"The Lachmann Legacy: An Agenda for Macroeconomics"

      South African Journal of Economics
      vol. 65, no. 4 (December), 1997, pp. 459-481
      Redrafted here as Chapter 2 of Time and Money
    An Agenda for Macroeconomics

    Roger W. Garrison
     

    Adopting a means-ends framework for macroeconomic theorizing is a way of emphasizing the critical time dimension—the time that elapses between the employment of means and the achievement of ends. In a modern, decentralized, capital-intensive economy, the original means and the ultimate ends are linked by a myriad of decision of intervening entrepreneurs. As the market process moves forward, each entrepreneur is guided by circumstances created by the past decisions of all entrepreneurs and by expectations about the future decisions of consumers and of other entrepreneurs. These are the decisions that give rise to what Ludwig Lachmann has called an intricate latticework of capital and what we will call—to emphasize the time dimension—the intertemporal structure of capital.
            Austrian macroeconomics, then, concerns itself with two critical aspects of economic reality: the intertemporal capital structure and entrepreneurial expectations. Mainstream macroeconomics has long ignored the first-mentioned aspect but has become keenly attentive—almost obsessively attentive—to the second. On my interpretation, Lachmann's writings argue for a better balance of attention and suggest that the mainstream's overemphasis of expectations is directly related to its underemphasis of capital structure. 

    What About Expectations?
    There is some dispute concerning the Austrian school's attention to expectations as evidenced by conflicting views about the writings of Ludwig von Mises: "Mises always emphasized the role of expectations" (Phelps, 1970, p. 129); "Mises hardly ever mentions expectations" (Lachmann, 1976, p. 58). Is it possible that these seemingly opposing pronouncements are somehow both true? The "always" and even the "hardly ever" (Lachmann didn't say "never") make us suspect that both involve overstatement. But the validity of each derives from the different alternative treatments of expectations to which Misesian economics is being compared. Phelps was providing a contrast to the 1960s view of the tradeoff between inflation and unemployment. The idea that this tradeoff is a stable one and that it provides a menu of social choice for policymakers requires a wholesale neglect of expectations. Lachmann was providing a contrast to the 1930s view of investment in an uncertain world. Equilibration, according to the Swedish economists, involves a playoff between expected and realized values of the level of investment; persistent disequilibrium, according to Keynes, is attributable to the absence of any relevant and timely connection between long-term expectations and underlying economic realities. In comparison to Keynes and even the Swedes, Mises underemphasized expectations. This was Lachmann's judgment.
            In a letter of August 1989, Lachmann posed to me a direct question about Mises's and Hayek's neglect of expectation (a neglect he referred to in a subsequent letter as "a simple matter of historical fact"). "Do you agree with me that in the 1930s Hayek and Mises made a great mistake in neglecting expectations, in failing to extend Austrian subjectivism from preferences to expectations?" His particular phrasing of this question links it directly to his 1976 article, in which he traced the development of subjectivism "From Mises to Shackle." Also, Lachmann's question was a leading question, followed immediately with "What, in your view, are the most urgent tasks Austrians must now address?" Lachmann himself had spent several decades grappling with expectations. He recognized in an early article (1943) that expectations in economic theorizing presents us with a unique challenge. They cannot be regarded as exogenous variables. We must be able to give some account of "why they are what they are" (p. 65). But neither can expectations be regarded as endogenous variables. To do so would be to deny their inherent subjectivist quality. This challenge always emphasized but never actually met by Lachmann has been dubbed the "Lachmann problem" by Roger Koppl (1998, p. 61.)
            My response to Lachmann did not deal head-on with the Lachmann problem but focused instead on Hayek and Keynes and derived from considerations of strategy. Hayek was trying to counterbalance Keynes, whose theory featured expectations but neglected capital structure. Without an adequate theory of capital, expectations became the wild card in Keynes's arguments. Guided by his "vision" of economic reality, a vision that was set in his mind at and early age, he played this wild card selectively—ignoring expectations when the theory fit his vision, relying heavily on expectations when he had to make it fit. Hayek's countering strategy is made clear in his Pure Theory of Capital (pp. 407ff.): "[Our] task has been to bring out the importance of the real factors [as opposed to the psychological factors], which in contemporary discussion are increasingly disregarded." But in countering Keynes's "expectations without capital theory," Hayek produced—or so it could be argued—a "capital theory without expectations." In response to Lachmann's question about the most urgent tasks, I suggested that we need to put capital theory (with expectations) back into macroeconomics and that my inspiration for working in this direction was Lachmann's own writings.
            What I saw then as inspiration I see now as legacy. Though exhibiting increasing emphasis on the uncertain future and decreasing confidence that the market's equilibrium tendencies will prevail, Lachmann's writings—from his 1943 "Role of Expectations" article, to his 1956 Capital and Its Structure to his 1986 The Market as An Economic Process—were focused sharply on both capital and expectations. During the three decades that separated the two books, his own thinking grew ever closer to Shackle's. The macroeconomy to him became the kaleidic society. The existence of equilibrating forces was not in doubt. But neither was the existence of disequilibrating forces. And there was no way to know which, in the end, would win out. Among Austrian economists, Lachmann was virtually along in his agnosticism about the ability of the market economy to coordinate.
            If Lachmann's legacy is to bear fruit, today's Austrian macroeconomists will have to allow their thinking to be guided by the question "What about capital"? But as a preliminary task, they will have to respond effectively to the question that has become the litmus test for modern macroeconomic theorizing: "What about expectations?"
            So: what about expectations in today's macroeconomics? In earlier decades, this question could be asked out of concern about emphasis—too little or too much? But more recently the question is posed impishly—with serious doubts that any theory that does not feature so-called rational expectations can survive a candid response. The question has gotten the attention in recent years of defenders as well as critics of Austrian theory and particularly of the Austrian theory of the business cycle. But as we have seen, the challenge itself is not new to the Austrians. Hayek (1939) dealt early on with "Price Expectations, Monetary Disturbances, and Malinvestments." Lachmann (1943 and 1945) raised the issue anew—and with a hint of impishness—arguing that the treatment (or neglect) of expectations in Mises's account of business cycles constitutes the Achilles' heel of the Austrian theory. Mises's glib response (1943), in which he acknowledged an implicit assumption about expectations (their being fairly elastic), suggested that he did not take Lachmann's critical assessment to be a particularly hard-hitting one. More recently, however, critics within the Austrian school (e.g., Butos, 1997) have charged that modern Austrian macroeconomists ignore expectations or, at least, do not deal adequately with them.
            Modern defenders of the Austrian theory are often put on the spot to respond to these critics in a way that (1) recognizes the treatment of expectations as the sine qua non of business cycle theory it has come to be in modern macroeconomics, (2) reconciles the Austrian view with the kernel of truth in the rational expectations theory, and (3) absolves modern expositors of Austrian business cycle theory for not giving expectations their due. There is no direct answer, of course, that will satisfy the modern critic who issues the challenge in the form of the rhetorical question: "What about expectations?"—hence the impish tone with which it is posed.
            While my response to Lachmann in 1989 focused on the strategic considerations made by Hayek in his battle with Keynes, my reply to the imps of the 1990s hinges on the fact that Hayek lost the battle. Reflection reveals that this question, or, more accurately, the context in which it is asked, is wholly anachronistic. Modern treatments of expectations, which can be understood only in the context of the macroeconomics that grew out of the Keynesian revolution, cannot simply be grafted onto the Austrian theory, whose origins predate Keynes and whose development entailed an explicit rejection of Keynes's aggregation scheme. Accordingly, a brief history of macroeconomic thought is prerequisite to a satisfactory answer to any question about the role of expectations in the Austrian theory of the business cycle. 

    The Keynesian Spur
    It was in the 1930s that macroeconomics and, with it, business cycle theory, broke away as a separate subdiscipline. To describe the breakaway, some writers use terms such as Keynesian detour or Keynesian diversion, which suggest that the path of development was, for a time, less direct than it might have been; my Keynesian spur (analogous to a spur line of a railway) suggests development in the direction of a dead end. As Keynesianism worked its way through the profession, macroeconomics came to be defined not as a set of issues concerning the overall performance of the economy but as a particular way of theorizing about the economy. For purposes of gauging the economy's ability to employ resources, the new macroeconomics focused on the aggregate demand for output relative to the economy's potential output. For purposes of dealing with the issue of stability and charting the dynamic properties of the economy (such as those implied by the multiplier-accelerator process), the output of investment goods was separated from the output of consumption goods: Investment is the unstable component, and consumption is the stable component of aggregate demand. The summary treatment of inputs was even more severe. Consistent with the strong labor-market orientation, inputs were treated as if they consisted exclusively of labor or could be reckoned in labor-equivalent terms. The structure of capital was assumed fixed, the extent of its actual utilization changing in virtual lockstep with changes in the employment of labor. Income earned by workers was reckoned as the going wage rate times the number of (skill-adjusted) worker hours, and changes in labor income were taken to imply proportional changes in total income.
            Dropping out of the macroeconomic picture was any notion that labor income may move against other forms of income, as the classical economists had emphasized, as well as the notion that changes in the structure of capital—more of some kinds, less of other kinds—may figure importantly in the economy's overall performance. These changes in relative magnitudes, by virtue of their being relative changes, were no part of the new macroeconomics. In fact, it was the masking of all the economic forces that assert themselves within the designated aggregate magnitudes, particularly those that are at work within the investment aggregate, that allowed macroeconomics to make such a clean break from the pre-Keynesian modes of thought.
            Analytical simplicity was achieved in part by the aggregation per se and in part by the fact that the featured input aggregate was labor rather than capital. All the thorny issues of capital—involving unavoidable ambiguities in defining it, measuring it, and theorizing about it—were set aside as the simpler issues of labor became the near-exclusive focus. The preeminence of labor in this regard seemed almost self-justifying not only on the grounds of its relative simplicity but also on the grounds that it is our concern for workers, after all, and their periodically falling victim to economywide bouts of unemployment that justify our study of macroeconomic phenomena. Despite its being descriptively accurate, "labor-based macroeconomics" is a term not in general use today but only because virtually all modern macroeconomics is labor-based.
            A few noncontroversial propositions about spending on consumption goods as it relates to consumers' incomes is enough to establish a clear dependence of aggregate demand and hence aggregate income on investment spending, which—absent capital theory—seems to be rooted in psychology rather than in economics (Keynes, 1936, p. 161-63). It follows in short order that an economy dominated by such a dependency and constricted by an assumed fixity of the wage rate is inherently unstable. Movements in the investment aggregate, up or down, give rise to magnified movements—in the same direction—of income and consumption. Classical theory is reduced to the minimal role of identifying the level of income that constitutes full employment, implying that changes in the Keynesian aggregates are real changes for levels below full employment and nominal changes for levels above.
            A comparison of the Keynesian analytics with those that predate the breakaway of macroeconomics confirm that what counts in classical theorizing is the interplay among landlords, workers, capitalists, and entrepreneurs. Relative and sometimes opposing movements of the incomes associated with these four categories gives the economy its stability. For Keynes, all such relative movements were downplayed or ignored. It is as if an automotive engineer, in his quest for analytical simplicity, modeled a four-wheeled vehicle as a wheelbarrow and then declared it inherently unstable. To impose stability on the Keynesian wheelbarrow, some external entity would have to have a firm grip on both handles. Those handles, of course, took the form of fiscal policy and monetary policy. The mixed economy, whose market forces are continually countered by policy activism, could achieve a level of performance that a wholly private macroeconomy could never be able to achieve on its own. If sufficiently enlightened about the inherent flaws of capitalism, the fiscal and monetary authorities could keep the Keynesian wheelbarrow between the hedgeposts of unemployment and inflation.
            Although simple in the extreme, highly aggregative, labor-based macroeconomics was ripe for development. Questions about each of the aggregates and their relations to one another gave rise to virtually endless variations on a theme. What about consumer behavior? Beyond the simple linear relationship with current income, consumers may behave in accordance with the relative-income hypothesis (Duesenberry), the life-cycle hypothesis (Modigliani), or the permanent-income hypothesis (Friedman). What about the interest elasticity of the demand for money and of the demand for investment funds? Different assumptions, as might apply in the short run and the long run, allowed for some reconciliation between Keynesian and Monetarist views. What about wealth effects? What about investment lags? What about differential stickiness between wages and prices?
            The "what-about" questions served to enrich the research agenda of macroeconomics in all directions. The highly aggregative, labor-based macroeconomics survived them all, even thrived on them, by providing answers that set the stage for still more what-about questions. Even the critical question, "What about the real-cash-balance effect?," whose answer initially separated the Keynesians from the classicists, ultimately worked in favor of policy activism. The Keynesians embraced the notion that the economy could settle into an equilibrium characterized by persisting unemployment. Critics such as Haberler, Pigou, and eventually Patinkin argued that falling wages and prices would increase the real value of money holdings and that the spending out of these real cash balances would restore the economy to full employment. That is, even with all the other equilibrating forces buried deep in Keynes's macroeconomic aggregates, there remained a single margin (between money and output) on which to achieve a full-employment equilibrium. Real cash balances became, in effect, a balancing act that allowed the market economy to ride the Keynesian wheelbarrow as if it were a unicycle! Keynesians could concede the theoretical point while making the classically oriented critics look impractical if not downright foolish. If the critics willingly accepted Keynes's aggregation scheme, they would have to accept the policy implication of his theory as well. Considerations of practicality strongly favor a policy activism that takes the macroeconomy to be a Keynesian wheelbarrow rather than a policy of laissez-faire that presumes it to be a classical unicycle.
            The one exception to the agenda-expanding queries was the question that eventually came to be dreaded by practitioners of the new macroeconomics: What about expectations? In the face of the Monetarist counterrevolution and particularly the introduction of the expectations-augmented Phillips curve, it was no longer acceptable to assume that workers expect stable prices even as their real wage rate is being continually and dramatically eroded by inflation. The notion of a stable downward-sloping Phillips curve was no longer possible to maintain. Allowing workers to adjust their expectations of next year's rate of inflation on the basis of last year's experience did not much improve the theory's logical consistency or preserve its policy implications. The short-run Phillips curve was not exploitable in any welfare-enhancing sense. Even half-serious attempts to answer the question about expectations led to a contraction rather than an expansion of the research program. Logically consistent and rigorous answers led to a virtual implosion. If macroeconomists could provide simple answers to the what-about questions, why couldn't market participants? Some entrepreneurs and speculators could literally figure out the same things that the macroeconomists had figured out. Others could mimic these macro-savvy market participants, and still others could catch on if only by stumbling around in an economy where the highest profits go to those most in the know. Any theory about systematic macroeconomic relationships and certainly any policy recommendation would have to be based on the assumption of rational expectations.
            Embracing the rational-expectations theory had the effects of bringing long-run conclusions into the short run (Maddock and Carter, 1982), denying the possibility of using fiscal and monetary policy to stimulate or stabilize the economy (Sargent and Wallace, 1975 and 1976), and—despite the fact that these ideas were an outgrowth of Monetarism—questioning the importance of money in theorizing about the macroeconomy (Long and Plosser, 1983). The sequential attempts to deal with expectations became more and more directed towards preserving the internal logic of macroeconomics at the expense of maintaining a link between macroeconomic theory and macroeconomic reality. All too soon, the very idea of business cycles was purged of any meaning that might connect this term with actual historical events.
            Macroeconomics in the hands of the new classical economists, who tend to judge all other macroeconomic theories in terms of their treatment of expectations, lost the flavor but not the essence of its highly aggregative forerunners. The 1970s witnessed a search by macroeconomists for their microeconomic moorings. That is, recognizing that macroeconomics had pulled anchor in the 1930s and had been adrift for four decades, they sought to reanchor it in the fundamentals. The actual movement back to the fundamentals, however, affected form more than substance. The macroeconomic aggregates were replaced by representative agents. But the illusion of these agents forming expectations, making choices and otherwise doing their own thing is just that, an illusion. What the representative agent represents is the aggregate. Further, the things that the agent is imagined to be doing leave little scope for theorizing at either a microeconomic or a macroeconomic level. Phelps (1970, p. 5), who pioneered this search for microfoundations, clearly recognized the nature of the new classical theorizing: "On the ice-covered terrain of the Walrasian economy, the question of a connection between aggregate demand and the employment level is a little treacherous." The terrain is featureless, and the individuals, a.k.a. agents, are indistinguishable from the representative agent. (One is reminded of the once-popular poster showing ten thousand penguins dotting an ice-scape—with an anonymous penguin in the back ranks belting out the title bar of I Gotta Be Me.) In typical new classical models, the ice-scape is an especially bleak one, allowing for the existence of only one commodity. And to rule out such considerations as decisions about storing the commodity, leasing it, or capitalizing the value of its services, the single commodity is itself conceived as a service indistinguishable from the labor that renders it. This construction eliminates the need to distinguish even between the input and the output. In order to keep such an economy from degenerating into autarky, with each penguin rendering the service to himself, we are to think in terms of some particular service which, due to technological—or anatomical—considerations, one penguin has to render to another. "Back-scratching services" is offered as the paradigm case (Barro, 1981, p. 83).
            In their zeal to isolate the issue of expectations and elevate it to the status they believe it deserves in macroeconomics, the new classical economists have produced models whose sterility is matched by no other. Theorizing centers on the question of whether or not a change in the demand for the commodity is a real change or only a nominal change. The expectation that a change will prove to be only a nominal one implies that no real supply-side response is called for; the expectation that a change will prove to be a real one implies the need for a corresponding reallocation of the representative penguin's time—between scratching backs and consuming leisure.
            In order even to raise the issue of cyclical variation in output, new classical macroeconomists, whose models are constructed to deal explicitly and rigorously with expectations, must contrive some time element between (1) the observation of a change in demand and (2) the realization of the true nature (nominal or real) of the change. A construction introduced by Phelps (1970, p. 6.) involves a multiplicity of islands, each with its own underlying economic realities but all under the province of a single monetary authority. (Here, we overlook the fact that the very existence of money on the new classical ice-scape presents a puzzle in its own right.) In accordance with the fundamental truth in the quantity theory of money, a monetary expansion has a lasting influence only on nominal variables. Thus, in Phelps' construction, real changes are local; nominal changes are global. The representative penguin on a given island observes instantly each change in demand for the service but discovers only later (on the basis of information from distant islands) whether the change is nominal or real. The microeconomics of maximizing behavior in the face of uncertainty allows us to conclude that even before discovering the true nature of the change in demand, the representative penguin will respond to the change as if it were at least partially real. Monetary manipulation, then, can cause temporary changes in real magnitudes. This is the model that underlies the new classical monetary misperception theory of the business cycle.
            An alternative development of new classicism, one that avoids the contrived and theoretically troublesome notion of monetary misperception, simply denies the existence of business cycles as conventionally conceived—or as modeled with the aid of the distinction between local and global information. According to real business cycle theory, what appear to be cyclical variations in macroeconomic magnitudes are actually nothing more than market adjustments to randomly occurring technology shocks to the economy—even if the shocks themselves cannot always be independently identified. Changes in the money supply have nothing to do with these adjustments (or are an effect rather than a cause of them). Further, the adjustments take place at an optimum (profit-maximizing) pace. Whereas conventional macroeconomics attempts to track the cyclical variation of the economy's output around its trend-line growth path, real business cycle theory denies that trend-line growth can be meaningfully defined. It holds that actual variations in output reflect variations in the economy's potential. According to this strand of new classicism (and despite its being labeled real business cycle theory), movements in the macroeconomy's input and output magnitudes are not actually cyclical in any economically relevant sense.
            Still another alternative development closely tied to the idea of rational expectation is one that recognizes the possibility of macroeconomic downturns but denies any role to misperceptions. The variations in output can be attributed to certain obstacles (costs) that prevent the instant adjustment of nominal magnitudes. Technology shocks need not be the only source of change. Changes in the money supply can affect the economy, too. There are no significant information lags, but penguins cannot translate changes in demand instantaneously into the appropriate changes in nominal magnitudes. Prices are sticky. The stickiness, however, can be explained in terms of optimizing behavior and rational expectations. So-called menu costs (the costs of actually producing new menus, catalogs, and price tags) stand in the way of instantaneous price adjustments. These are the ideas of new Keynesian theory (Ball et al., 1988)—"Keynesian" because of price stickiness; "new" because the stickiness is not indicative of irrational behavior. (We will argue in Chapter 11 that new Keynesian ideas in the context of a complex decentralized capital-intensive economy are worthy of attention.)
            In response to the question "What about expectations?" we get new classical monetary misperception theory, real business cycle theory, and new Keynesian theory. This is the state of modern macroeconomics. While each of these theories include rigorous demonstrations that the assumptions about expectations are consistent with the theory itself, none are accompanied with persuasive reasons for believing that there is a connection between the theoretical construct and the actual performance of the economy over a sequence of booms and busts. Applicability has been sacrificed to rigor. The Keynesian spur has led us to this dead end. 

    Meeting the Challenge to Austrian Theory
    The very fact that the Austrian theory of the business cycle is offered as a theory applicable to many actual episodes of boom and bust—from the Great Depression to the Bush recession—seems to raise the suspicions of modern critics. If the theory has maintained its applicability, it obviously has not suffered the implosion that follows from the attempt to deal adequately—rigorously—with expectations. The critic imagines that he can stand flat-footed in front of an Austrian business cycle theorist, ask "What about expectations?" and then step back to watch the Austrian theory degenerate into some story about back-scratching penguins. The questioner expects that the Austrian theorist will first grapple ineffectively for an acceptable answer and then finally realize the true significance of this implosion-inducing question.
            Some modern Austrians (Butos and Koppl, 1993) have argued that dealing effectively with expectations may be a matter of doing the right kind of cognitive psychology. They suggest that Hayek's Sensory Order (1952), which deals with sensory data in the context of the structure of the human mind, may be relevant here. In this view, dealing with expectations consists not of choosing among alternative hypotheses (static, adaptive, rational) but of providing a theoretical account of the mental process through which expectations are formed and then integrating this theory with the theory of the business cycle. It is as if we must begin our story with photons striking the retinas of the entrepreneurs and end it with the ticker tape reporting the consequent capital gains and losses. Any actual integration of a theory of expectations in this sense with the theory of the business cycle would have to begin with a Crusoe economy and then, after introducing Friday, proceed to a barter economy, to a market economy, to a market economy with a central bank, and finally to an episode of credit expansion in this last-mentioned setting. This exercise may well have some payoffs, although it is difficult to anticipate even the nature of the possible payoffs. But surely it is doubtful that such a merging of cognitive psychology and macroeconomics, if in fact it could be done, would provide answers that would satisfy either the critics or the defenders of the Austrian theory.
            In light of the evolution of modern macroeconomic thought (from its break with the rest of economics and particularly with capital theory, to its simplification on the basis of the now-conventional macroeconomic aggregates, to its blossoming in the hands of practitioners exploring the many variations on a theme, to its eventual implosion in the face of embarrassing questions about expectations), the Austrians are ill-advised to take the question about expectations at face value. "What about expectations?" proved to be an embarrassing question for conventional macroeconomists; it need not be an embarrassing one for Austrian economists, whose theory has not suffered the same evolutionary fate. Further, the Austrians can hardly be expected to resist embarrassing the modern business cycle theorists by simply turning the impish question around and asking: "Expectations about what?" About changes in the overall levels of prices and wages? About price and quantity changes in a one-commodity world as perceived by a representative agent? About real and nominal changes in the demand for back scratching? It should go without saying that a satisfactory answer to the "Expectations about what?" question is a strict prerequisite to a satisfactory answer to the "What about expectations?" question. And for the Austrians, the prerequisite question is to be answered in terms of the macroeconomics that predates its breaking away from the fundamentals.
            In the Austrian view, the issues of macroeconomics are inextricably bound up with the issues of microeconomics and particularly with capital theory. The entrepreneurs, no one of whom is representative of the economy as a whole, influence and are influenced by one another as they bid for resources with which to carry out or possibly to modify their production plans. Conflicting plans involving the provision of immediately consumable services, such as Barro's back scratching, can be quickly reconciled as potential consumers make decisions about whether to purchase this service or to consume leisure and as they choose among the alternative providers of it. If an economy could be usefully modeled as the market for a single service provided by a representative supplier, there would not likely be any issues that would give macroeconomics a distinct subject matter. Important macroeconomic issues arise precisely to the extent that the economics of back scratching is not the paradigm case, which is to say, to the extent that inputs and outputs are not temporally coincident. If resources must be committed well before the ultimate satisfaction of consumer demand, then capital goods in some form must exist during the period that spans the initial expectations of the entrepreneur and the final choices of consumers. These capital goods can be conceived to include human capital as well as capital in the more conventional sense and to include durable capital goods as well as capital in the sense of goods in process.
            It is useful to think of the production process as being divided into stages of production such that the output of one stage is sold as input to a subsequent stage. Hayek (1967) employed a simple right triangle to depicted the capital-using economy—which gave him a leg up on Keynes, who paid no attention to production time. This little piece of geometry will become an key element of our capital-based macroeconomic model in Chapter 3. One leg of the triangle represents consumer spending, the macroeconomic magnitude that had the attention of both Keynes and Hayek; the other leg tracks the goods-in-process as the individual plans of producers transform labor and other resources into the goods that consumers buy. In Hayek's construction, human capital and durable capital are ruled out for the sake of keeping the theory tractable and developing a heuristic model, leaving us with the relatively simple conception of capital as goods in process with a sequence of entrepreneurs having command over these goods as they mature into consumable output. Still, there is a nontrivial answer to the "Expectations about what?" question. Complicating matters, however, is the fact that the sequence of stages is far from linear: There are many feedback loops, multiple-purpose outputs, and other instances of nonlinearities. Further, each stage may also involve the use of durable—but depreciating—capital goods, relatively specific and relatively nonspecific capital goods, and capital goods that are related with various degrees of substitutability and complementarity to the capital goods in other stages of production. These are the complications emphasized by Lachmann in his Capital and Its Structure.
            It is this context in which the Austrians can address the "Expectations about what?" question. The proximate objects of entrepreneurial expectations relevant to a particular stage of production include prices of inputs, which are the outputs of earlier stages, and prices of outputs, which are inputs for subsequent stages. The expected price differentials (between inputs and outputs) have to be assessed in the light of current loan rates and of alternative uses of existing capital goods. And judgments have to be made about possible changes in credit conditions and in the market conditions for the eventual consumer goods to which a particular stage of production contributes. Price, wage, and interest-rate changes will have an effect on entrepreneurs' decisions, and their decisions will have an effect on prices, wages, and interest rates. This interdependency is what justifies the general conception of the market as an economic process.
            The market process facilitates the translation of the underlying economic realities—resource availabilities, technology, and consumer preferences (including intertemporal preferences)—into production decisions guided by the expectations of the entrepreneurs. The process plays itself out differently depending upon whether the interest rate on which it is based is a faithful reflection of consumers' time preferences or, owing to credit expansion by the central bank, a distortion of those preferences. In the first case, the economy experiences sustainable growth; in the second, it experiences boom and bust. This is the essence of the Austrian theory of the business cycle (Mises et al., 1996; Garrison, 1986).
            Two "assumptions" (a more appropriate term here might be "understandings") about expectations are implicit in the Austrian theory: (1) the entrepreneurs do not already know—and cannot behave as if they already know—the underlying economic realities whose changing characteristics are conveyed by changes in prices, wages, and interest rates, and (2) prices, wages, and interest rates tend to facilitate the coordination of economic decisions and to keep those decisions in line with the underlying economic realities. Thinking broadly in terms of a market solution to the economic problem, we see that a violation of the first assumption implies a denial of the problem, while a violation of the second assumption implies a denial that the market is a viable solution. Taken together, these two assumptions do not allow us to categorize the Austrians' treatment of expectations as static, adaptive, or rational, as these terms have come to be used. But they do allow for a treatment of expectations that is consistent with the view that there is an economic problem and that the market is, at least potentially, a viable solution to that problem. And dealing with expectations in the context of a market process does give some basis for dealing with the Lachmann problem identified early in this chapter. Expectations can be regarded as largely endogenous—in a macroeconomically relevant sense—when the market process has been set against itself and largely exogenous otherwise.
            Consistency provides a standard by which the alternative treatments of expectations can be compared. After all, the idea of rational expectations stemmed from the recognition that the assumptions of static expectations and even of adaptive expectations were often inconsistent with the theories in which they were incorporated. Lucas (1987, p. 13) refers to the rational expectation hypothesis as a consistency axiom for economics. As such, the adjective "rational" refers neither to a characteristic of the market participant whose expectations are said to be rational nor to a quality of the expectations per se. It refers to only to the relationship between the assumption about expectations and the theory in which it is incorporated. The new classical assumption of rational expectations may well be consistent with the monetary misperception theory as set out in a Barro-style back-scratching model. But note that both the assumption and the model are inconsistent with there being a significant economic problem for which the market might provide a viable solution. Accordingly, a rational-expectations assumption plucked from a new classical formulation and inserted into Austrian theory—or into any other pre-Keynesian theory that affirms the existence of an economic problem—would involve an inconsistency, and hence, by the standard of consistency, would no longer be "rational." That is, it is not logically consistent to claim (1) that there is a representative agent who already has (or behaves as if he already has) the information about the underlying economic realities independent of current prices, wage rates and interest rates and (2) that it is prices, wage rates and interest rates that convey this information.
            The distinction between local and global information together with the information lag that attaches to global information allows for a telling point of comparison of new classical and Austrian views. In the new classical construction, this knowledge problem is contrived for the sake of modeling misperception. The representative agent sees changes in money prices immediately but sees evidence of changes in the money supply only belatedly. The agent does not know immediately, then, whether the change in the money prices reflects a real change or only a nominal change. In the Austrian theory, the treatment of the knowledge problem rests upon a different distinction between two kinds of knowledge—a distinction introduced by Hayek for the purpose of calling attention to the nature of the economic problem broadly conceived. Hayek (1945) distinguishes between the knowledge of the particular circumstances of time and place and knowledge of the structure of the economy. Roughly, the distinction is one between market savvy and theoretical understanding. It is not a contrivance for the purposes of modeling misperception but rather an acknowledgment of the fundamental insight most commonly associated with Adam Smith: The market economy works without the market participants themselves having to understand just how it works.
            The strong version of rational expectations employed by new classicism exhibits a certain symmetry with the notion of rational planning conceived by advocates economic centralization. The notions of both rational expectations and rational planning fail to give adequate recognition to Hayek's distinction between the two kinds of knowledge. Both employ the term "rational" to suggest, in effect, that reasonable assumptions about one kind of knowledge can (rationally) be extended to the other kind. Central planning could be an efficient means of allocating resources if the planners, who, we will assume, have a good theoretical understanding of the calculus of optimization, also had (or behaved as if they had) the knowledge that is actually dispersed among a multitude of entrepreneurs and other market participants. Symmetrically, monetary policy would have no systematic effect on markets if entrepreneurs and other market participants, whose knowledge of the particular circumstances of time and place are mobilized by those markets, also had (or behaved as if they had) a theoretical understanding of macroeconomic relationships. To recognize Hayek's distinction and its significance is simply to acknowledge that central planning is, in fact, not efficient and that monetary policy can, in fact, have systematic effects.
            Dealing with expectations in the context of Hayek's distinction rather than in the context of the contrived distinction between global and local knowledge adds a dimension to Austrian economics that can be no part of new classicism. While the global/local distinction is stipulated to separate two mutually exclusive kinds of knowledge, the two kinds of knowledge identified by Hayek exhibit an essential blending at the margin. Market participants must have some understanding of how markets work, if only to know that lowering a price is the appropriate response to a surplus and raising a price is the appropriate response to a shortage. Suppliers of particular products as well as traders in organized markets have a strong incentive to understand much more about their respective markets—about current and expected changes in market conditions and the implications for future prices. They know enough to make John Muth's (1961) treatment of expectations as applied to the hog market seem not only "rational" but eminently plausible. Symmetrically, economists-cum-policymakers must have some knowledge about the particulars of the economy in order to apply their theories to various existing circumstances. And to prescribe policies aimed at a particular goal, such as a specific unemployment rate or inflation rate, they would have to have a substantial amount of market information—about how changes in actual market conditions affect, for instance, the demand for labor and the demand for money.
            Further, the extent of the overlap is itself a matter of costs and benefits as experienced differentially by policymakers and by market participants. For policymakers, additions to their theoretical understanding are likely to be strongly complementary to existing understandings and may even have synergistic effects, while additional knowledge of the particular circumstances of time and place would likely involve high costs and low benefits. A symmetrical statement can be made about entrepreneurs with respect to costs and benefits of increased market savvy as compared to increased theoretical understanding. In general, specialists in one kind of knowledge experience sharply rising costs of—and sharply declining benefits to—the other kind of knowledge. Putting the matter in terms of costs and benefits suggests that the actual and/or perceived costs and benefits can change. Undoubtedly, the extent to which policymakers and market participants make use of both kinds of knowledge is dependent on the institutional setting and the policy regime. Reform in the direction of increased policy activism on the part of the central bank, for instance, will increase the benefits to entrepreneurs and other market participants of their understanding the short- and long-run relationships linking money growth to interest rates, prices, and wages. Stated negatively, entrepreneurs who experience a sequence of episodes in which the central bank is implementing stabilization policy or attempting to "grow the economy" may face a high cost of not understanding how money-supply decisions affect the market process.
            There is an overlap between the two kinds of knowledge and the extent of the overlap is itself a result of the market process. These aspects of Austrian theory have no counterpart in new classical theory. Expectations will be based on the knowledge of particular circumstances of time and place plus the understanding that corresponds to the overlap. Expectations are not rational in the strong sense of that term, but they do become more rational with increased levels of policy activism and with cumulative experience with the consequences of it. Equivalently stated, expectations are adaptive, but they adapt not just to changes in some particular price, wage rate, or interest rate, but also to the changing level of understanding that corresponds to the overlap. Finally and significantly, further development of the issue of expectations in the context of two kinds of knowledge and the market as an economic process will involve an expansion rather than an implosion of the Austrian research program. 

    What About Capital?
    If we think in terms of market solutions to economic problems, we must accord expectations a crucial role. But that role is overplayed if it is assumed that expectations come ready made on the basis of information that is actually revealed only as the market process unfolds; it is underplayed if it is assumed that expectations are and forever remain at odds with economic realities despite the unfolding of the market process. Either assumption would detract from the equally crucial role played by the market process itself, which alone can continuously inform expectations. On reflection, we see that the near-obsessive focus on expectations in modern labor-based macroeconomics owes much to the sterility of the theoretical constructions. There is simply not much of anything else to focus upon.
            What about capital? Much of Austrian theory is aimed—either directly or indirectly—at providing a satisfying answer to this question. And macroeconomists who think in terms of a entrepreneurial decisions in the context of a complex intertemporal capital structure have at the same time written much "about" expectations—even if that very word does not appear in their every sentence. Ludwig Lachmann's attention to expectations was always explicit as was his attention to capital and its structure. Accordingly, we can credit him for setting an important agenda for macroeconomics. Capital-based macroeconomics with due attention to entrepreneurial expectations and the market process can compare favorably with the modern labor-based macroeconomics and its assumption of rational expectations.