Production, growth and business cycles: I. The basic neoclassical model

This 1988 Journal of Monetary Economics paper by Robert King, Charles Plosser, and Sergio Rebelo presents the neoclassical one-sector model of capital accumulation, augmented with a labor-leisure choice, as the basic organizing framework for real business cycle (RBC) analysis, and works out its implications for steady-state growth, for perfect-foresight capital-accumulation dynamics, and for fluctuations driven by technology shocks. The authors show that steady-state growth restricts admissible preferences to a constant-relative-risk-aversion-in-consumption class in which the income and substitution effects of labor-augmenting technical progress on labor supply exactly offset (Sec. 2.3, eqs. 2.8a-2.8b), and that off the steady state capital and work effort move in opposite directions along the transition path even though a higher capital stock raises the marginal product of labor, because the high real return associated with a low capital stock induces households to substitute intertemporally toward more work and more investment. With purely temporary (serially uncorrelated) productivity shocks, the model calls forth large contemporaneous responses of labor and investment – a one percent shock raises hours by about 1.33 percent in the baseline calibration – but produces essentially no serial correlation in output or employment, even though consumption and capital inherit substantial persistence from consumption-smoothing; reproducing the persistence seen in actual business cycles therefore requires building substantial serial correlation into the technology shocks themselves; the model’s own internal propagation is not enough. When technology shocks are instead assumed highly persistent (a quarterly autoregressive coefficient of 0.9), the baseline calibration (labor’s share 0.58, 10% annual depreciation, a 6.5% annual real return, and steady-state hours equal to 20% of the time endowment, matching 1948-1986 U.S. averages) reproduces the observed ranking of relative volatility across investment, wages, consumption and hours, but it overstates how closely consumption, investment and hours co-move with output relative to actual 1948.1-1986.4 U.S. quarterly data, and in particular the model’s near-unit correlation between hours and output is far higher than the near-zero raw correlation in the U.S. data (though correlations computed within five-year subperiods, or after low-frequency filtering, run substantially higher, around 0.3 to 0.8). The paper concludes that the basic neoclassical model, once persistent technology shocks are assumed, captures some central stylized facts of fluctuations – procyclical employment, consumption and investment, and the ranking of relative volatilities – but leaves the ultimate source of persistence unexplained (it derives mainly from the assumed persistence of shocks rather than the model’s internal dynamics) and abstracts from government and agent heterogeneity, motivating the companion essay’s stochastic-growth and multi-sector extensions.

Sergio T. Rebelo | Macro Paper Warehouse

Production, growth and business cycles: I. The basic neoclassical model