"In the second place, a system would involve more variables than those which may be supposed to enter into demand relationships, and so the preliminary work of estimating the necessary time series would also be increased. In the third place, in the simultaneous method of estimation the estimates of the original parameters come to depend on most of the remaining estimates, and so it is quite likely that they will have large sampling variances."

“A choice may have to be made between a biased estimator which is believed to be relatively accurate and an unbiased estimator which is demonstrably highly inefficient, and it is by no means obvious that the latter is always to be preferred.”

“I have tried to make clear in this brief sketch of the nature of deductively formulated theories, at least as they occur in economics, that, although their origin lies in common observation and introspection, nevertheless they are capable of a purely mathematical development through the exact statement of a of a set of postulates which, however, since they must inevitably contain assumptions about human behaviour, require to be tested by reference to actual events.”

“In 1945, the war ended and I was chosen to be the first director of the newly-established Department of Applied Economics in Cambridge. Between leaving the government service and taking up my new post, I had a break of about three months which I spent at the Institute of Advanced Study in Princeton. I intended to use my time there writing up my ideas on a social accounting system for the measurement of economic flows, a thing I had wanted to do for years but had not had time for during the war. What happened was that, in Princeton, I met, the Director of Intelligence at the League of Nations, who wanted a paper on the problems of defining and measuring the national income and related totals for consideration by the League's Committee of Statistical Experts. He asked me if I would undertake the work and naturally I accepted. I soon had a memorandum ready and it was discussed in Princeton while I was still there by a subcommittee convened by Loveday. Their report was eventually published by the United Nations in Geneva in 1947 under the title, Measurement of National Income and the Construction of Social Accounts, with my memorandum as an appendix.”

“The usefulness of observation and measurement in testing economic theories arises because the theorems of economics are supposed to relate to the actual world... Any economic theorem rigorously deduced from given postulates may be regarded as a hypothesis about the actual world which xperience may show to be false.”

“There is no hope, as far as I can see, of pushing the postulates of economic theory back to a set of irreducible axioms.”

“We usually find a considerable degree of correlation between indicators and also between different sets of weights that we should ever think of using. Consequently we are likely to find that alternative index numbers using different weights show similar and often approximately proportionate movements. That this is so in most economic applications is due to the nature of the actual world and not simply to the formal characteristics of index number techniques.”