This paper analyzes the impact of trade on the stability properties of trading countries and on stationary welfare. We consider a two-country two-good two-factor overlapping generations model where countries differ in terms of their technology. In the autarky equilibrium and the free-trade equilibrium, indeterminacy relies, under dynamic efficiency, on a capital intensive consumption good and intermediate values of the elasticity of intertemporal substitution in consumption. Opening the borders to trade can be a source of a global destabilizing effect. Indeed, considering a free-trade equilibrium in which one country is an exporter of the consumption good and the other country is an exporter of the investment good, indeterminacy can occur with trade even though the two countries are determinate in autarky. Finally, opening to trade increases the stationary welfare of the country that exports the investment good and deteriorates the one of the other country.

We examine the impact of balanced-budget labor income taxes on the existence of expectation-driven business cycles in a two-sector version of the Schmitt-Grohé and Uribe (SGU) [(1997) Journal of Political Economy 105, 976–1000] model with constant government expenditures and counter-cyclical taxes. Our results show that the destabilizing impact of labor income taxes strongly depends on the capital intensity difference across sectors. Local indeterminacy is indeed more likely when the consumption good sector is capital intensive, as the minimal tax rate decreases, and less likely when the investment good sector is capital intensive, as the minimal tax rate increases. The implication of this result can be quantitatively significant. Indeed, when compared to SGU, local indeterminacy can be either completely ruled out for all OECD countries when the investment good is sufficiently capital intensive or drastically improved, delivering indeterminacy for a larger set of OECD countries, if the consumption good is sufficiently capital intensive. Focusing however on recent estimates of the sectoral capital shares corresponding to the empirically plausible case of a capital intensive consumption good, we find that there is a significant increase of the range of economically relevant labor tax rates (from a minimum tax rate of 30% to 24.7% for which local indeterminacy arises with respect to the aggregate formulation of SGU.

We analyze the importance of the frequency of decision making for macroeconomic dynamics, in the context of a simple, well-known business cycle model with balanced budget rules. We explain how the frequency of decision making (period length) and the measurement unit of time (calibration frequency) differ and examine how local stability is affected by changes in the period length. We find that as the period grows longer, indeterminacy occurs less often. This may have significant quantitative implications: for the model at hand, there is a wide range of economically relevant labor tax rates (from 30% to 38%) for which the continuous-time model gives indeterminacy, whereas the discrete-time model has determinate dynamics.

We consider an overlapping-generations economy with two consumption goods. There are two sectors that produce a pure consumption good and a mixed good that can be either consumed or used as capital. We prove that the existence of Pareto-optimal expectations-driven fluctuations is compatible with standard sectoral technologies if the share of the pure consumption good is low enough. Following Reichlin's [Journal of Economic Theory 40 (1986), 89–102] influential conclusion, this result suggests that some fiscal policy rules can prevent business-cycle fluctuations in the economy by driving it to the optimal steady state as soon as they are announced.

The present paper studies the influence of variable labor utilization on local indeterminacy and expectations-driven fluctuations, in one-sector models with (nearly) constant returns to scale. It is shown that, in comparison to the configuration of constant input utilization, considering variable utilization reduces the actual possibilities of factor substitution and, consequently, the range of input substitution elasticities that are compatible with endogenous fluctuations. In particular, local indeterminacy and expectations-driven fluctuations occur only if utilization rates are sufficiently inelastic, whereas local determinacy prevails when utilization is highly elastic. However, accounting for the fact that variable utilization reduces the effective elasticity of capital/labor substitution leads us to argue that expectations-driven fluctuations are more plausible because they require larger elasticities of apparent input substitution. In contrast with the recent literature, the analysis does not rely on significantly increasing returns to scale in production. Accordingly, the results are not at variance with recent empirical studies emphasizing the importance of variable utilization and denying the evidence of large increasing returns.