Plate x in print

The Shock

real business cycles; the new-classical claim that booms and slumps are the economy's efficient response to real shocks, and idleness a choice.

To restore the old level of work, first ask what the work can now produce and what stands in the worker's way.

— the argument of this plate, after the Real plan, the earlier Lag, and the Prescott–Summers debate

An economy can produce less because the work has become less productive. A harvest fails, a machine breaks, or an essential input becomes harder to obtain; continuing to devote the old effort to the old activity can then waste resources. Real business cycle theory turns this familiar possibility into a much larger challenge. It asks how much of the movement in output and employment could arise from people adjusting sensibly to changes in productive opportunities. The earlier Lag briefly introduced the challenge, and the separate Real plan made it a question for this volume. A slump may contain a necessary response to worse circumstances. Establishing how much is necessary requires us to examine both the circumstances and the opportunities people actually have.

The real-business-cycle research associated with Kydland and Prescott puts changing productivity inside a model of growth, work, consumption, and saving. Households choose when to consume and how much time to devote to work; firms use labor and accumulated capital; disturbances and investment transmit effects through time. In the competitive benchmark, prices clear markets and the allocation is efficient under the model's assumptions. Prescott's 1986 account compared simulated fluctuations with the statistical behavior of postwar American aggregates and argued that productivity disturbances could explain much of the pattern. He also identified discrepancies, measurement problems, and unfinished extensions. The model makes an important possibility precise: fluctuations can occur even when the machinery of exchange functions as the model intends.1

1. Edward C. Prescott, Theory Ahead of Business Cycle Measurement (1986), pp. 9–21, particularly The Growth Model and Summary and Policy Implications. This is a university-hosted copy of the Minneapolis Federal Reserve's published article. Its argument concerns specified model economies and the historical comparison studied.

An invented one-person economy shows why restoring old output need not restore well-being. Suppose an hour produces ten units of a consumable good, while the personal cost of working rises with total hours: in the example's common valuation units, cost equals half the square of hours. Ten hours then produce one hundred units at a cost of fifty, leaving a net benefit of fifty. Now productivity falls to eight units an hour. The best choice becomes eight hours, producing sixty-four units at a cost of thirty-two. Insisting on the former ten-hour day produces eighty units but costs fifty, leaving thirty rather than thirty-two. Output rises under the insistence, while the person's welfare falls. The lost productive opportunity remains lost.2

2. An invented static illustration with consumption c = Ah and welfare U = c − h²/2, h ≥ 0. Maximizing Ah − h²/2 gives h = A. For A = 10, h = 10 and U = 50; for A = 8, h = 8 and U = 32. Imposing h = 10 after the shock gives U = 30. This deliberately simple example omits saving, uncertainty, income effects and exchange; it illustrates one welfare distinction, not the full dynamic RBC model.
⁂

The efficiency conclusion changes when the people choosing work cannot obtain the exchanges the calculation assumes. Our solitary producer controls the hours worked and receives every unit produced. A job-seeker facing an employer who cannot finance orders inhabits a different problem. There may be customers who value the goods and a worker willing to make them, yet no completed transaction. Similarly, uninsured losses, market power, or costs imposed on others can separate private choices from the best feasible allocation. Each addition changes the question the model must answer. Calling an outcome an equilibrium says that the specified choices fit together; calling it efficient requires further conditions. Neither designation, by itself, tells us how fairly the costs are distributed or whether the model describes the person outside the factory gate.3

3. This is the limitation exposed by the one-person example. Lawrence H. Summers's Some Skeptical Observations on Real Business Cycle Theory (1986), pp. 23–27, makes exchange failures a central objection to Prescott's account. The published article is retained here through a university-hosted copy; its historical empirical claims are not assumed to settle every subsequent model or episode.

A fall in measured productivity also needs an explanation independent of the name assigned to it. Suppose a factory records the same workforce and equipment as before but produces ninety units instead of one hundred. A damaged machine could explain the decline. So could a shortage of orders that leaves part of the equipment unused while the firm keeps its trained workers. A calculation based on recorded workers and equipment alone can register lower productivity in both cases. The measurement does not choose between them. Summers pressed this identification problem, along with the evidence for the labor responses the model required. The challenge is to establish the underlying disturbance without treating the outcome the theory seeks to explain as sufficient proof of its cause.4

4. The factory is an invented measurement example: with recorded inputs unchanged, output per recorded input falls 10 percent in either story. Summers, pp. 24–25, questions the identification of measured productivity changes with technology shocks and the evidence for intertemporal labor substitution. No historical recession is classified by this example.

Evidence capable of distinguishing the accounts must reach beyond their shared prediction that output can fall. It matters whether usable capacity disappeared, whether working time responded to changing returns, whether prices and wages moved as the proposed mechanism requires, and whether firms lost willing customers or lost the means to serve them. A convincing explanation should expose these implications to evidence that was not used merely to choose its parameters. It should also allow different mechanisms to meet. A genuine interruption in supply can reduce incomes and weaken demand elsewhere; a financial disruption can delay replacement of the equipment that would restore supply. Once such links are admitted, the practical question becomes how much of the decline reflects unavoidable scarcity and how much reflects additional failures that can be repaired.5

5. The distinction between a disturbance and the mechanisms carrying it forward belongs to the earlier Shock, on Frisch and Slutsky. That plate is a separate treatment. This expanded leaf develops the RBC question introduced in The Lag and the Real plan.

The real shock makes policy answer for the resources it asks people to use. If the economy has lost a productive opportunity, purchasing power alone cannot recreate it. If people have also lost access to worthwhile exchanges, accepting the entire decline as necessary leaves a different loss unexamined. The useful lesson is the discipline of comparing attainable alternatives: what work is possible, what it costs, who can undertake it, and which obstacle a proposed intervention would remove. That comparison respects the force of a real constraint while keeping the unemployed person's circumstances in view. The Stabilizer now turns to the government's choice—whether an intervention can improve the allocation, and whether it can arrive while the improvement is still available.