia's consumer confidence index has lost its forecasting power

Economist Indrek Saul argues that Estonia's consumer confidence index, published monthly by Statistics Estonia, has lost its ability to predict GDP. He calls on the agency to either rebuild the index using only its most predictive components or publish a validity indicator alongside it.

Every month, Statistics Estonia releases a consumer confidence index that is widely cited in media and policy circles as a barometer of economic health. But according to economist Indrek Saul, the number may be close to meaningless — at least as a forecasting tool.

Saul argues that the index, in its current form, no longer reliably predicts gross domestic product. This is not a minor technical quibble. If a headline economic indicator fails to track the thing it is supposed to measure, publishing it without caveat risks misleading businesses, policymakers, and the general public.

What should change

The solution, in Saul's view, is straightforward: Statistics Estonia should either reconstruct the index using only those sub-components that have demonstrated predictive power over recent periods, or it should publish a transparency metric alongside the index — one that shows readers how well the indicator has actually performed against GDP in recent quarters.

Without such a reform, Statistics Estonia is effectively sending out a monthly number whose forecasting value is unknown to anyone, including the agency itself. Saul's criticism points to a broader problem in official statistics: indicators developed decades ago are rarely reassessed for continued relevance, even as economic structures and household behaviour evolve.

A call for transparency

The argument is not that consumer sentiment surveys are useless in principle. Confidence data can still capture mood and behavioural intentions. But if the link to real economic output has weakened, that fact should be communicated explicitly. Presenting an index without its track record is, at best, uninformative — and at worst, actively misleading.