ONE important question arises out of current attempts to provide a foundation for aggregate wage rigidity and unemployment. How mobile are unemployed workers? Recent theoretical models of unemployment can be divided into two basic categories: the new microeconomic search theories attribute unemployment to the job search and job changing behavior of workers who become permanently separated from their jobs. The newer contract theories attribute unemployment to the periodic employment reductions (via temporary layoffs) that are necessary to accommodate demand fluctuations when workers remain indefinitely attached to specific firms. The relative importance of these two approaches in explaining cyclical unemployment hinges on the share of unemployment over the business cycles that is associated with real labor turnover. Unfortunately no data are currently collected on the fraction of unemployment with or without job change. This paper derives estimates of temporary layoff unemployment for U.S. manufacturing from Bureau of Labor Statistics (BLS) establishment turnover data. After a brief discussion of current unemployment data, a time dependent distributed lag model of manufacturing rehires is developed. The model allows estimation of both the percentage of each month's layoffs that end in rehire and the average duration of unemployment before rehire. Together with the layoff rate, these two statistics determine an estimate of manufacturing unemployment without job change.

T HE presence of adjustment costs for a firm altering its labour force or capital stock creates considerable problems for econometric work involving firm behaviour, either on the employment or the investment side. The usual way of handling these costs is to derive a so-called desired level for the choice variables based on an optimization assuming there are no adjustment costs and then to add an ad hoc lagged adjustment process, the nature of which however is not derived explicitly from the actual costs of adjustment. There are two problems with this. First, as noted by Nerlove (1972), the resulting paths would not in general be optimal for a profitmaximizing firm so there is an inconsistency between equilibrium behaviour and adjustment behaviour. Second, by allowing the adjustment process to be data determined without imposing the restrictions implied by optimization over adjustment costs, one makes it harder to discriminate between models on the basis of the data. This makes it difficult to test the appropriateness of assumptions about the underlying economic model (for example, the form of the production function used). The problem with including the adjustment costs in the optimization is of course that it leads to formulations that are intractable for empirical work. This paper is an attempt to avoid that problem by looking more carefully at labour costs and to see what can be said about the underlying production model without using ad hoc dynamic assumptions. In particular it recognizes the fact that, even for small firms, the wage rate for operatives is a step function in hours worked; up to a certain number of normal hours an operative is paid at one hourly wage rate but above that number of hours is paid at a higher, overtime rate. Of course, a firm would only use labour at the overtime rate if it cost less overall than hiring more labour at the normal rate. But since firms do regularly employ overtime labour, this suggests that they must face either a rising supply curve for labour or else a fixed cost in hiring additional employees; Since the latter enables one to retain price-taking behaviour in the labour market and can also provide an explanation for short-time working, that is what will be investigated here. What is relevant to the production decisions of a firm is of course the marginal cost of an additional hour of labour, provided, that is, that the average cost is not so high that it is unprofitable to produce at all. For a firm that has employees on short-time working, the marginal cost is the normal hourly wage rate. For a firm with employees on overtime, the marginal cost is the hourly overtime rate. These observations are formalized in the next section of the paper. In succeeding sections they are applied to models based on the Constant Elasticity of Substitution (CES) production function and to the vintage model of Malcomson and Prior (1979).

THE Federal Home Loan Bank Board (FHLBB), a government regulatory agency for member Savings and Loan Associations (SLA), provides advance loans (advances) to members. Advances are considered a major policy tool for the FHLBB in stabilizing deposit, mortgage, and housing markets. Indeed, the FHLBB has characterized advances as providing 'a central credit resource, capable of expanding and contracting to meet the needs of its member institutions for housing credit.' Debt issues in the government agency capital market are the primary source for FHLBB funds.2 The quantity of advances outstanding has grown from $5.3 billion at year-end 1968 to $32.7 billion at year-end 1978, and with considerable variation in between. A key issue in evaluating the role of advances has been whether or not the FHLBB uses nonprice rationing in allocating advance loans. Without nonprice rationing, the FHLBB sets the interest charge for the advances and member SLAs determine the quantity of loans they wish to borrow. In this case, an appraisal of FHLBB policy is relatively straightforward and can be based on the level of the advances rate. With nonprice rationing, in contrast, an appraisal of FHLBB policy is more complicated because the unobserved availability of advances must also be considered. The relevance of this issue is underscored by a recent report of the existence of nonprice rationing.3 Existing econometric models of the advances market treat the question of nonprice rationing in different ways. Hendershott (1977) has the quantity of advances determined by SLA demand without regard to interest rates, so there is neither price nor nonprice rationing. As a consequence, advances policy is viewed as purely passive. Kearl and Rosen (197-4) have the quantity of advances determined by a FHLBB reaction function, again with no role for interest rates, so they have a pure nonprice rationing system. In the MPS model (see Gramlich and Jaffee (1972)) both the quantity and interest rate for advances are set exogenously by the FHLBB, which then implies a combination of interest rate and nonprice rationing. Finally, Silber (1973), in the most sophisticated of the studies, has two versions, one with interest rate and one with nonprice rationing. The version with interest rate clearing has a FHLBB reaction function determining the interest rate and an SLA demand function determining the quantity. The version with nonprice rationing has the FHLBB determine the quantity of advances and no equation for the interest rate, in much the same spirit as Kearl and Rosen. The treatment of rationing in these models is not satisfactory. First, the models specify a priori the presence or absence of rationing, but without providing a test of the hypothesis that rationing takes place. Second, among the rationing models, the specifications are deficient in assuming either that only rationing occurs (Kearl and Rosen, and Silber) or that price and nonprice rationing always occur together (the MPS model). In this paper we develop a model based on optimizing behavior on the part of the FHLBB. Whether market clearing or nonprice rationing behavior obtains in a given time period depends on both economic conditions and the nature of the FHLBB's objective function. As a polar case, the general model collapses into an ordinary market clearing simultaneous equation model and we are able to offer statistical evidence as to which model is to be preferred. In section II.A we develop a market clearing model and in II. B a rationing model. A geometric interpretation is given in section II.C. Section III provides the empirical specification for impleReceived for publication November 6, 1978. Revision accepted for publication July 12, 1979. * We are indebted to NSF Grant SOC77-07680 and the Federal Home Loan Bank Board for support, to Naoyaki Yoshino for helpful comments, and to David Romer for expert research assistance. An earlier version of this paper was presented at the Econometric Society Meetings, Vienna, September 1977. 1 FHLBB Journal, April 1972, p. 24. 2 See Jaffee (1976) for a recent survey of FHLBB structure and policies. 3 Wall Street Journal, May 4, 1979, p. 18.

and not from the explicit examination of substitutability between private and public provision of goods and services. If private and government investment expenditures are perfect substitutes (the limiting case) then increases in government investment goods purchases financed by additional debt creation reduces potential output because private net-of-deficit savings decline by more than the increase in government investment at the initial level of output. Since government debt is viewed as an addition to wealth there is a decline in the relative desire to accumulate capital goods, whether private or public. Thus, the observation that there may be a potential decrease in steady state output is not due to perfect (or any other degree of) substitutability between private and public expenditures but rather is due to a less than perfect symmetry to the wealth effects associated with tax and deficit financing. The introduction of less-thanperfect' expenditure substitutability mitigates against this revenue composition effect, as a one dollar increase in government investment goods would initially cause a less than one dollar decrease in private investment demand. In like fashion, von Furstenberg's assertion that the marginal propensity to save must be unity if fiscal actions are not to affect steady state output (p. 77) is correct only within the context of his particular private savings function. If future tax liabilities are perfectly discounted then the marginal propensity to save will be unity out of the obtained by the private sector from changes in the form of financing government expenditures. Thus, alluding to an observed savings rate of 8% in the United States does not constitute any substantive evidence about possible impacts on potential output of changes in government expenditures, as that average savings rate cannot be applied to marginal changes in disposable income when future tax liabilities are fully, or even partially, discounted.

FOR over three decades, debate has raged over the economic assumption that the large corporation, through the decisions of its managers, attempts to maximize its profits. Empirical analysis of the behavior of the corporation has led to conflicting claims. The inquiry into the determinants of executive compensation has been no exception. Statistical investigation of executive compensation has been dominated by a search for one decisive explanation. Is the size, measured by either sales or assets, or profitability, measured by net corporate income or by the rate of return on assets, the key variable in establishing the level of the executive's reward? Proponents on both sides of this issue-the managerialists who support the corporate growth hypothesis and the neoclassical economists who favor the profit maximization assumption-seem to argue that the contest can be resolved by the presentation of unambiguous evidence that will award victory to one side and vanquish the other. This spirit of antagonism has distorted the essential element of the executive compensation question. The behavior of the corporation and the market forces that shape this behavior can be explained or illustrated only by the use of a series of intercorrelated variables. Not one of the available measures of corporate success, be it net income, sales or assets, is an exact measure of economic profits or firm size, nor is it independent of the other variables. This study focuses on the resolution of the serious econometric problems encountered in the process of estimating the determinants of executive compensation. Later we will show how the successful elimination of problems of simultaneous equations bias, multicollinearity and heteroscedasticity leads to the conclusion that the managerialist and neoclassical models of the firm are complementary, rather than substitute, explanations for the pattern of executive compensation.

The results presented below demonstrate that the structural modeling approach to interest rate determination not only stands apart from the sectoral disaggregation question conceptually but also performs fairly well without sectoral disaggregation empirically. This paper presents estimation and dynamic simulation results for an aggregated equivalent to the disaggregated model of the determination of bond yields developed in Friedman (1977; 1979). Instead of six bond demand and two bond supply equations, here there are but one demand and one supply equation. The empirical results show that, while disaggregation is of value in structural interest rate modeling (that is, the disaggregated model outperforms the aggregated one), even the aggregated structural model performs very well in comparison with familiar unrestricted reduced-form term structure equations.

straints that are imposed, the price vectors chosen to generate the data, and the functional form used for the cost function approximation. It is also fair to say that if input-output coefficients are the ones of interest, since these are obtained for each of the points used in the generation, it may be better to seek some approximating functions for these directly rather than seek an approximating function for the cost function and then derive the input-output coefficients. All these comments do not imply that the pseudo data approach should be given up. It is, however, important to sort out the aims of the analysis. There are some problems (like studying the effects of changes in environmental regulations) where one cannot get any answers from time series and one has to use the process analysis models that have detailed specification of technology and the constraints. But for this the appropriate thing is to do a simulation analysis of the process model itself and not seek a single equation approximation of the complex technology. What distin,guishes Griffin's approach from the garden variety simulation analyses of process models is this distillation in a single equation.

DISAGREEMENT about the responsiveness of housing demand to variations in relative prices persists, despite extensive empirical analyses. Two factors underlie this disagreement: the multidimensional character of housing makes direct observation of prices (as distinct from expenditures) impossible; and, the significant search, transactions, and moving costs associated with changing dwellings imply that, at any instant, a given household's consumption may deviate significantly from its utility maximizing level in a static equilibrium. While a number of ingenious attempts have been made to circumvent these problems, each is quite indirect and relies upon strong, and untestable, assumptions (cf. Mayo (1978)). This paper provides direct estimates of price elasticities, based upon an explicit model of housing consumption dynamics and utilizing the experimental manipulations of housing prices incorporated in the Housing Allowance Demand Experiments.1 While the data are limited to two years of longitudinal data and pertain only to low income renters, they nevertheless permit the direct estimation of this key parameter of housing demand. This analysis focuses on the price responsiveness of households, but clearly other changes in household circumstances (such as in income or family size) affect desired housing consumption. In fact, given the limited longitudinal data, information about other demand adjustments provides valuable insights into consumption dynamics. Price changes can be viewed as but one of a variety of exogenous influences on housing demand. A complete structural model of housing demand would consider the joint influence of household preferences, relocation costs, and prices on search and moving behavior and, conditional on this, their subsequent influence on housing consumption. However, both household preferences and relocation costs are generally unobserved, and estimation of such a complete model is simply beyond our current capabilities. We concentrate upon the more modest goal of modelling the reduced form relationship between housing consumption and housing stock disequilibrium (defined below). Consumption dynamics are represented by variants of a linearized stock adjustment process . This formulation is based on the simple observation that adjustments will generally be a monotonic function of the magnitude of disequilibrium in housing consumption. As indicated by past work (Hanushek and Quigley, 1979), this is both a convenient and powerful characterization of short run dynamics. Two basic formulations of consumption dynamics are considered. Let Htd represent the desired, or static equilibrium, quantity of housing demanded by a given household, and let Ht be the actual (observed) housing consumption at time t. In the simplest form, households are assumed, on average, to close the gap between desired and equilibrium housing consumption at a constant rate a, so that

Jon P. Nelson, George R. Neumann, Ronald M. Crandall, A Comparison of Alternative Econometric Models of Iron and Steel Investment Behavior, The Review of Economics and Statistics, Vol. 62, No. 1 (Feb., 1980), pp. 122-127