i. in his recent book Keynesian Revolution, a significant contribution to current literature, Lawrence Klein says (p. 91 ) that there are two Keyneses in the matter of the savingsinvestment equation. This statement has frequently been made before and while much has been written about it, perhaps something useful can still be added. In Klein's exposition, savings and investment as observables are always equal, being the point of intersection of the schedules of savings and investment. There is, of course, as Klein points out, no inconsistency between (i) the definition which makes actual or observable savings and investment always equal, and (2) the concept of savings and investment as intersecting schedules. Nevertheless the question remains whether these schedules are to be regarded as (a) very short run schedules which might at times shift about violently and capriciously in response to temporary conditions, or (b) longerrun schedules determined by fairly stable factors involving the behavior pattern of a community with respect to the ratio of savings to income at different income levels. 2. A related, but distinct, terminological matter should be mentioned. While the Keynesian theory of income determination can indeed be formulated in terms of savings and investment schedules, I am increasingly convinced that the presentation as given in the General Theory is more useful. In the General Theory, Keynes, in fact, did not explicitly work with savings and investment schedules,' even though it is perfectly true that his analysis can be put in these terms. General Theory, however, makes (i) actual investment and (2) the consumption function, the determinants of income. The propensity to consume and the rate of new investment determine between them the volume of employment. . (General Theory

this brief comment I am going to limit myself to what I regard as essentials. I am not going to discuss details, important enough in themselves, but which I think can, among reasonable men, be settled in a fairly satisfactory manner. I do not think it is worth while to waste a lot of time on whether a high level of is a better term than employment. Both phrases need definition. Everyone is agreed that in a dynamic market economy there will be seasonal, frictional, and transitional unemployment associated with changes in the seasons, the introduction of new products and of new methods of production. Moreover, in a free society where wage earners work whom they please, there will of necessity be a degree of labor turnover. Full employment in the United States, in my judgment, means perhaps 4 or 5 per cent unemployed at any one time; assuming a labor force of around 6o million, this would mean unemployment of 2 /2 to 3 million. If on the average 5 weeks should elapse before a new job was found, this would mean that 25 to 30 million people would shift jobs in each 12-month period. Thus an average of 4 to 5 per cent unemployed provides enormous flexibility in the labor market. Important as these matters of detail and definition are, I turn now to what I regard as more fundamental considerations. What is really important is that the Murray Full Employment Bill, if enacted into law, would, in common with the British and Canadian state papers on Employment Policy, represent a new attitude, purpose, and responsibility of the central government with respect to the problem of unemployment. Instead of palliative and ameliorating measures, these state papers announce a positive national policy with respect to the maintenance of employment, production, and national income. The British and Canadian state papers recognize that these are novel experiments. This involves a new approach and a new responsibility the State, says the British White Paper, and adds, In these matters we shall be pioneers. Similarly the Canadian state paper states, We must determine therefore to learn from experience, to invent and improve the instruments of our new policy as we move forward to its goal.... The Government is inaugurating policies which break new ground and is confident that these policies, with full public understanding and support, will achieve . . . satisfactory results of decisive importance. later years as experience grows they can be made to yield ever improved results, which will mark a new era in Canadian development. Apart from announcing a new goal of national policy and a new responsibility of government, these documents are of the utmost significance in that each commits the government in question to a periodical and continuous assessment of the employment situation. The Murray Bill makes it the duty of the President to transmit to Congress at the beginning of each regular session (and thereafter supplemental reports from time to time) a National Production and Employment Budget setting forth the estimated trends and prospective developments with respect to the size of the labor force, the gross national product, national income, private consumption expenditures, private investment expenditures, state and local outlays, and alternative ranges of federal expenditures. The British White Paper stresses the importance of establishing a central staff qualified to measure and analyze economic trends, and it lists the principal classes of statistics which must be obtained for the efficient operation of an employment policy. These procedures assure that the government will continually take the pulse and temperature, so to speak, of the economy in order to measure how well its policies are succeeding in achieving one of its primary responsibilities. Thus the

A “grid” bootstrap method is proposed for confidence-interval construction, which has improved performance over conventional bootstrap methods when the sampling distribution depends upon the parameter of interest. The basic idea is to calculate the bootstrap distribution over a grid of values of the parameter of interest and form the confidence interval by the no-rejection principle. Our primary motivation is given by autoregressive models, where it is known that conventional bootstrap methods fail to provide correct first-order asymptotic coverage when an autoregressive root is close to unity. In contrast, the grid bootstrap is first-order correct globally in the parameter space. Simulation results verify these insights, suggesting that the grid bootstrap provides an important improvement over conventional methods. Gauss code that calculates the grid bootstrap intervals-and replicates the empirical work reported in this paper'is available from the author's Web page at www.ssc.wisc.edu˜bhansen

to study policies affecting money and credit separately from fiscal policy, because the underlying cause-effect relations are so very different. In terms of a summary characterization of the new budget, what I have tried to convey is this: Let us not say that it is designed to show the impact of economic policies; let us rather say more modestly that it shows the impact of fiscal policies; and let us make clear that a detailed and comprehensive analysis of monetary and credit policies cannot be conducted within the framework of the NIP accounts.