Blog o’ Burns

Attfield, C.L.F., D. Demery and N.W. Duck. Rational Expectations in Macroeconomics. New York: Basil Blackwell Ltd, 1985.

Note: I focused my summary on the first three chapters of this book. I will edit my post sometime this week to include information from Chapter 5, which discusses rational expectations in an open economy.

CHAPTER 1: EXPECTATIONS IN MACROECONOMICS

Attfield et al (1985) begin their analysis of rational expectations with three examples in economic theory in which expectations of one variable or another play a vital role: 1) the IS-LM model; 2) the consumption function; and 3) the natural rate hypothesis.

1) In the Keynesian IS-LM model, the IS curve is volatile due to the instability of the expected profitability of investments, causing it to shift by large amounts as firms change the amount of investment they wish to undertake. However, the frequent shifts in the IS curve would not have much impact if the LM curve were steep, but quite the opposite, it is likely to be horizontal. Why is that? Demand for money depends on expectations of future interest rates. At the critical rate of interest, the interest rate reaches a point so low that people’s demand for money will be very responsive to shifts in it, thereby causing the LM curve to appear flat. If people expect a drop in interest rates, they will sell their bonds and their demand for money will go up.

2) One post-Keynes development in the consumption function has been the concept of permanent income, indicating that people consume based on expectations of future income rather than just current income. This has significant implications on government policy, because if policymakers attempt to stimulate the economy by increasing expenditure, the multiplier may not be as strong if people do not expect their future income to change.

3) In the natural rate hypothesis, it is assumed that workers are primarily concerned with real wages and will form expectations with respect to inflation when determining whether to supply more labor. If aggregate spending rises, prices will rise, and employers will offer higher nominal wages to attract more labor. Workers will see that higher nominal wages do not translate into higher real wages and will stop supplying labor. In this case, an increase in aggregate spending would have little effect on employment. However, the opposite would occur if workers did not expect prices to rise.

As demonstrated in the three examples above, incorporating expectations into economic models has changed how economists view potential outcomes. However, little data exists to measure this phenomenon. Attfield, Demery, and Duck ask whether it is possible to have a general theory explaining how expectations are formed that can be easily tested and can demonstrate the relationships proposed in macroeconomic theory. The theory of rational expectations attempts to serve such a purpose.

CHAPTER 2: THE THEORY OF RATIONAL EXPECTATIONS

The theory of rational expectations assumes that people base their expectations on all available information, which is used to formulate a “process” (often represented by an algebraic equation) that reflects the pattern of a variable’s behavior. If the variable follows the exact pattern described by the process, people’s expectations will be perfectly accurate with zero forecasting error. However this would most likely not be the case since most processes are not completely deterministic — most have a stochastic (or random) element. The best expectation a rational person can make is that this error term will equal its mean of zero.

One criticism of the theory is the plausibility that people actually use complex processes when forming their expectations. Many people are unaware that certain economic variables exist let alone understand how they may determine the value of other variables. In addition, it may not be efficient to try to obtain all of the information necessary to form accurate forecasts. A rational person will only acquire information as long as the marginal return from doing so exceeds the marginal cost. Attfield et al argue that though people probably do not consciously calculate the expected values of variables, collectively they act as if they do. Futhermore it is costless to recognize when your expectations turn out to be wrong, so it is costless to use this information until the marginal return from doing so equals zero. A second criticism is in the fact that the actual values of determinant variables may not be known. In such a case, the authors argue, people will form rational expectations with respect to those values as well. As a final criticism, there may be limits to the applicability of the theory of rational expections. On occasion, events occur that appear unique or unusal. The authors believe that such events are often a function of some greater underlying pattern that has not been recognized yet. People are capable of making intelligent appraisals of current circumstances to for expectations.

CHAPTER 3: RATIONAL EXPECTATIONS AND A FLEXIBLE PRICE MACROECONOMIC MODEL

The authors derive the following model under the assumption that prices move freely to equate supply and demand.

To review a bit, to derive the aggregate demand curve, we must look at equilibrium aggregate demand determined by the IS and LM curves at different price levels. The IS curve shows combinations of real income and nominal interest rates for which aggegate demand for output equals actual output. The LM curve shows combinations of real income and nominal interest rates for which the quatity of money dmanded equals the quantity supplies. Where the two curves intersect indicates the equilibrium interest rate and level of aggregate demand. From this diagram, a downward-sloping aggregate demand curve can be derived.

Now let us assume that suppliers are primarily influenced by relative prices. If the relative price of goods in one region is higher than in another, suppliers in that region will respond to the higher prices by supplying more output. A higher relative price in one region translates into a lower relative price in another, causing that supplier to decrease its output. Thus, aggegate supply stays the same with an adjustment in relative prices. However, what if suppliers are unsure as to the value of relative prices? They will have to form rational expectations of the average, economy-wide price of the good and compare this expectation with the actual prices in their own market. If the exonomy-wide expectation of relative prices turns out to be above actual relative prices, aggregate output will be above its natural rate. The natural rate reflects the relationship between output and price level if the price level is accurately forecasted. The primary implication of this hypothesis is that changes in the price level which are expected have no effect on real output.

Now let us assume that aggregate demand shifts upward, but people expect price levels to remain the same. Suppliers will see the unexpected rise in their prices and assume that a shift in relative demand for their goods has occurred. Responding to these higher prices, they will supply more output. However, if people are rational, why couldn’t they predict the shift in aggregate demand? And if they could predict that it would rise, why didn’t they forecast the rise in average price level? At the heart of these questions is how rational people handle the “signal extraction problem.” A rational supplier wants to know the precise impacts of changes in aggregate demand and relative demand on his/her prices. However, he/she can only see the combined influence of these two factors on his/her prices. Thus, expectations will vary somewhere between the old price level and the actual price level.

The key conclusion to draw from the theory of rational expectations is the fact that only random shifts in aggregate demand will affect real output. Predictable shifts will only affect prices.