Real Business Cycle Theory

The model

Take a Ramsey-Cass-Koopmans economy. A representative household chooses consumption and labor supply; firms produce output Y using capital K and labor N according to a Cobb-Douglas function with stochastic productivity A:

Yt  =  At · Ktα · Nt1−α

The productivity process A is the only source of randomness. Kydland and Prescott assumed it follows an autoregressive process:

log At  =  ρ · log At−1  +  εt,    ε ~ N(0, σ2)

with ρ ≈ 0.95 and σ ≈ 0.007 calibrated to match U.S. productivity data. That's it. No demand shocks, no monetary shocks, no nominal rigidities, no fiscal policy. Households and firms optimize; markets clear; cycles emerge.

The propagation mechanism

A positive productivity shock arrives. Output rises directly. But that's only the impact effect — most of the cycle is what happens next. The Ramsey-style optimization creates four reinforcing responses:

1. Labor responds. A higher A raises the marginal product of labor, so real wages rise. Households substitute toward more work today (the intertemporal labor-supply margin), so hours N rise.
2. Investment surges. A higher A also raises the marginal product of capital. Households shift consumption forward to take advantage; saving and investment jump.
3. Consumption rises gradually. Permanent income rises (the shock is persistent), so consumption rises — but smoothly, because households spread the gain across time.
4. Capital accumulates. Today's investment becomes tomorrow's capital, so output stays elevated even after A starts reverting. This is the propagation that turns transient shocks into multi-quarter recessions and expansions.

All four variables — Y, N, C, I — rise together. They are procyclical. This co-movement is exactly what U.S. data show, and it is exactly what older Keynesian demand-shock models had trouble matching (demand shocks typically push C and I in opposite directions). RBC's empirical fit for procyclical co-movements was the central evidence cited for the theory.

Worked example — calibration and impulse response

Kydland-Prescott's headline calibration (1982, refined 1991):

• Capital share α = 0.33 (from national accounts).
• Quarterly discount factor β = 0.99 (annual real interest rate ≈ 4%).
• Capital depreciation δ = 0.025 quarterly (10% annual).
• Frisch labor elasticity 1 (the elasticity of hours with respect to wage holding marginal utility constant).
• Productivity autocorrelation ρ = 0.95; innovation std σ = 0.007.

The impulse response to a one-standard-deviation positive technology shock (about 0.7% in log A):

• Output Y rises ~1.0% on impact, peaks at ~1.3% after 4 quarters, decays back to baseline over 30+ quarters.
• Hours N rise ~0.5% on impact, peak around quarter 2.
• Consumption C rises ~0.3% on impact, smooth path, peaks around quarter 8.
• Investment I jumps ~3.0% on impact — five to seven times more volatile than output.

The volatility numbers match U.S. data with surprising fidelity: investment is indeed five to seven times more volatile than output; consumption is less volatile than output; hours and output have a correlation around 0.85. This calibration success was the empirical punch RBC delivered.

RBC vs other business-cycle theories

	RBC (Kydland-Prescott 1982)	Old Keynesian (Klein 1950s)	Monetarist (Friedman 1960s)	New Classical (Lucas 1972)	New Keynesian DSGE (Smets-Wouters 2003)	Austrian (Hayek 1930s)
Shock source	Real (technology, TFP)	Demand shocks	Money supply	Monetary surprises	Multiple: tech, demand, markup, monetary	Credit, malinvestment
Are cycles efficient?	Yes	No	No	If unanticipated	No (sticky prices)	No (capital misallocation)
Role of money	None or minimal	Strong	Central	Surprise only	Important (sticky)	Initial cause
Co-movement (C, I)	Procyclical, both	Trouble matching	Trouble matching	Procyclical	Procyclical	Procyclical, drift
Stabilization role	None / harmful	Strong (fiscal)	Money rule	Rules over discretion	Active (Taylor rule)	None (let cycles purge)
Foundation of	DSGE	1960s policy	Money supply rules	Modern macro methodology	Central-bank workhorses	Cycle-skeptics

Kydland and Prescott — calibration as method

Finn Kydland and Edward Prescott met at Carnegie-Mellon in the early 1970s. Their first major collaboration, the 1977 paper "Rules Rather than Discretion: The Inconsistency of Optimal Plans," won them the 2004 Nobel jointly with the RBC framework. That paper showed why discretionary monetary policy can produce inflation bias: even a benevolent central bank's promises become incredible if it has incentive to renege ex post. The Nobel committee cited both contributions together as constituting "dynamic macroeconomics."

Their methodological innovation was calibration. Where Cowles Commission econometricians estimated dozens of parameters from cyclical data, Kydland and Prescott pinned theirs down from long-run averages and microeconometric studies — then judged the model by whether its simulations matched cyclical moments. Critics (notably Larry Summers and Glenn Rudebusch) called this "shut up and estimate" econometrics replaced by "shut up and calibrate." Defenders argued calibration disciplines structural modelling against parameter overfitting. The methodological debate continues to this day.

Counterarguments

The Solow residual is not pure technology. Negative technology shocks would mean people forgot how to produce. That's hard to swallow. Empirically, the measured Solow residual is correlated with capacity utilization, inventory adjustment, labor hoarding, and oil prices — most of which are not pure technology. Basu, Fernald, Kimball (2006) constructed a refined utilization-corrected TFP series and found much smaller cyclical fluctuations.

Recessions don't look voluntary. RBC predicts that in a recession, workers choose to take more leisure because real wages temporarily fell. Survey evidence and unemployment statistics describe involuntary job loss — laid-off workers actively searching, not enjoying time off. The labor-market frictions literature (Mortensen-Pissarides) sought to fix this; modern DSGE typically includes search-and-matching.

Monetary shocks matter. Romer and Romer's narrative monetary-policy shocks, Christiano-Eichenbaum-Evans's VAR identifications, and the Volcker disinflation all show that monetary policy has real effects. Pure RBC says no — those effects must come from somewhere. New Keynesian DSGE accepts this and adds sticky prices.

2008 was not a negative technology shock. The financial crisis came from credit and balance-sheet stress, not a fall in productivity. Pure RBC has no story. Credit and financial-frictions DSGE (Bernanke-Gertler-Gilchrist 1999; Christiano-Motto-Rostagno 2014) extend the framework but go well beyond pure RBC.

Common pitfalls

• Reading RBC as "cycles are good". RBC says cycles are Pareto-optimal given the shocks. The shocks themselves can still be unwanted; the theory just says given them, the cycle is the efficient response.
• Confusing the Solow residual with technology. The residual is anything not explained by measured K and L. RBC theory uses it as a technology proxy, but most modern work treats this as a serious caveat.
• Forgetting capital accumulation. Most of the persistence in the RBC impulse response comes from capital accumulation, not from the AR(1) productivity process. Drop capital and you get much less propagation.
• Treating RBC as obsolete. It's superseded as a complete theory but its methodology — calibration, microfoundations, impulse-response evaluation — is alive and well in every modern DSGE model.
• Ignoring the Frisch elasticity. Hours response is highly sensitive to the assumed labor-supply elasticity. RBC needs a Frisch elasticity around 2-4 to match data; microeconometric estimates are typically below 1, generating tension.