Existential Social Theory liberation. Reading Heidegger in relation to Hegel Friedrich (1844–1900); Sociology, History of; Weber, he produced in Being and Nothingness (1956 [1943]) Max (1864–1920); Weberian Social Thought, History an impressive restatement of the entire project of Of modernityasthediscoveryofhumanautonomy.Herethe commitment to freedom as the fundamental pre- requisite for any valuedform of human beingis restated with unsurpassed determination. The result is less a Bibliography secular version of a Christian notion of existence Camus A 1954 The Rebel [English translation by Bower A]. (Kierkegaard) than the consecration of a particular Knopf, New York kind of Romanticism (Fichte). This is characteristic Heidegger M 1962 [1927] Being and Time [English translation by also of Albert Camus (1913–60), whose The Rebel Macquarrie J, Robinson E]. Harper, New York (1954) returned to nineteenth-century Russian lit- Jaspers K 1933 Man in the Modern Age [English translation by erature for an understanding of nihilism as the most Paul E, Paul C]. Routledge, London consistent and fearless of existential views of the Jaspers K 1963 General Psychopathology [English translation human world. Only the free act could be considered by Hoening J, Hamilton M]. University of Chicago Press, authentically human, and as conformity to any exist- Chicago, IL Kierkegaard S 1978 [1845] Two Ages [English translation by ing convention was tainted with coercion, acts of Hong H V, Hong E H (eds.)]. Princeton University Press, rebellion bore the only indubitable sign of freedom Princeton, NJ and, therefore, of authenticity. Kierkegaard S 1980 [1844] The Concept of Anxiety [English Sartre increasingly sought for his existentialism a translation by Hong H V, Hong E H (eds.)]. Princeton larger and more dramatic stage; history and the University Press, Princeton, NJ collective subject, rather than the small individual acts Kierkegaard S 1987 [1843] Either\Or [English translation by of everyday life, supplied the setting for the working Hong H V, Hong E H (eds.)]. Princeton University Press, out of the great refusal; the negation of everything Princeton, NJ intolerable in modern life. The attempted synthesis of Langiulli N 1997 European Existentialism. Transaction, New existentialist and Marxist traditions resulted, in Crit- Brunswick, Canada and London Nietzsche F 1986 Human, All Too Human [English translation by ique of Dialectical Reason (1976 [1960]), in a social Hollingdale R J]. Cambridge University Press, Cambridge, theory paradoxically ‘founded’ on ‘anti-foundational’ UK existential insights. Through this Sartre sought to Sartre J-P 1956 [1943] Being and Nothingness [English translation establish praxis relevant to contemporary life rather by Barnes H]. Philosophical Library, New York than an explanation of its various features. Sartre J-P 1976 [1960] Critique of Dialectical Reason [English His views, thus, remain rooted in existential insight; translation by Sheridan-Smith A]. New Left Books, London ‘The only concrete basis for the historical dialectic is Weber M 1948 From Max Weber: Essays in Sociology [English the dialectical structure of individual acts.’ The radical translation by Gerth H H, Wright Mills C (eds.)]. Routledge freedom central to Being and Nothingness is somewhat and Kegan Paul, London muted here—reciprocity and otherness, mediated H. Ferguson through a series of ‘objective’ forms (dyad, triad, serial group, fused group, statutory group, organization, class), are conceived as pre-given and as essential aspects of, and constraints upon, every human project. However the subversive genius of its empty form is also invoked: ‘For freedom is nothing other than a Expectations, of choice which creates for itself its own possibilities.’ The human remains the being for whom ‘The upsurge of freedom is immediate and concrete.’ ‘Expectations’ in economics refers to the forecasts or Existential social theory seeks to realize two distinct views that decision makers hold about future , and seemingly incompatible values; the absolute and sales, incomes, taxes, or other key variables. The untrammeled freedom of the subject; and the actualiz- importance of expectations is due to their often ation of authentic selfhood. Much of the development substantial impact on the current choices of firms and of existentialism is an attempt to reconcile these two , and hence on current prices and the values and to imagine a social world in which they overall level of economic activity. could be harmonically interrelated. The urgency and difficulty of such a task is the source of the pathos, which distinguishes all serious existential social theory. 1. Expectations in the History of Thought Modern economic theory recognizes that the central See also: Marxist Social Thought, History of; Mod- difference between economics and natural sciences lies ernity; Modernity: History of the Concept; Modern- in the forward-looking decisions made by economic ization and Modernity in History; Nietzsche, agents. Therefore expectations are a basic building

5060 Expectations, Economics of block of economic theories. For example, in con- and it is currently the benchmark paradigm in both sumption theory the paradigm life cycle and per- micro- and . Some recent research manent income approaches stress the role of expected has gone beyond RE by developing models of learning future incomes. In decisions present behaŠior with explicit theories of data collection and calculations are conditional on expected future prices forecasting. and sales. Equity prices, rates, and exchange In this article we review developments in the rates all clearly depend on expected future prices. modeling of expectations formation, with an emphasis A central aspect of economic theories is that on and learning behavior. RE expectations influence the time path of the economy, modeling is the subject of many books, e.g., Sargent and conversely one might reasonably hypothesize that (1987) and Farmer (1999). Evans and Honkapohja the time path influences expectations. The current (2001) is a treatise on the learning approach. standard methodology for modeling expectations is to assume rational expectations (RE), which is in fact an equilibrium in this two-sided relationship. 2. Traditional Models RE modeling is a recent key step in a long line of dynamic theories which have emphasized the role of We will illustrate the modeling of expectations with expectations. The earliest references to economic some well-known simple models. The first example is expectations or forecasts date to the ancient Greek the . Consider a single competitive philosophers and the Bible. Systematic economic in which there is a time lag in production. analyses in which expectations play a major role began Demand is assumed to depend negatively on the as early as Henry Thornton’s treatment of paper prevailing market : credit, published in 1802, and E; mile Cheysson’s 1887 l k jŠ formulation of a framework which had features of the dt mI mp pt "t (1) ‘cobweb’ cycle. The role of expectations was given some attention by the classical , but their while supply depends positively on the expected price: method of analysis was based on the stationary state in l j ejŠ which perfect foresight prevails. Expectations were st rI rp pt #t (2) equated with actual outcomes, which downplayed " their significance. where mp, rp 0 and mI and rI denote the intercepts. is credited with the notion of ‘static We have introduced shocks to both demand and Š Š expectations’ of prices. The ‘cobweb’ model of a market supply. "t, #t are exogenous iid (identically and with a production lag was one of the first formal independently distributed) random variables with models with expectations in the 1930s. In the same mean zero and constant variance. decade, the temporary equilibrium approach, initiated The interpretation of the supply function is that by the Stockholm school, explicitly introduced expec- there is a one-period production lag, so that supply tations of future prices influencing current demands decisions for period t must be based on information and supplies. (1961) was the first to available at time tk1. For simplicity we make the formulate the notion of rational expectations and did representative assumption that all agents have so in the context of the cobweb model. the same expectation. Extensions to heterogeneous In macroeconomic contexts the importance of the expectations have been analyzed in the literature. state of long-term expectations of prospective yields We assume that markets clear. The observed price is for investment and asset prices was emphasized by then obtained by equating st and dt, which leads to (1936) in his General Theory. l µjα ejη Keynes stressed the central role of expectations for the pt pt t (3) determination of output and employment, but did not have an explicit model of how expectations are formed. where µ l (m kr )\m and α lkr \m ! 0. η l Š kŠ \ I I p η "p p σ# t η He even sometimes suggested that attempting to ( "t #t) mp so that we can write t iid(0, η), i.e., t forecast very distant future events can virtually over- is an iid random variable with mean zero and variance # whelm rational calculation. In the 1950s and 1960s ση. expectations were introduced into almost every area of Equation (3) is an example of a temporary equi- macroeconomics, including , investment, librium relationship. It illustrates the central role of demand and inflation using adaptiŠe expec- expectations by showing how the current market- tations or related schemes. clearing price depends on expected prices. Develop- Rational expectations made the decisive appearance ments since the Stockholm School and Keynes can be in macroeconomics in the work of Robert E. Lucas Jr. seen as different theories of expectations formation, and Thomas J. Sargent in the beginning of 1970s. i.e., how to close the model so that it constitutes a fully Many of the key contributions are collected in Lucas specified dynamic theory. We now describe briefly and Sargent (1981) and Lucas (1981). The RE hy- some of the most widely used schemes with the aid of pothesis became widely used in the 1970s and 1980s this example.

5061 Expectations, Economics of

2.1 Static Expectations be readily available. The optimal forecast method depends on the stochastic process of the variable being Naive or static expectations were used widely in the forecast and this implies interdependency between the early literature. In the context of the cobweb model forecasting method and the . On this they take the form approach we write e l pt pt−" (4) e l pt Et−"pt (6) l Once this is substituted into Eqn. (3) one obtains pt µjα jη for the cobweb example. Et−"pt denotes the math- pt− t, which is a stochastic process known as " ematical expectation of pt conditional on variables an autoregressive process of first order (AR(1)). This observable at time tk1 (including past data). and related stochastic processes are studied in stan- RE is an equilibrium concept. The actual stochastic dard textbooks on time series analysis and econo- process followed by prices depends on the forecast metrics. rules used by agents, so that the optimal choice of the In the early literature there were no random shocks, l µjα forecast rule by any agent is conditional on the choices yielding a simple difference equation pt pt−". of others. An RE equilibrium imposes the consistency This immediately led to the question of whether the condition that each agent’s choice is a best response to generated sequence of prices converged to the station- the choices by others. In the simplest models we have ary state over time. The convergence condition is, of representative agents and these choices are identical. course, QαQ ! 1. Whether this is satisfied depends on the l µjα jη For the cobweb model we have pt Et−"pt t. relative slopes of the demand and supply curves. Taking conditional expectations E of both sides In the stochastic case this condition determines l µjα t−" yields Et−"pt Et−"pt so that expectations are whether the price converges to a stationary stochastic l kα −"µ l given by Et−"pt (1 ) and we have pt process. (Loosely speaking, a stationary stochastic pro- kα −"µjη (1 ) t. This is the unique way to form expec- cess is a non-explosive process with statistical prop- tations which are ‘rational’ in the model (3). erties that are unchanging over time.) Two related observations should be made. First, under RE the appropriate way to form expectations depends on the stochastic process followed by the Š η 2.2 Adapti e Expectations exogenous variables, t in the example. If these are not The origins of the adaptiŠe expectations hypothesis can iid processes then the RE will themselves be random be traced back to . It was formally variables, and they often form a complicated stoch- introduced in the 1950s by Phillip Cagan, Milton astic process. Second, neither static nor adaptive Friedman, and Marc Nerlove. In terms of the price expectations are in general rational, except in special level the hypothesis takes the form cases. Next, we consider some key aspects of RE models. e l e jλ k e pt pt−" (pt−" pt−") (5) For the cobweb model it can be shown that both 3.1 Forward-looking Character expectations and prices converge to stationary stoch- In the cobweb model expectations pertain only to astic processes, provided the stability condition Q kλ kα Q ! current prices. Most economic models have forward- 1 (1 ) 1 is met. looking elements, so that expectations about future can equivalently be written periods appear. A simple example is the Cagan model as a distributed lag with weights declining exponen- kλ of inflation which postulates that the demand for tially at rate 1 . Besides adaptive expectations other money depends linearly on expected inflation: distributed lag formulations were used in the literature to allow for extrapolative or regressive elements. m kp lkψ( pe kp )jζ , ψ " 0 (7) Adaptive expectations played a prominent role in t t t+" t t macroeconomics in the 1960s and 1970s. For example, ζ where mt is the log of the at time t, t is inflation expectations were often modeled adaptively an iid money with zero mean, pt is the in the analysis of the expectations augmented Phillips e log of the at time t, and pt+" denotes the curve. expectation of pt+" formed in time t. Solving for pt we obtain the equation l α e jβ jŠ 3. Rational Expectations pt pt+" mt t (8) α l ψ jψ −" β l jψ −" Š l The RE revolution begins with the observations that where (1 ) , (1 ) , and t k jψ −"ζ ! α ! β l kα adaptive expectations, or any other fixed weight (1 ) t. Note that 0 1 and 1 . distributed lag equation, may provide poor forecasts The same formal model arises in other contexts, e.g., in certain contexts and that better forecast rules may the standard model of with risk neutrality

5062 Expectations, Economics of takes the form of Eqn. (8) with different values for α If data on expectations, for example from surveys, and β. Under RE the model is are available, then this provides a straightforward way to test for . Suppose we have time series p l αE p jβm jŠ (9) t t t+" t t data on, say, the average one-period ahead forecast ft of p , where the forecast f is made based on I , the Assume that m follows some general exogenous t t t−" t information available at time tk1. Under RE f l stochastic process. One can show that the expression t E(p Q I ) and we therefore test whether E(e Q I ) l 0, t t−"l k t t−" _ where et pt ft. In practice, the test is implemented l Š jβ αi pt t Etmt+i (10) by a regression: i=! l φh j et zt−" ut (11) solves the model, provided that the sum converges. Given an exogenous stochastic process for mt, where z is a vector of variables in I . Under the null t−" t−φ"l techniques are available for computing the forecasts hypothesis of RE the restrictions H!: 0 hold and ut Etmt+i and hence the RE solution (10). is serially uncorrelated. H! can be tested using the An important feature of the solution (10) is its standard F-test under some additional assumptions. forward-looking character. The current price level is a The intuition for the tests is straightforward. If φ  0 sum of expected values of future money stocks. If, for then forecast errors are systematically correlated with example, there is a change in which is zt−" and the information in zt−" is not being fully anticipated to lower the money stock from some exploited. specified future date T " t, this will be reflected in the Tests can also be conducted using forecast data with price level already at time t. Previously unexpected multiple period horizons and\or panel data, although changes in policy will affect pt from the moment they these lead to econometric complications. In practice, become known. tests of rationality using survey data on forecasts typically lead to rejections of rationality, although the interpretation of these results is controversial. Sup- 3.2 Solutions to Linear RE Models porters of RE often argue that the individuals surveyed Equation (9) is a simple example of a linear RE model. are not the relevant group of agents or that the More general models allow for a dependence on lagged expectations are measured with error. However, rejec- values of the endogenous variable p , expectations tions of rationality also raise the possibility that agents t do deviate from full RE owing to learning, forecasting formed at different times and over various horizons, \ and more general exogenous processes. Generaliz- costs, and or strategic . See Pesaran (1987) ations to multivariate frameworks are particularly for further discussion of econometric tests of the important in practice. rational expectations hypothesis. A number of techniques are available for obtaining solutions under RE. Undetermined coefficient meth- ods are often particularly convenient. Frequently, as 4.2 The in the model (9) with QαQ ! 1, there is a unique During the 1950s and 1960s, applied macroecono- nonexplosive solution, given nonexplosive exogenous metric models explicitly or implicitly assumed that variables. In this case a method called the expectations were given by adaptive expectations or Blanchard–Kahn technique can be used to compute by some related lag scheme. The estimated systems this solution. were used for policy analysis by evaluating the effects of alternative policies. This implicitly assumed that the estimated coefficients reflect a structure that is in- 4. Econometric Issues with Rational Expectations variant to alternative policies. Lucas (1976) criticized this approach, arguing that if expectations are 4.1 Testing for Rationality rational, then the coefficients relating target variables to policy instruments will change when the policy If the RE hypothesis is correct, then expectations obey process changes. very strong assumptions. The law of mathematical As an example, consider the Cagan model (9) with conditional expectations, from probability theory, β l kα Q Q l Q 1 and assume that money supply follows the states that E(E(p I) S) E(p S)ifS and I are rule alternative information sets with S 9 I. Here E(p Q Ω) denotes the mathematical expectation of p conditional m l µjρm jw (12) on an information set Ω. This result immediately t t−" t Q l l k Q QρQ ! implies E(e S) 0, where e p E(p I). where 1 and wt is an (unforecastable) mean zero This result, called the orthogonality principle, implies iid shock. In this case the solution (10) reduces to that the RE forecast error e must be uncorrelated with l δjγ jŠ every variable in the information set. pt mt t (13) 5063 Expectations, Economics of

l where bubble, which is any process Bt satisfying EtBt+" α−" α ! Bt. The bubble solutions are ‘explosive’ since 1, δ l (1kαρ)−"αµ and (14a) and macroeconomists are often content to assume γ l (1kρα)−"(1kα) (14b) away ‘explosive bubbles,’ although rational bubbles that burst periodically can be constructed. In models The ‘traditional’ approach to policy would estimate δ with carefully elaborated , explosive and γ from historical data and use the estimated solutions can often be ruled out theoretically. None- version of Eqn. (13) to evaluate the effects of alterna- theless, the empirical issue of whether rational (or almost rational) bubbles affect asset prices is still tive policy sequences mt on the price level. However, it is clear from Eqn. (14) that γ is not invariant to the debated. policy rule since, for example, a change in the policy Other models have multiple RE solutions, none of parameter ρ affects the value of γ. Thus, under RE, the which are explosive. As a simple example, consider the l α jβ jŠ QαQ " response of target variables to policy variables depends model pt Et pt+" mt t with 1. (This case on the form of the policy rule. can arise from linearized versions of the overlapping The effect of the ‘Lucas critique’ has been to shift generations model of money under some specifications - attention to estimation of deep parameters (such as α of preferences.) If mt is iid with mean m, from Eqn. (10) l kα −"αβ - jβ jŠ in the above example), which do remain invariant to there is the RE solution pt (1 ) m mt t. policy changes and to use these estimates to evaluate However, there are also solutions of the form alternative policy rules under RE. l α−" kβα−" kα−"Š jε pt pt−" mt−" t−" t (15)

4.3 Estimation and Testing of RE Models where ε is an arbitrary process, observable at time t t ε l and satisfying Et−"( t) 0. When data on expectations are not available, it is still These solutions are well-behaved nonexplosive pro- often possible in the context of structural models to ε cesses. The variable t is often called a ‘sunspot’ both estimate the models under the assumption of RE variable and the corresponding equilibrium a ‘sunspot and to test the RE assumption. In this case the RE solution.’ The term sunspot is used to indicate the assumption is being tested jointly with the model. A arbitrary nature of the variable. The solution depends simple example is the Cagan model (9), with β l 1kα, ε on t only because everyone believes that it does. It together with the money supply rule (12). Equations does not appear in the basic model structure. (12) and (13) can both be estimated, but Eqns (14a,b) Guesnerie and Woodford (1992) provide a review of imply that there are nonlinear cross-equation par- the results on sunspot equilibria and endogenous ameter restrictions that are imposed by the RE fluctuations. hypothesis. The system can be estimated under these It has been demonstrated formally that sunspot overidentifying restrictions. The restrictions can also solutions can exist in overlapping generations models be statistically tested, in effect testing RE jointly with in which the microfoundations are carefully con- the model. structed. More recently it has been shown that, with Another strand that has arisen recently is the increasing and monopolistic com- estimation and testing of Euler equations. These are petition, sunspot solutions can exist in variations of dynamic equations, relating, e.g., current consump- common models. These models raise the tion to expected future consumption and interest rates, possibility that business cycle fluctuations may be due which arise as optimality conditions from dynamic to random but rational fluctuations in and maximization. Under RE testable overidentify- firm expectations as a ‘self-fulfilling prophecy.’ The ing restrictions again arise. empirical significance of self-fulfilling prophecies re- mains an open issue. There are numerous other examples of multiple 5. Multiple Equilibria and RE equilibria under RE. One example is provided by a l α jβ variation of the Cagan model pt Et pt+" mt, Many RE models can have multiple equilibria. The where now money supply is assumed to follow a existence of rational speculative bubble solutions l - jξ j feedback rule mt m pt−" ut. (Here also for con- besides the fundamental equilibrium illustrates this venience Š l 0). This leads to the equation phenomenon. The standard model of asset pricing is t formally identical with the Cagan model (8) if we set l β ` jα jβξ j⠊  β l α l j −" pt m Et pt+ pt− ut (16) t 0, 1, (1 r) , where r is the real " " . Here mt represents the dividend at t and For many parameter values this equation yields two pt is the price of the asset. Then Eqn. (10) is called the ‘fundamental solution,’ i.e., the present value of RE solutions of the form anticipated future dividends. A ‘rational bubble sol- l j j ution’ is the sum of the fundamental solution and a pt k" k#pt−" k$ut (17) 5064 Expectations, Economics of where the ki depend on the original parameters which there are two solutions of the form (17). For the α, β, ξ, m- . In some cases both of these solutions are pure RE approach this is a conundrum. Which stochastically stationary. solution should we and the agents choose? Similarly, l e Other types of multiplicity of RE solutions arise in in nonlinear models of the form yt F(yt+") with nonlinear models. Many nonlinear models can be put multiple steady states, cyclic solutions or , in the general form one can ask which equilibria are stable under learning. The adaptive learning approach provides answers to l e yt F ( yt+") (18) all these questions, as discussed in the next section. Finally, the transition under learning to RE may where random shocks have here been left out for itself be of interest. The process of learning adds simplicity. If y is increasing in ye , i.e., Fh " 0, there t t+" dynamics that are not present under strict rationality may be multiple steady states that satisfy y- l F(y- ), and they may be of empirical importance. In the cases i.e., that occur at the intersection of the graph of F(.) just described these dynamics disappear asymptoti- and the 45m line. This possibility can arise in models cally. However, there are various situations in which with increasing returns to scale in production, one can expect learning dynamics to remain important , or monopolistic . y is often a over time. For example, if the economy undergoes measure of output or aggregate economic activity and structural shifts from time to time then agents will the low steady states represent inefficient ‘coordination need periodically to relearn the relevant stochastic failures’ (Cooper 1999). Other specifications of Eqn. processes. (18) have multiple perfect foresight equilibria taking the form of regular cycles in addition to a steady state. Nonlinear models can also exhibit sunspot equi- libria, taking the form of a finite state Markov process. 6.1 Statistical Approach to Learning For example, in models with two steady states it can be In the adaptive learning approach economic agents shown that there are RE solutions, depending on a behave like statisticians or econometricians when two-state Markov sunspot process, in which the forecasting economic variables needed in their decision solution alternates between values close to the perfect making. The economy is taken to be in a temporary foresight steady states. equilibrium in which the current state of the economy depends on expectations. The learning approach to 6. Learning expectations formation makes the forecast functions and the estimation of their parameters fully explicit. A The RE approach presupposes that economic agents novel feature of this situation is that the expectations have a great deal of knowledge about the economy. and forecast functions influence future data points. Even in the simple examples, in which RE are constant, As an illustration, consider again the cobweb model computing these constants requires the full knowledge (3). Assume that agents believe that the stochastic l of the structure of the model, the values of the process for the market price takes the form pt parameters and that the random shock is distributed constantjnoise, i.e., the same functional form as the iid. In empirical work economists, who postulate RE, RE solution. The sample mean is the standard way for do not themselves know the parameter values and estimating an unknown constant, and in this example must estimate them econometrically. A more plausible it is also the forecast for the price. Thus agents’ e l " t−" view of rationality might thus be that the agents also expectations are given by pt t i=! pi. Combining act like statisticians or econometricians when doing this with Eqn. (3) leads to a fully specified stochastic the required forecasting. This insight is the starting dynamic system. It can be shown that the system under point of the adaptiŠe learning approach to modeling learning converges to the RE solution if α ! 1. expectations formation. If the economic model incorporates exogenous or Taking this approach immediately raises the ques- lagged endogenous variables, it is natural for the tion of its relationship to RE. In many cases learning agents to estimate the parameters of the perceived can provide at least an asymptotic justification for the process for the relevant variables by means of least- RE hypothesis. For example, in the cobweb model (3), squares regressions. As an illustration suppose that an if agents estimate an unknown constant expected value observable exogenous variable wt−" is introduced into by computing the sample mean from past prices one the cobweb model, so that Eqn. (3) now takes the form can show that expectations will converge over time to l µjα ejδ jη the RE value. In more general models convergence to pt pt wt−" t (19) RE can occur if agents use the appropriate econo- metric functional form and run regressions in the same It would now be natural to forecast the price as a linear way that an econometrician might. function of the observable wt−". In fact, the unique RE Another major advantage of the learning approach solution is of this form: arises when there are multiple equilibria. Consider the l ` j ` jη above example with a monetary feedback rule in pt a bwt−" t (20) 5065 Expectations, Economics of where a- l (1kα)−"µ and b- l (1kα)−"δ and the cor- 6.3 Other Approaches to Learning responding rational forecast is p l a- jb- w . t t−" There are some other recent approaches to modeling Under least-squares learning agents would forecast expectation formation. ‘EductiŠe learning’ and according to ‘rational learning’ are closer to RE than adaptive pe l a jb w (21) learning. In the former agents engage in a process of t t−" t−" t−" reasoning using common-knowledge assumptions and the learning takes place in logical or notional time. where at−" and bt−" are parameter estimates obtained Rational learning takes place in real time, but retains by a least-squares regression of pt on wt−" and an intercept, using the data available through time tk1. the RE equilibrium assumptions at each point in time. In contrast, the adaptive learning approach assumes It can be shown that (at, bt) converges to the unique REE (a- , b- )ifα ! 1. that agents possess a form of bounded rationality which may, however, approach rational expectations over time. Besides the statistical formulation outlined 6.2 Stability Under Learning above, alternative models of bounded rationality learning have been studied. Concepts from compu- In general, when expectations are modeled by least- tational intelligence, such as genetic algorithms, neural squares learning there is convergence to the REE as networks and classifier systems, have been used as t !_ provided that a stability condition is met. The models of adaptive learning in economics. stability condition can usually be obtained by the Finally, we note that the literature has also con- expectational stability approach. sidered learning in situations with structural shifts To illustrate this approach for the cobweb model and\or misspecification of the perceived stochastic (19), suppose that expectations are based on the process. Certain learning rules, such as constant gain Š l j jη percei ed law of motion (PLM) pt a bwt−" t and algorithms, can track structural changes more ac- e l j hence given by pt a bwt−", where (a, b) may not be curately than least-squares algorithms, although they the REE values. Substituting this into Eqn. (19) we often fail to converge fully to an RE equilibrium if the obtain the corresponding actual law of motion (ALM): economic structure is in fact constant. Such pro- l µjα j δjα jη cedures can lead to new forms of persistent learning pt ( a) ( b)wt−" t (22) dynamics which may be empirically important. This yields a mapping from the PLM parameters (a, b) into the ALM parameters T(a, b) l (µjαa, δjαb). See also: Asset Pricing: Derivative Assets; Bounded Only at the REE values does one have T(a, b) l (a, b). Rationality; Keynes, John Maynard (1883–1946) Expectational stability (E-stability) looks at whether the REE is the stable outcome of a process in which the parameters of the PLM are adjusted slowly toward the parameters of the ALM that they induce. For- Bibliography mally, this adjustment is described by a differential Cooper R 1999 Coordination Games, Complementarities and equation and E-stability corresponds to local stability Macroeconomics. Cambridge University Press, Cambridge, of the REE under these dynamics. For details, see UK Evans and Honkapohja (2001). Evans G W, Honkapohja S 2001 Learning and Expectations in In the cobweb model the E-stability condition is Macroeconomics. Princeton University Press, Princeton, NJ given by α ! 1. Even in fairly complicated models it is Farmer R E 1999 The Economics of Self-fulfilling Prophesies, often straightforward to compute E-stability con- 2nd edn. MIT Press, Cambridge, MA Guesnerie R, Woodford M 1992 Endogenous fluctuations. In: ditions and these appear generally to govern con- Laffont J-J (ed.) AdŠances In Economic Theory: Sixth World vergence of statistical learning rules. Congress. Cambridge University Press, Cambridge, UK, Local stability under adaptive learning, or E- Vol. 2 stability, provides a selection criterion in models with Keynes J M 1936 The General Theory of Employment, Interest multiple RE equilibria. For example, in the Cagan and Money. Macmillan, London model with policy feedback (16) there are two RE Lucas R E Jr 1976 Econometric policy evaluation: A critique. In: solutions of the form (17), but only one of these is E- Lucas R E Jr (ed.) Studies in Business Cycle Theory. MIT stable and hence stable under least-squares learning. Press, Cambridge, MA (originally published in 1976) The other solution, although rational, is therefore Lucas R E Jr 1981 Studies in Business Cycle Theory. MIT Press, unlikely to arise in practice. Similarly, in the nonlinear Cambridge, MA l e Lucas R E Jr, Sargent T J (eds.) 1981 Rational Expectations and model yt F(yt+"), only the steady states with slope h - ! Econometric Practice. University of Minnesota Press, Minne- F (y) 1 are E-stable. This model can also have apolis, MN sunspot solutions and it has been shown that an Muth J F 1961 Rational expectations and the theory of price appropriate form of adaptive learning can converge to movements. Econometrica 29: 315–35 rational sunspot equilibria, provided the correspond- Pesaran M H 1987 The Limits to Rational Expectations. Black- ing E-stability condition is satisfied. well, Oxford, UK

5066 Experiential Psychotherapy

Sargent T J 1987 Macroeconomic Theory, 2nd edn. Academic both for health and sickness, and good and bad. What Press, New York is essential, however, is that people are seen as having innate worth, the ability to know the difference G. W. Evans and S. Honkapohja between good and evil, and the capacity to choose. Free will thus is the foundational principle and therapy involves facilitation of the actualization of the po- tential for healthy functioning. Health is seen as arising Experiential Psychotherapy from the ability to symbolize experience in awareness and to integrate different parts of the self to create 1. Introduction equilibrium and coherent meaning. Experiential theory of personality proposes a pro- The term experiential is defined in the Concise Oxford cess model of the self. The self is seen as a dynamic Dictionary (1990) as ‘involving or based on experi- experiential system that is in a continual process of ence’ and experience is defined as the ‘actual ob- self-organization. Experiential theory is constructivist servation of, or practical acquaintance with, facts or in nature, and adopts a dynamic and dialectical view events’ or ‘to feel or be affected by (an emotion, etc.).’ of functioning. It emphasizes that meaning is created A distinction between two ways of knowing, knowl- by human activity, in dialogue with others, that change edge by acquaintance and knowledge by description, is an inherent aspect of all systems and that this occurs first made by St. Augustine, and later emphasized in by a synthesis of parts. Meaning construction is viewed the epistemologies of William James and Bertrand as being constrained by a bodily felt emotional Russell, helps describe the essence of experiential experience but as influenced by language and culture therapy. Experiential therapy is an approach to and ultimately as created by a synthesis of experience therapy that focuses on promoting knowledge by and symbol (Gendlin 1962, Greenberg et al. 1993, acquaintance. Here a person does not come to know Greenberg and Pascual-Leone 1995). Emotional ex- something about him- or herself conceptually but perience is seen as both creating and being created by rather has an emotional experience of it. In experi- its conscious symbolization and expression. This view ential therapy the clients’ experiencing process is kept casts people as creators of the self they find themselves as a continuous point of reference for all therapist to be. In addition, experiential therapy, by adopting a responses and change is seen as occurring by the relational view of functioning, sees the self as coming promotion of new in-session experience. into existence at the boundary between inside and Although the term experiential therapy first arose to outside, between organism and environment, by a refer to Whitaker and Malone’s approaches to psycho- synthesis of bodily experience, symbol and inter- dynamically oriented family and individual therapy personal validation. (Whitaker and Malone 1953), this approach has Emotional experience, although seen as a basically subsequently grown out of a variety of ‘third force,’ healthy resource, is viewed as capable of either humanistic, approaches. These include client-centered providing healthy adaptive information based on its (Rogers 1959), Gestalt (Perls et al. 1951), and ex- biologically adaptive origins or, in certain instances, as istential therapy (May and Yalom 1995), as well as having become maladaptive through learning and other approaches that emphasize working with the experience. The most basic process for the individual clients in-therapy experiencing process (Greenberg et in therapy is thus one of developing awareness of al. 1998). Experiential therapy consists of those ap- emotion and discriminating which emotional res- proaches that, within the context of a facilitative ponses are healthy and can be used as a guide, and human relationship, focus on the creation of new which are maladaptive, and need to be changed. meaning by the symbolization of experience in aware- ness. The two major foci of experiential therapy practice are the personal presence of the therapist in 2.1 Goals the provision of an emphatic and facilitatively genuine The general goals of treatment are to promote more therapeutic relationship, and increasing clients’ aware- fluid and integrative self-organizations. Therapy ness of internal experience by focusing them on their focuses on the whole person, i.e. it is person centered moment by moment subjective experience. rather than problem or symptom focused, but within this wholistic focus, the underlying determinants of 2. View of Human Nature and Functioning different types of self-dysfunction are also an im- portant point of focus. Thus both a change in manner Experiential therapy adopts an existential view of of functioning of the whole self and changes in human nature, one that sees existence as preceding particular problems in self-organization are viewed as essence. People are seen as active agents in the important. For example, a client may be seen as construction of their own realities and choice is seen as changing the manner of functioning both by becoming the final arbiter in human functioning. Individuals are more empathic to the self and by being able to seen as being born morally neutral with a penchant symbolize bodily felt experience, as well as changing

5067

Copyright # 2001 Elsevier Science Ltd. All rights reserved. International Encyclopedia of the Social & Behavioral Sciences ISBN: 0-08-043076-7