For an argument about why it is legitimate to argue in terms of aggregates, and why it may sometimes be inappropriate to try to reduce everything to individual behaviour (even if we could in principle do so), see the relevant chapter of Kevin Hoover's excellent book, The Methodology of Empirical Macroeconomics (Cambridge University Press 2001). This may not amount to a disagreement with Anthony Waterman's carefully worded defence of methodological individualism, but it seems an important point to add. Roger Backhouse