Question

Using one of the schools of thought we’ve discussed in class (Keynesian/ New Keynesians, Monetarism, New Classical, RBCT, ABCT) discuss the merits of rule-based vs. discretionary fiscal and monetary policy with regard to stabilizing the macroeconomy. Be sure to include the assumptions made by the school of thought you use in your answer.

Answer 1

Since the 1930’s expectations (anticipation’s or views about the future) have played an important role in economic theory. This is because economics is generally concerned with the implications of current actions for the future. In recent times attention has switched from more or less mechanical forms of expectations generation (extrapolative or adaptive) which has become essentially ad-hoc to the theoretically attractive approach of the rational expectations hypothesis. This states that agents use economic theory to form their expectations, and should not make systematic errors in their forecast of the future.

The reason for this switching however, is not hard to find. It derives partly from the sad state in which macroeconomic theory found itself in the early 1970’s, with the phenomenon of stagflation[1] confounding earlier Keynesian optimism and with the Philips curve apparently experiencing increasing instability and collapse. It also relates to the fact that the adaptive expectations model associated closely with the name, Cagan (1956) became increasingly untenable as a model of expectations formation under conditions of accelerating inflation which typified the 1970’s.

THE NEW CLASSICAL APPROACH

The new Classical approach to explaining business cycles in relation to the role of expectations has something in common with monetarist, in that the shock sets off the cycle is a change in the money supply.

New Classical economists assume that the actors in the private sector of the economy have rational expectations. This assumes that agents form expectations based upon all available information about the future at the time they take the decision. Therefore, agents make only random errors in forecasting the future course of economic variables. This means that the expectational errors that trigger cycles cannot be systematic. If they were systematic, agents could learn from the pattern of mistakes and improve their forecasts.

According to the New Classical approach, only unanticipated policy changes lead to changes in real national income. Systematic policy changes will be predictable and will have no real effects. Most economists do not accept the proposition that only unexpected policy changes will have real effects. One reason is that there a lot of fluctuations in price and wage setting behaviour that very few contracts can be renegotiated as soon as a policy change in relation to interest rates is announced. Hence the policy –makers certainly have some leverage over real activity, even when making policy changes that are predictable.

Another reason is that the massive complexity of the economy makes it impossible for individual agents to know how some change in policy will affect all the relevant prices and quantities that matter to them over any specified period of time. This is where the presumption that private agents have expectations of what policy makers are going to do, and that this influences private behaviour, is the subject matter here. Without, assuming randomly, just reasonable approximate expectations, private anticipation of government action can affect the outcome of policies.

THE MONETARISTS

Monetarism is closely allied with the Classical school of thought. It is essentially an extension of Classical theory which was developed in the 1960s and 1970s to try to explain a new economic phenomenon, stagflation.

This sees expectations as determined by essentially unexplainable psychological forces. It states that inflation is determined by excess aggregate demand and price expectations; that expectations are generated by past price history and hence by previous excess demand; that excess demand results from excessive monetary growth; and therefore that excessive monetary growth, past and present is the root cause of inflation.

According to the rational expectations hypothesis, individuals will tend to exploit all pertinent information about the inflationary process when making their price forecasts. If true, this means that forecasting errors ultimately could arise only from random shocks occurring to the economy. For if the public is truly rational, it will quickly learn from these inflationary surprises and incorporate the new information into its forecasting procedures. As incorporated in monetarist models, the rational expectations will always be correct and the economy will always be at its long-run steady-state equilibrium.

Monetary advocates of the strict rational expectations view argue that it carries some radical implications for stabilization policy. Specifically it implies that systematic policy actions cannot influence real variables even in the short run, since rational agents would already have anticipated and acted upon these policies. To have an impact on output and employment authorities must be able to create a divergence between actual and expected inflation. This follows from the monetarist view that inflation influences real variables only when it is unanticipated.

The authorities must be able to alter the actual rate of inflation without simultaneously causing an identical change in the expected future rate. Thus, the only way that monetary policy can have even a short-term influence on real variables is for it to be completely unexpected. However, with stagflation the monetarist view becomes questionable as stated in the latter.

It is argued that using empirical data to test the validity of the rational expectations hypothesis, two difficulties are immediately encountered. Firstly, much of the evidence for rational expectations is sought in macroeconomic models which incorporate other assumptions, particularly price clearing scenarios. Negative findings concerning such models do not therefore invalidate rational expectations perse. Secondly, there is the problem of observational equivalence by which we mean that for any rational expectations model which fits the data there will always be non-rational expectations model which fits the data equally well. It is in recognition of these difficulties, that various approaches have been adopted in carrying out empirical test of this theory (Shaw, 1987).

Barro (1977) tested the rational expectations hypothesis. He attempted to show that it is only the unanticipated component of monetary growth that affects employment, real output, and the price level. He used annual data from the USA covering the period 1941 to 1973. In accordance with certain theoretical considerations and after some empirical experimentation, he obtained a measure of anticipated monetary growth. He then computed the unanticipated component of monetary growth in each period as the difference between actual monetary growth in the period and the anticipated component of monetary growth in that period. His statistical tests all seem to support one of the main predictions made by the simple rational expectations model: that it is unpredictable monetary growth that is important in the determination of the level of unemployment and that predictable monetary growth is irrelevant.

THE NEW KEYNESIAN

Keynesian theories of expectations assume that expectations are slow to change. The theory of extrapolative expectations says that expectations depend on extrapolations of past behaviour and respond only slowly to what is currently happening to costs. In one simple form of the theory, the expected future inflation rate is merely a moving average of past actual rates.

The rationale is that, unless a deviation from past trends persists, firms and workers will dismiss the deviation as transitory. They will not let it influence their wage and price setting behaviour.

The theory of adaptive expectations in this context states that the expectation of future inflation rates adjusts to the error in predicting the current rate. Thus, if you thought the current rate was going to be 5% and it turned out to be 10% you might revise your estimate of the next period’s inflation rate upwards by, say, half of your error, making the new expectation 7.5%.

The two theories make expectations about future inflation depend on past actual rates. In an obvious sense such expectations are backward-looking, since the expectation can be calculated using data on what has happened already. Backward-looking expectations are overly naïve. People do look ahead to the future and assess future possibilities rather than just blindly reacting to what has happened before.

In modelling expectations of a variable under the Keynesian theory, the simplest assumption is that the expected rate of change of the variable over the next time period will be the same as the change which has occurred over the previous period, so that,

EtX t + 1= Xt(1.1)

Where

X=Price

EtX t + 1=expected rate of change of X from period t to t + 1

Xt=actual rate of change of the X from t – 1 to t

This was used by among others, Turnovsky (1972). A slightly more general model is provided regressive or extrapolative expectations hypothesis:

EtX t + 1=Xt + O(Xt – X t – 1)(1.2)

Now if the parameter O is 0, equation (1.1) is obtained equation (1.2) can also be rearranged to give:

Et Xt + I=(I + 0) Xt – 0 Xt – I(1.3)

Where the expectation is a weighted average of the two most recent actual values used. This can be regarded as a particular case of

Et Xt + I=bo Xt + b1 Xt – 1 + b2 Xt – 2 + …………(n) (1.4)

Where the expectation is determined by the current and all past actual values. A common restriction for equation (1.4) and which has theoretical attractions, is to assume

Bi=(I – 1) 1i 0 < 1 < I(1.5)

Which gives the geometric distributed Lag or the Koyck Lag. Substitution of (1.5) into (1.4) and by lagging the resulting equation one period we obtain:

Et Xt + I – Et – I Xt = I – 1 (Xt – Et – I Xt)(1.6)

This was used by Cagan (1956), who evaluates the right hand side for different values of 1.

In equation (1.6), the current expectation is a weighted average of the previous expectation and the current actual rate of X. Alternatively, (1.6), which is the version commonly known as the adaptive expectations model or the error learning mechanism expresses the change in the expectation as an adjustment depending on the error between the actual rate of X from t – 1 to t and the expectations for that period.

Carlson and Parkin (1975) modified equation (1.6) by the addition of a of a second error term. Frenkel (1975) suggested a model which combines both regressive and adaptive components. These variations require the appropriate adjustment coefficients to be constant.

A criticism which applies to these models id that information other than past actual value of X and past expectations is ignored. Therefore, a much wider information set will thus be relevant in determining current expectations.

DEMAND MANAGEMENT POLICIES

In relation to monetarist policies effect on demand, it is argued that a rough correlation between changes in the money supply and changes in the level of economic activity is accepted by most economists. But there is controversy over how this correlation is to be interpreted. Do changes in money supply cause changes in the level of aggregate demand, and hence of business activity or vice-versa? Friedman and Schwartz (1963) argue that changes in the money supply cause changes in business activity. There argument was based on the Great Depression of the 1930’s, where a major contraction in the money supply, shifted the aggregate demand curve far to the left. This led the monetarists to advocate a policy of stabilizing the growth of money supply.

Keynesians believe that both monetary and non-monetary forces are important in explaining cycles. Although they accept serious monetary mismanagement as one potential source of economic fluctuations, they do not believe that it is the only, or even the major, source of such fluctuations.

Thus they deny the monetary interpretation of business cycle history given by monetarists. They believe that most fluctuations in the aggregate demand curve are due to variations in the desire to spend on the part of the private sector and are not induced by government policy.

The New Classical School suggests that if private agents are watching the government and trying to form expectations of its future behaviour, not only does it matter what the government does, but it also matters what agents think it will do in the future. This means that a government needs more than just correct current policies. It also needs to establish credibility that it will follow the correct policies in the future.

The New Classical approach gives no role to aggregate demand in influencing business cycles; it provides no role for stabilization operating through monetary and fiscal policies. Indeed, the models used by this school to date predict that the use of such demand management policies might prove to be harmful to an economy as a whole.

SUMMARY AND CONCLUSION

This paper has looked at the three different economic schools of thought namely the New Classical, the monetarist and the Keynesian approach with regard to their view on expectation formation. Finally, the effect of the three different economic schools of thought on demand management policies have been covered.

It will be worthwhile to conclude that the development of the role of expectations formation will make a more significant contribution to economic theory as a whole, than prejudged assumptions which has been the critical aspect of most research carried out in the past with regard to the three schools of thought. This relates in particular to the New Classical approach in which demand management policies is deemed as irrelevant.