Schettkat & Jovičić (2017) sta pripravila čudovit paper, v katerem sta pokazala, da predpostavke, na katerih temelji znamenita “Lucasova kritika”, povsem napačne. Zelo na kratko, Lucas & Sargent (1978) sta argumentirala, da so ekspanzivne makroekonomske politike neučinkovite, saj so ekonomski agenti sicer lahko kratkoročno zavedeni z dvigom posameznih nominalnih kategorij (cene, plače), vendar hitro opazijo, da se dvigne tudi splošna raven cen (inflacija), zaradi česar se realne kategorije ne spremenijo. Output se zato vrne na prvotno raven. Vse, kar so ekspanzivne makroekonomske politike povzročile, je zgolj višja inflacija – kar prikazujemo z vertikalno Phillipsovo krivuljo na ravni “naravne” (strukturne) stopnje brezposelnosti. In še naprej, ker imajo agenti racionalna pričakovanja, “gredo na led” samo enkrat, samo enkrat reagirajo na spremembe v ekonomski politiki, naslednjič pa že vedo, da s tem ne dosežejo realnih sprememb, pač pa le višjo inflacijo. Makroekonomske politike so zato povsem neučinkovite.
Ta Lucasova kritika, temelječa na mikrofundiranem ekonomskem modelu, je zrevolucionirala in za dobra tri desetletja dominirala v makroekonomiji in povzročila številne negativne učinke. Kot je denimo politika varčevanja in strukturnih reform namesto ekspanzivnih makroekonomskih politik v času povpraševalnega šoka. No, Schettkat & Jovičićeva sta pokazala, da so te predpostavke Lucasa & Sargenta absolutno napačne, saj čeprav govorita o učenju racionalnih agentov, tega zares ne upoštevata. Agente zgolj premikata iz enega stacionarnega stanja v drugega (iz enega dolgega roka v drugega) v “logičnem času”, zanemarjata pa vse vmesne faze (vse kratkoročne učinke) v “realnem času”. Schettkat & Jovičićeva pokažeta, da če v Lucas-Sargentovem modelu zamenjata zgolj eno zadevo – “logični čas” zamenjata z “realnim časom” in tako dopustita dejansko kratkoročno učenje agentov, se Lucasova kritika sesuje. Namreč, kar agenti izkusijo na kratek rok zaradi denimo ekspanzivne fiskalne ali monetarne politike, je, da so zaradi njiju dobili službo in dohodek, ki ga lahko realno potrošijo. Zanje makroekonomske politike niso neučinkovite, ampak imajo realne pozitivne učinke, zato pozitivno (z realnimi učinki) reagirajo nanje. Vertikalna Phillipsova krivulja kot simbol neučinkovitosti makroekonomskih politik ne obstaja.
In še nekaj, o čemer Schettkat & Jovičićeva ne govorita: ko gospodarstvo doleti negativni povpraševalni šok (negativen output gap), gospodarstvo po definiciji ni več na ravni optimalnega (strukturnega) outputa, pač pa precej nad strukturno brezposelnostjo (in pod strukturno ravnijo outputa), zato imajo makroekonomske politike, ki vplivajo na spodbujanje agregatnega povpraševanja, zelo velike realne učinke na povečanje outputa in zmanjšanje brezposelnosti. Prav zato je odsvetovanje ekspanzivnih makroekonomskih politik v času povpraševalnih šokov in njihovo zamenjevanje s strukturnimi reformami (ki imajo vpliv le na potencialni output) takšna bedarija.
“The ideas of economists and political philosophers, both when they are right and when they are wrong, are more powerful than is commonly understood. Indeed, the world is ruled by little else.”
The rational expectations hypothesis (REH) which led to the policy ineffectiveness hypothesis (PIH) in the new classical model is a tragic example of Keynes’ statement. The REH has shaped common views on macroeconomic policies classifying unemployment as structural, which thus cannot be pushed below its ‘natural’ level by expansionary macroeconomic policy, fiscal or monetary. All expansionary macroeconomic policy can achieve is a higher price level. These are the major conclusions of the PIH, widely accepted among central bankers, politicians, and many economists. Unfortunately, the PIH did not remain in academics but rather guided and still guides economic policies emphasizing structural reform and austerity in Europe with tragic, long-lasting negative effects on growth prospects.
That the predictions of Keynesian theory “… were wildly incorrect, and that the doctrine on which they were based is fundamentally flawed, are now simple matters of fact, involving no novelties in economic theory,” wrote Lucas and Sargent (1978: 49); they argued, “For policy, the central fact is that Keynesian policy recommendations have no sounder basis, in a scientific sense, than recommendations of non-Keynesian economists, or for that matter, noneconomists” (Lucas/Sargent 1978:57). The core of the so-called “Lucas critique” is changing coefficients in a macroeconomic model because economic agents do not simply repeat past responses to economic policies but they learn from experience, correct their expectations accordingly, and avoid systematic errors (the REH). Expansionary macroeconomic policy measures, unless unexpected, do not affect the real economy (Sargent/Wallace 1975).
Certainly, no Keynesian economist denies that future-oriented decisions are based on expectations, but Lucas and Sargent assume a stochastic version of perfect foresight (Arrow 1986: 316) where the economy settles at a specific equilibrium; that is, the economy is assumed to be stationary, an ergodic system (see Davidson 1982). According to the PIH, only when confronted with an unexpected, surprising policy shock, when information is imperfect, can macroeconomic policies show real effects. However, experience will perfect the knowledge; that is, economic agents are learning, and they will consequently not react to expansionary macroeconomic policy in the future once they experience the inflationary cycle laid out in the writings of the new macroeconomics. Rising prices is all that expansionary macroeconomic policy can achieve; it pushes the economy out of the general equilibrium but fails to stimulate production.
When experiencing expansionary macroeconomic policy the first time, economic agents lack knowledge about the economic process – they are caught by surprise – but experience will lead them to discover the ‘true’ model – learning – and henceforth expansionary macroeconomic policy will be ineffective unless unexpected. Unexpected expansionary macroeconomic policy is treated like an exogenous shock, hitting the economy in optimum (i.e., pushing the economy out of equilibrium) because economic agents – according to the argumentation – misinterpreted nominal variables as real. They confused general price increases with relative price rises for their good, igniting adjustments to reach the mistakenly perceived new optimum. However, they will learn that they confused nominal and real and the economy returned to the initial position, the optimum, at a higher price level.
Although extremely influential in economics and economic policy, we show that the Lucas/Sargent analysis is fundamentally flawed for several reasons:
(1) The Phillips curve, the trade-off between inflation and unemployment, is taken as the structural form of the Keynesian model. Actually, rightward shifts of the demand function cause higher output, higher employment, and lower unemployment in the Keynesian model, with the probable side effect of some inflation. Thus, in the Keynesian model, the causation runs from rising demand not from inflation to quantities, which may occur if the capacity limits are reached.
(2) Furthermore, in Lucas/Sargent, only suppliers (of products or labor) are misinterpreting a general price rise, but firms experience prices in intermediate products and workers observe prices as consumers. Agents do not notice price changes for the products they buy but rather interpret a general rise in prices as specific for the goods they supply?
(3) Applying comparative statics – applying isolated equilibrium analysis for different situations in ‘logical time’ (Robinson 1980) – is inadequate for analyzing the adjustment process and misleading even if the underlying microeconomic assumptions are accepted. As Robinson (1974, 1980) clearly explained, markets converging4 to equilibrium requires time, a dimension missing in the quantity-price diagram5 and only superficially used in the Lucas/Sargent analysis. Moving ahead in time, one ‘short run’ follows the ‘next short run’ (Kalecki 1968). Responses to economic incentives, even if based on misperceptions, affect the next short run and may shift the economic potential in the “short runs” to come.
(4) Lucas/Sargent claims only one economics – not micro and macro – which needs to be based on ‘first principles’ (equilibrium and optimization), the so-called micro foundations, but their analysis treats micro units superficially. They do not investigate the consequences of actions at the micro level; they ignore output and income effects. Micro-foundations are simply the assumptions of equilibrium and maximization.
(5) Using ‘learning’ as a metaphor for discovering an assumed stationary model and ignoring the adjustment process is insufficient and misguiding; it is, in this case, evidentially flawed. Lucas and Sargent assume that economic agents learn that the economy converges to the ‘long-run’ solution of their specific model. Actually, agents learn from their experience, which may well be that expansionary macroeconomic policy improves the real economy, that the potential increases, and that their incomes rise permanently. Does the economic adjustment process when the economy is ‘out of equilibrium’ affect the outcome? Yes, it leads to completely different conclusions.
Keynes (1936) showed in chapter 19 of the General Theory of Employment, Interest, and Money that assuming perfect markets but a decline in real wages brought about by a reduction in nominal wages is inconsistent. Similarly, we argue that Lucas and Sargent, given their own neoclassical assumptions, are inconsistent. In this paper, we show that if economic agents behave as they argued, they will not discover their static model but rather they will learn the virtues of a dynamic economy. Expansionary macroeconomic policy will be perceived as stimulating, pushing the economy to a higher potential, and it will be perceived as a virtue rather than a mistake.
Thus, economic agents are assumed to have learned (or to know) the right model ‒ which is supposed to be the new classical model ‒ and they know (or have learned) the values of the
parameters reproducing the true relationships. According to Lucas and Sargent (1978), this reasoning is why rising prices and high unemployment (stagflation) are observed in the 1970s. Even if the analysis is based on neoclassical micro-foundations, Lucas and Sargent’s conclusion is not as clear-cut as the analysis with comparative statistics suggests. Although new classical macroeconomics ostensibly provides consistent microeconomic foundations, the foundations are actually inconsistent with the neutrality of macroeconomic policy in historical time. Only rational expectations ‒ the shortcut to the long-run neoclassical solution ‒ solves the problem but assumes away the intermediate effects on real variables; it assumes away frictions in the adjustment process. Continuing the analysis as if the period of ‘mistakes’ had not occurred, as it did not have any effect on the real economy (other than price changes), however, is misleading.
If workers take jobs, even if misled by nominal wage rises as a result of expansionary policy, what do they do in the period when they are misled? If employers hire additional workers, what do they have them producing? Unless productivity drops substantially when additional workers enter the workforce, there will be additional (real) output. Although prices may rise, additional output will dampen inflation. In addition, what happens to unearned income ‒ capital or transfers ‒ does it rise with nominal wages?
In this paper, we take the process, experience, and learning seriously, (i.e., we investigate what economic agents actually experience and therefore learn about the effectiveness of expansionary macroeconomic policy in ‘historical time’). ‘Historical time’ is the only deviation from the neoclassical assumptions claimed to be essential for rigorous economic analysis by Lucas and other proponents of the new classical macroeconomics.
We must conclude that even if neoclassical micro-foundations are rigorously applied but learning, the process in historical time, is taken seriously, the predicted ineffectiveness of expansionary macroeconomic policies is a false conclusion. If the economy behaves as assumed in the Lucas/Sargent model and economic agents are learning, expansionary macroeconomic policy is perceived as a virtue. The Lucas/Sargent analysis is flawed because it uses learning as a metaphor but ignores the consequences of the postulated behavior. Sadly enough, the widely application of the Lucas model in economic policy has been extremely costly.
Vir: Schettkat & Jovičić (2017), Macroeconomic revolution on shaky grounds: Lucas/Sargent critique’s inherent contradictions