The national accounts of Norway uses a matrix of weights, the consumption matrix, to calculate the development of household consumption groups based off of retail trade sales. However, the observation of both these weights and household consumption values are based on infrequent surveys. With new information on the target, household expenditures, this method allows updating the weights without observing them directly, given the values of retail-trade sales.
This is the indirect method for benchmarking a matrix of weights that links one vector to another through a matrix product. The method makes the smallest necessary changes to the weights in order to produce the target vector with the matrix product of the input vector and the new and endogenous weights. For reasonability, requirements are also put on the weights that they can individually only be between 0 and 1, as well as all the weights for a given industry summing up to 1. This allows for more updated information in intermittent estimates of the target vector.
Indirect Weight Benchmarking
This document describes an indirect method for benchmarking a matrix of weights by adjusting the weights by the smallest change that reproduces newly observed levels when the underlying cells cannot be observed directly.
Documents 2026/26
Published: 3 July 2026
ISBN (electronic):978-82-587-2120-5
ISBN (electronic):978-82-587-2120-5