PROFESSORS Friedman and Meiselman' recently have reported that a simple theory model describes aggregate consumption more accurately than a simple autonomous expenditure model. They believe this result is evidence that the quantity theory is a better description of the American economy than the autonomous expenditure or Keynesian theory.2 If their interpretation were correct, the Friedman-Meiselman paper would be one of the most significant economic studies in many years. But it is not correct. Friedman and Meiselman have represented the autonomous expenditure theory in a very unorthodox form. Their statistical comparisons are extremely sensitive to how the autonomous expenditure theory is represented. Below, I employ a more conventional representation of the autonomous expenditure theory and demonstrate why Friedman and Meiselman's tests are misleading. Further, using this conventional model and some of their data, little empirical evidence is found which favors the theory. Finally some other conceptual weaknesses of the Friedman-Meiselman tests are illustrated. Briefly, Friedman and Meiselman compare simple, partial, and multiple correlation coefficients obtained from the following equations, estimated from annual (1897-1958) and quarterly (1945-1958) data for the United States: C=al+8(A (1) C=a2 +82M (2) C = a3+/33A +13P (3) C = a4 +84M+y4P (4) C = a5 + 35A + 85M (5) C = a6 + 86A + 86M + Y6P (6)

UMEROUS theories of commercial bank N behavior have been proposed. In all of them, some form of profit maximization has been posited, either explicitly or implicitly, as the motivating force. Therefore, the rates of return, positive and negative, which a bank realizes from its assets and liabilities are important determinants of a bank's portfolio composition. Conversely, the composition of a bank's portfolio is an important determinant of its profits.' The relevant rates of return on earning assets, of course, are not easily observed nominal rates; servicing and processing costs must be deducted. Similarly, the relevant rates for liabilities are not observable interest payments per dollar; servicing costs net of service charges to depositors must be added. The relevant rates of return are net rates, and to apply a theory of bank behavior it is necessary to have estimates of these net rates. The purpose of this paper is to provide empirical estimates of the net rates of return which banks realize on various elements of their portfolios. Regression methods are utilized to allocate revenue and cost among the elements of bank portfolios.2 Thus, given observations of a cross section of banks. least-squares regressions of net current operating income (and other variants of profit) on various assets and liabilities are computed. The coefficients are estimates of net rates of return. The first section of the paper explains the analytical framework underlying the study and describes the data. The second and third sections report applications of the model to different samples of banks. By far the richest set of data concerns member banks in the Tenth Federal Reserve District, which is analyzed in section II. In section III, data for Connecticut commercial banks are studied. The fourth section of the paper briefly compares the estimates of net rates of return at Tenth District and Connecticut banks.