The case for state intervention does not immediately follow from the mere existence of these limitations of the market. One can for instance think of co-operative arrangements within civil society itself to overcome cases of isolation-related market failure without necessarily having recourse to state intervention.

This entire "close universe" approach sees state intervention essentially as a means of supplementing the functioning of the market rather than of suppressing it. Two perceptions follow first, one can have a stable combination of a market economy with state intervention, and secondly, one can have in principle any combination of a market economy with state intervention. Or putting it differently, the fact of state intervention creates no tensions in a market economy, a proposition that clearly underlies Keynes' writing.

As mentioned earlier, however, this approach, and the perception of the state-market relationship that follows from it, is based on the assumption that the territorial jurisdiction of the state is co-terminus with the geographical expanse of the market. But this is never the case. Capitalism from its very beginning has been an international system, and the "market" usually refers to the international market and the "state" in question is basically a nation-state.

The fact that Keynesian demand management by some world state is not possible would not matter if each nation-state intervened in demand management in its own economy and if there existed some mechanism whereby the current account deficits on the balance of payments of the deficit countries could be automatically financed by the surplus countries (since the aggregates of deficits and surpluses must exactly match, this should be theoretically possible). But this is not the case; surplus countries are not obliged to lend to deficit countries, and hence demand their "pound of flesh" as a condition for lending. Likewise, the fact that re-distributive measures are not undertaken at the world level would not matter if labour was freely mobile across to intervene to ensure that the position of the population living within the territory under their jurisdiction is improved relative to the equilibrium that the unfettered operation of the world market would have achieved for them.

This need is further strengthened because of the movements of capital. With restrictions on international labour mobility, where exactly productive capital flows, i.e. which region experiences a faster rate of growth of productive capital stock, becomes a crucial determinant of prosperity over time. And hence the need arises for the nation-state to intervene to improve the position of its population. What is more, the movements of short-term speculative capital that have autonomy of there own, the movements of crucial bearing on the single-period equilibrium within any country. This equilibrium in other words loses any claim to having anything sacrosanct about it; it emerges merely as the outcome of the caprices of speculators. Hence again the need arises for state intervention.

A world of autonomous financial flows is much more prone to social instability: when finance flows in it creates social instability owing to increased unemployment, but when finance flows out it creates social instability owing to "unacceptable" rates of inflation. The basic proposition about an economy open to such inflows being in a worse state than one not open to such inflows still remains valid.

Against this conclusion it may be argued that keeping the exchange rate fixed and letting the effect of financial flows fall on the level of foreign currency reserves would get rid of the problem. But any expansion of reserves in a period of large financial inflows would improve the state of credit in the economy and increase domestic expenditure and absorption, usually through larger consumption, which would deplete these enlarged reserves. In a period of large financial outflow therefore the level of reserves would be unlikely to meet the outflow, necessitating either a reduction in the level of activity or currency depreciation with "unacceptable" inflationary consequences. True, in such a case, society as a whole would appear to have splurged in one period to be squeezed in the next; but the beneficiaries of the splurge and the victims of the squeeze would by no means be identical, so that the social consequences of the two would be asymmetrical.

Likewise the policy of stabilizing the level of activity and the exchange rate by expanding the fiscal deficit in a period of large financial inflow and contracting it in a period of large financial outflow would not work, not only because of "ratchet effects" "distribution-related" and "isolation-related" cases. This "territory-related" case is of paramount importance in the context of developing economics and I venture to add transitional economics.