Macroeconomics is at a crossroads. The call of real science is drawing it forward to a degree that is without precedent in the history of the field. But the field's origins are related to the exceptional policy relevance of macroeconomics, and the field continues to be pressed for answers to difficult policy problems that sometimes are beyond the current capabilities of the field. Continuing tensions exist between policy demands and the constraints of systematic, rigorous scientific development. A commonly mentioned example is the increasing size of structural macroeconometric models. While highly regarded in many governments, those massive models are greeted with skepticism by the best economics journals.

We use a general equilibrium model to study the impact of fully funding social security on the distribution of consumption across cohorts and over time. In an initial stationary equilibrium with an unfunded social security system, the capital/output ratio, debt/output ratio, and rate of return to capital are 3.2, 0.6, and 6.8%, respectively. In our first experiment, we suddenly terminate social security payments but compensate entitled generations by a massive one-time increase in government debt. Eventually, the aggregate physical capital stock rises by 40%, the return on capital falls to 4.4%, and the labor income tax rate falls from 33.9 to 14%. We estimate the size of the entitlement debt to be 2.7 times real GDP, which is paid off by levying a 38% labor income tax rate during the first 40 years of the transition. In our second experiment, we leave social security benefits untouched but force the government temporarily to increase the tax on labor income so as gradually to accumulate private physical capital, from the proceeds of which it eventually finances social security payments. This particular government-run funding scheme delivers larger efficiency gains (in both the exogenous and endogenous price cases) than privatization, an outcome stemming from the scheme's public provision of insurance both against life-span risk and labor income volatility.

We explore alternative approaches to numerical solutions of large rational-expectations models. We discuss and compare several current alternatives, focusing on the trade-offs in accuracy, space, and speed. The models range from representative-agent models with many goods and capital stocks, to models of heterogeneous agents with complete or incomplete asset markets. The methods include perturbation and projection methods. We show that these methods are capable of analyzing moderately large models even when we use only elementary, general-purpose numerical methods.

We examine a decision theoretic model of portfolio choice in which investors face income risk that is not directly insurable. We consider the sensitivity of savings and portfolio allocation rules to different assumptions about utility, the stochastic process for income and asset returns, and market frictions (transactions costs and short-sale constraints). Under CRRA time additive utility, habit persistence utility, and for a broad range of parameterizations, the model predicts that investors wish to borrow and invest all of their savings in stocks. This qualitative implication is robust to the introduction of significant transaction costs in the stock market, and contrasts sharply with portfolio allocation models in which there is no labor income.

We construct portfolios of stocks and bonds that are maximally predictable with respect to a set of ex-ante observable economic variables, and show that these levels of predictability are statistically significant, even after controlling for data-snooping biases. We disaggregate the sources of predictability by using several asset groups — sector portfolios, market-capitalization portfolios, and stock/bond/utility portfolios — and find that the sources of maximal predictability shift considerably across asset classes and sectors as the return horizon changes. Using three out-of-sample measures of predictability — forecast errors, Merton's market-timing measure, and the profitability of asset-allocation strategies based on maximizing predictability — we show that the predictability of the maximally predictable portfolio is genuine and economically significant.

Efficient Method of Moments is used to estimate and test continuous-time diffusion models for stock returns and interest rates. For stock returns, a four-state, two-factor diffusion with one state observed can account for the dynamics of the daily return on the S&P Composite Index, 1927–1987. This contrasts with results indicating that discrete-time, stochastic volatility models cannot explain these dynamics. For interest rates, a trivariate Yield-Factor Model is estimated from weekly, 1962–1995, Treasury rates. The Yield-Factor Model is sharply rejected, although extensions permitting convexities in the local variance come closer to fitting the data.

We consider an economy with an incomplete securities market and heterogeneously informed investors. Each investor trades in the market to hedge the risk to his endowment and to speculate on future security payoffs using his private information. We examine the efficiency of the securities market in allocating risk and transmitting information under different market structures, as defined by the set of securities traded in the market. We show that the introduction of derivative securities can decrease the market's efficiency in revealing information on security payoffs, and increase the equity premium and price volatility in the market.

The behavior of the real interest rate in a general equilibrium multisector model with irreversible investment is examined. It is shown that in such a model purely sectoral shocks can lead to substantial variation in the real interest rate and other aggregate time series. A source of variation in aggregate time series that is not found in one-sector models is thus examined, and the implications of this source of variation for the behavior of the interest rate are highlighted. Such a model seems to better capture the relationship among the real interest and output or investment than the standard one-sector stochastic growth model. It is also shown that, because of a desire to smooth consumption, with irreversible investment a rise in uncertainty concerning the future return to capital tends to lead to more current investment and a lower real interest rate.

This study examines the effect of short-sale constraints on a stock market, in particular, on stock prices, trading volume, and the relationship between stock price movements and output cycles. The economic model features incomplete markets and heterogeneous agents. The short-sale constraint is endogenously determined in the economy and is a function of agents' risk aversion, time preference, and exogenous driving forces. The dynamic model is solved using a policy function iteration algorithm. We find that, for an array of reasonable time-preference parameters and risk-aversion coefficients, the short sale limits range from 27 to 45% of total outstanding shares. Imposing short-sale constraints causes stock prices to move upward. Trading volume is high when some agents have a large amount of stock holdings but incur a negative shock on their nonfinancial income and is low when some agents have few stock holdings and also incur a negative shock to their nonfinancial income. Stock prices are found to be countercyclical and the expected stock returns are procyclical. These countercyclical stock-price movements are shown to be related to the imposition of a short-sale constraint.

We review the properties of algorithms that characterize the solution of the Bellman equation of a stochastic dynamic program, as the solution to a linear program. The variables in this problem are the ordinates of the value function; hence, the number of variables grows with the state space. For situations in which this size becomes computationally burdensome, we suggest the use of low-dimensional cubic-spline approximations to the value function. We show that fitting this approximation through linear programming provides upper and lower bounds on the solution to the original large problem. The information contained in these bounds leads to inexpensive improvements in the accuracy of approximate solutions.