The paper concentrates on empirical testing of the stability and cross-sectional invariance of the factor patterns underlying the Arbitrage Pricing Models of two neighbouring security markets. In previous studies, the method of transformation analysis has been used to address these issues at the individual asset level. In the present study three alternative transformation analysis models are used to study the stability and invariance problems. The tests are carried out at an aggregated level, such that subsets of asset returns are combined in equally weighted portfolios in the spirit of Fama and MacBeth. Portfolio formation is motivated by the observed anomalies of individual asset return series. Even if some anomalies may be present on the aggregated level also, their impact will be weaker. The amount of different issues of the same company in the database is varied in order to study the impact of parallel issues on the empirical results.