The conventional approach to estimating how fast economics converge examines the cross-economy relationship between the growth rate of per-capita output over some time period and its initial level. This approach produces consistent estimates only under highly restrictive assumptions, which are violated by the data. The paper develops an alternative approach that produces consistent estimates under weak assumptions. This approach yields estimates substantially larger than those reported in the literature and also sufficiently large to be broadly consistent with the predictions of neoclassical growth theory.

This paper derives three empirical implications of a well-specified model of consumer behavior, a model that nests both Ricardian equivalence and an alternative, non-Ricardian theory. Ricardian equivalence is then tested against this alternative. The tests reveal that Ricardian equivalence cannot be rejected.

In October 1979, the Federal Reserve began a policy of disinflation and also stopped stabilizing interest rates. Subsequently, interest rates became much more volatile. Using a methodology advanced by Barro, this paper estimates what this change in policy did to output in the United States. The paper finds that the stagnation of the U.S. economy since 1979 stemmed almost as much from increased interest rate volatility as from reduced money growth. The paper also assesses the extent to which increased volatility of money growth has contributed to stagnation. It finds no evidence of any appreciable contribution. The paper concludes that stabilizing interest rates is probably sensible monetary policy.

This paper estimates a general equilibrium model of suppressed inflation using quarterly U.S. data for the sample period 1942:IV-48:II. It finds that the pricelevel was reduced by at least 30.4 percent and that this suppression of inflation reduced employment by at least 11.7 percent and output by at least 7.1 percent. It also finds that monetary policy could have been equally successful in combating inflation but would not have had the side effect of reducing employment and output.

This paper investigates the relationship between nominal interest rates and prices using nearly two centuries of data from ten industrial countries. Both a positive relationship between interest rates and price levels (that is, a positive Gibson effect) and a negative relationship between interest rates and subsequent price changes (that is, a negative Fama-Fisher effect) prevailed until World War I. We propose a simple explanation wherein this doubly paradoxical juxtaposition of effects arises when money is supplied inelastically and prices are flexible. This double paradox disappeared after World War II when economies became mostly characterized by elastic money and sticky prices. During that period, a positive Fama-Fisher effect emerged while the Gibson effect largely dissipated.

This paper investigates whether the forty-eight contiguous U.S. states converge and, if so, whether convergence is absolute. Economies are shown to converge if, and only if, technology is stationary around a common trend. If convergence does occur, it is unlikely to be absolute unless the economy fixed effects in technology, capital's share, and the rental rate vanish. Examining data on the level of technology, capital's share, and the rental rate provides strong evidence that the continuous U.S. states converge rapidly to levels that are far apart. The rapidity of convergence suggests that factors and technology are highly mobile across the contiguous U.S. states.

Using panel data for the forty-eight contiguous U.S. states in each year between 1970 and 1986, this paper investigates the extent to which government capital and current government services contribute to private production. The paper finds fairly strong evidence that current government educational services are productive but no evidence that the other government activities considered are productive. Indeed, government capital often has statistically significant negative productivity. The results are robust across the many specifications considered.