John and Jim Thanks for your necessary correctives to my outburst. I accept all you say. Certainly, I have no difference with the insight that the profit acquired by a capitalist as a result of owning any asset whether or not it has intrinsic value and whether or not it is productively deployed can be decomposed into two components: (i) the new value which, as a result of the action of human labour, is added to social capital and thereafter added or transferred to the individual capital concerned (profit or return on capital) (ii) the previously existing value which is transferred between capitals as a result of the action of the price mechanism in combination with the effects of technical change. (depreciation, capital gains, windfall profits and losses) I also take John's point about capital-output ratio. I think that the issue is commonly conceived in the following terms: since the physical output per 'unit' of capital is in general increasing as a result of technical change, it appears completely paradoxical that the (net) value output of capital per unit of value of capital should in general decrease as a result of technical change. It is an aspect of commodity fetishism that this 'surface' appearance is so widely taken for truth. Okishio's theorem for me is the mathematically pure expression of this fetishistic view. In this sense, Marx's analysis is a genuinely scientific insight inasmuch as it explains the observed fact that the (new/net) value output per unit of value of capital *does* decrease even when there are favourable changes in the physical proportions concerned. However, this decrease happens under *all* circumstances except disinvestment in value terms. Sometimes, therefore, the rate of profit falls with a rising capital-output ratio, and sometimes with a falling capital-output ratio. I hope this point is not lost sight of. Can I suggest the following formulation: what has been established is that the rate of profit falls *independently* of the capital- output ratio? Previous views suggest a connection; there is no connection - precisely because, under circumstances of rapidly declining asset values due to technical improvements, the lags between asset prices at current and historic cost overwhelms the productivity gains when the rate of profit is arrived at. Indeed, the very process of technical innovation which lowers the current value of all assets, is the same process which produces the capital loss effect which negates the productivity gains when expressed as a money rate of profit. This does not at all rule out the possibility that in actual fact, output rises in use-value terms in proportion to capital stock [though I am not quite sure how this could be measured because of the aggregation problem] while in value terms, it falls. Therefore John's point is entirely well-taken. Actually, with regard to the empirical discussion, surely the question is this: on the basis of Okishio's theorem, it is quite remarkable that the rate of profit ever falls at all, never mind if this fall is persistent. As far as I can see within Okishio's framework the only possible reason for a fall in the rate of profit, given cost-reducing technical change, can be a rise in the real wage. Therefore, to refute the theorem empirically, one has only to show periods in which the real wage falls or stays constant, while the rate of profit falls. This is much easier than showing that the rate of profit falls throughout history, which frankly I doubt. There is a lot of empirical work out there which suggests that on this basis the Okishio theorem is empirically invalid. Can I respond with a further corrective: it isn't my intention to bury Okishio or indeed anyone. Okishio's record as a person is that of a dedicated and active supporter of the working class. He, I am informed, at the time viewed the policy implications of his theorem thus: since the rate of profit will rise with technical change, there is no reason to hold back the wage struggle simply to keep profits up. On the contrary, if productivity fails to rise in line with real wages (a neat reversal of the usual way this is put) then this must be the fault of the capitalists for not investing in new technology. The theorem thus stands as a justification for trade union struggle and a refutation of the normal capitalist view that wage rises must be stopped if profitability is to be sustained. Now, this argument makes a lot more sense in Japan than it does in the UK or the US. In the Anglo-Saxon world it was used in the reverse sense, to establish that the cause of the profit squeeze was greedy workers. I think this tends to give us Anglos a certain edge in our prejudice against the theorem. Actually, I think the theorem is a mathematical marvel. I think it is one of the most sublime theorems of the age. The question to ask is therefore: what faith should we place in mathematics per se, when such marvellous things are demonstrably false? This God may be beautiful but is She good? Kali, perhaps, has two faces. Alan