Expectations

Adaptive Expectations Hypothesis: People rely on past experience and information to form expectations. They only gradually modify their expectations.

Future expectations are the same as current observations. So if todayís inflation rate is 2.9% with adaptive expectations agents think that next periodís inflation rate will be 2.9%.
 

Inflation Rate
Time	Actual	Expected	Unanticipated
0	0%	0%	0%
1	2%	0%	+2%
2	4%	2%	+2%
3	4%	4%	0%
4	2%	4%	-2%
5	1%	2%	-1%

in period 1, actual inflation exceeds expected inflation. Thus the unanticipated inflation rate is positive.

If the inflation rate is rising, the public systematically underpredicts what the inflation rate will be next period.

If the inflation rate is falling, then the public systematically overpredicts the rate of inflation for next period.

Thus, with adaptive expectations the public can be systematically fooled if inflation is accelerating or decelerating.

Up to and during the 1960ís Phillips had noticed that there existed a relationship between inflation and unemployment going all the way back to the 1800ís. When inflation rose, unemployment fell. This observation is consistent with the view that business cycles are due primarily to demand side shocks. Show this with the AS/AD framework on your own. In the 1968 presidential campaign Hubert Humphrey (his son just lost his bid to be Governor of Minnesota to Jesse [formerly known as "the body"] the mind Ventura) and Musky argued that the Phillips curve represented a menu of options and that we should take advantage of the Phillips curve to get employment below the natural rate permanently. If there existed a stable trade off between inflation and unemployment, then as a society we could choose to have, say, 6% inflation per year if it meant that unemployment fell from, say, 4.5% to 3%. The problem is, as we showed above, that even with adaptive expectations the inflation rate would continually rise year after year.
 
 

• if inflation rises to 4% people expect p =2%. This is a movement along SRPC₀ from point 0 to point 1. Unemployment falls to U₁<U_N.
• In the next period agents adjust their price expectations and expect inflation equal to 4%, but the Fed can increase inflation to 6% (via monetary expansion) and move along the SRPC₁ which would keep the unemployment at U₁<U_N.
• the Same process continues into period 3.

The idea is that if agents have adaptive expectations and respond to rising inflation with a lag the policy makers could use monetary policy to permanently reduce the unemployment rate below the natural rate. There is a difficulty with this argument beyond the rational expectations critique that we shall develop more below. The problem is that if one follows this argument out it would imply an ever rising inflation rate and we know from historical experience (Germany 1920ís, Latin America, Hungary) the immense destruction to which ever rising inflation can lead (e.g. barter)

For this idea to make any sense one must assume that the inflation that occurs is predictable. As we discussed before, when we asked the question, "Why is inflation bad?" We answered it is bad if it is unpredictable and if it is so high that the currency is unable to transport purchasing power across time, i.e., currency becomes poor store of value. If inflation were perfectly predictable then it would be no problem if the inflation rate rose 2% a year (assuming that inflation is low). We could include this fact in all contracts that are written in the economy. There would be no damage done to those who lend. So predictability here is the key.

Question: What would a sensible person expect to happen in period 4? Would the person expect inflation to be 8% or more like 10%? If you are a wage negotiator who forms expectations adpatively, youíll systematically lose out.

Rational Expecttions Hypothesis: Expectations are based on all available information, not just past information but also predictions about current and future policy.

Gathering information about the economy is costly. People will not have complete knowledge, but people will use the information they do have including predictions about the future. People may make mistakes in predictions but will adjust expectations to policy changes. People will not make systematic mistakes. Errors tend to be random.

• If the Fed by 2% (to 4%), then people adjust expectations if Fed policy is systematic and could be predicted (shifted SRPC₀ to SRPC₁ with U unchanged as U_N).
• As the Fed 

Any systematic Fed policy will not be able to change U if people have rational expectations. In the example above people did notmake sytematic mistakes.

Now, suppose that Fed policy is random:

• on average Fed allows , but randomly increases and decreases MS.
• When people guess wrong (p^e=6%, but p_actual=7%), then you can have real effect of U_N falling to U_b, but to get this the Fed must have a random policy.

• The Fed 
• Since expectations are adaptive, the shift in AD is along a stable SRAS₀. Remember that for SRAS₀ agents expect the price level to be p₀ which is where SRAS₀ intersects LRAS.
• With adaptive expectations, agents update their price expectations with a lag. Next period, agents adjust their expectations and the SRAS₀ shifts to SRAS₁. Now the agents expect the price level to be p₁ which was last periodís average price level. Note that SRAS₁ intersects LRAS at p₁. Also, the Fed once again 

• Given adaptive expctations agents can be systematically fooled into underestimating the price level which means that output can be permanently maintained beyond the potential level of output.

Rational Expectations Applied to AS-AD

Consider systematic increases in MS

1. Fed announces an increase in the MS. People can fully anticipate the changes in the price level. The increase in the MS increases AD from AD₀ to AD₁. The change in price expectations shifts AS₀ to AS₁ simultaneously. Output is unchanged at Q_p.
1. Same thing happens in the next period. The Fed once again announces that it will increase the MS. Agents anticipate the change in the price level. The resulting simulataneous shifts in AD and AS offset one another, leaving output at Q_p and yielding a higher price level at p_2.

• On average the policy results in P=100
• if the Fed randomly changes MS then AD randomly shifts
• since the best prediction by economic agents is p=100 if the Fed increases the MS then people will underestimate the price level and real output will rise above the potential level to Q₂.
• If the Fed decreases the money supply agents will overestimate and real output will fall below the potential level to Q₁.

This is really an argument against permanently lowering output below the potential level of output and, equivalently, permanently lowering the rate of unemployment below the natural rate of unemployment. But, note that the force of the argument does not address the issue of reducing unemployment and raising output when the economy is performing below the natural rate of unemployment and below potential output. If you believe that expectations take time to adjust, then when there are slack resources the Fed can use expansionary monetary policy to try to increase AD and move the economy back towards Q_P and L_N.

Evidence in Favor of Rational Expectations

End of hyperinflations

• if adaptive expectations, then a drop in the money supply would result in a recession. Output would fall below the potential level of output and the price level would not rise by as much as it would have risen otherwise.
• if rational expectations, then a well anounced, credible drop in the money supply would lead to a lower price level but no recession.