The Reconciliation of Inconsistent Economic Data: the Information Gain

This paper demonstrates that the posterior estimates given by two common formulations of ‘data reconciliation’ problems result in better quality data. ‘Better quality data’ is interpreted in terms of the closeness of the posterior estimates to the true but unknown values. It is also shown that more or improved information further improves the quality of the posterior estimates. These results provide a deterministic rationale for the reconciliation of inconsistent economic data.