In this paper, we use the weak separability criterion to check for the existence of six different monetary aggregates reported by the Center of Financial Stability (CFS). We implement an extended version of the semi-nonparametric tests introduced by Barnett and de Peretti on US monthly data from January 1967 to December 2012. The test, first, checks for the necessary existence conditions of an overall utility function and a monetary subutility function, and then tests for the separability of the latter. On different subsamples, our results suggest that only the DM1 aggregate meets the separability criterion. Implemented on macroeconomic data, we have tested a joint assumption about separability and the existence of a representative agent. Thus, the rejection of the null could also be due to the rejection of stringent Gorman's conditions. More advanced tests for weak separability are clearly required to confirm the results found in this paper.

In aggregation theory, the admissibility condition for clustering components to be aggregated is blockwise weak separability, which also is the condition needed to separate out sectors of the economy. Although weak separability is thereby of central importance in aggregation and index number theory and in econometrics, prior attempts to produce statistical tests of weak separability have performed poorly in Monte Carlo studies. This paper introduces a new class of weak separability tests, which is seminonparametric. Such tests are both based on a necessary and sufficient condition and are fully stochastic, allowing to take into account measurement error. Simulations show that the tests perform well, even for large measurement errors.

The Generalized Axiom of Revealed Preference (GARP) can be violated because of random measurement errors in the observed quantity data. We study two tests proposed by Varian (1985) and de Peretti (2004), which test GARP within an explicit stochastic framework. Both tests compute adjusted quantity data that are compliant with GARP. We compare and contrast the two tests in theoretical terms and in an empirical application. The empirical application is based on testing a large group of monetary assets for the United States over multiple sample periods spanning 1960–1992. We found that both tests provided reasonable results and were largely consistent with each other.

This paper introduces a general procedure that tests the significance of the departures from utility maximization, departures defined as violations of the general axiom of revealed preference (GARP). This general procedure is based on (i) an adjustment procedure that computes the minimal perturbation in order to satisfy GARP by using the information content in the transitive closure matrix and (ii) a test procedure that checks the significance of the necessary adjustment. This procedure can be easily implemented and programmed, and we run Monte Carlo simulations to show that it is quite powerful.