The RILE index of the party manifesto dataset is the most used, but also the most controversial estimate of parties' left-right positions. Both its validity and reliability have been thoroughly criticised, yet a basic question has not been addressed. The current paper will argue that for the index to be valid, certain patterns of association must be present in the data. It will apply canonical correlation analysis (CCA) to test these associations within and between the sets of variables that are used to calculate the left-right (RILE) index. The analysis will conclude that for countries that have not experienced a communist/socialist past, the relationship is present, although substantively very weak. More important, however, is the result that for post-communist countries, the required associations in the data are clearly not there and thus it must be concluded that the RILE index is an invalid measure of left-right position in this set of countries.