At the 82nd CNB Economic Workshop, Ivana Vidaković of the General Economic Statistics and Statistical Information Systems Department presented her paper "Mind the Gap: Who Holds the Missing Deposits".

Household surveys are a key source of information for analysing the distribution of wealth, but their results often differ from the values recorded in sectoral balance sheet statistics. Such discrepancies may result from non-response, under-reporting of asset values and insufficient representation of households at the upper end of the wealth distribution, making it difficult to reconcile survey estimates with macroeconomic aggregates. This problem is particularly pronounced for household deposits, where discrepancies between survey estimates and sectoral aggregates are often substantial. Using Croatia as a case study, the paper develops an approach that addresses two potential sources of bias: the under-reporting of deposits within the survey sample and the under-representation of households with high deposit holdings.

The paper builds on the existing framework of distributional wealth accounts and develops a complementary two-stage procedure for adjusting household deposit estimates, combining data from the Household Finance and Consumption Survey with bank data. In the first stage, reported deposit values are adjusted for households whose income, assets and employment status suggest potential under-reporting. In the second stage, the remaining discrepancy is addressed through an adjustment of the upper tail of the distribution, combining moderate reweighting of observed households with a limited number of synthetic households.

The paper's main contribution lies in the reconstruction of the upper tail of the household deposit distribution. Alongside survey data, the approach uses aggregate deposit benchmarks and bank-based concentration indicators measured at the level of individual depositors. This makes it possible to improve survey-based estimates even without directly linking survey households to individual bank depositors. The resulting estimates are reconciled with the aggregate amount of household deposits and brought closer to bank-based concentration indicators, while preserving the household characteristics observed in the survey. More broadly, the paper illustrates how administrative data specific to individual components of wealth can help improve estimates of their distribution.