This paper is based on a long-term research program with Rachel Kranton on the implications of identity

37

reality.  One common amendment to standard economic method yields easy explanation of all

these anomalies.  Employees may care about nominal aspects of wages, not just about their

wages in real terms.  In each and every case they have a notion of what their nominal, and not

just their real wage, should be.  In terms of utility maximization, employees lose utility insofar as

there is a gap between what they think their nominal wages should be and their actual value. 

Adding such concerns to workers’ utility functions will explain each of the six anomalies.  In

some cases it is difficult to see how the anomaly could possibly be explained without such

addition.

Let’s first consider nominal wage rigidity.  A belief that workers should not take wage

cuts is sufficient to explain downward nominal wage rigidity.   This would account for the

asymmetric agglomeration of wage cuts precisely at zero.  If there is another possible

explanation for this finding, it is obscure.  In agreement with common sense, Bewley’s findings

say it is the workers (rather than the firms) with such beliefs.  Many different informants told

Bewley in many different ways that firms do not reduce wages in recessions when labor is easily

available and also cheaper because they are afraid of alienating their own workers by reducing

their money wages.  We also saw that the absence of accelerating decline of wages and prices in

the Great Depression could be easily explained by such nominal considerations in worker utility

functions.  

Dislike of money wage cuts, however, may be just the tip of the iceberg.  It is, of course,  

the form of nominal illusion that is clearest in the statistics.  But if there is one way in which

nominal wages enter utility functions, because employees have a notion of what their employer

should or should not do, there could also be many other ways. 

---

52

A variant on the original question would involve changing the numbers.  How much more than 1%

difference in the real wage would be necessary to make Ann happier than Barbara?

38

The next set of findings—from the questionnaires—takes us back to Barbara and Ann. 

There is an easy way to account theoretically for the findings that respondents thought that

Barbara would be happier than Ann, even while acknowledging that she is worse off than

Barbara economically: Barbara and Ann both have a nominal component in their utility function. 

They think that they should get a nominal raise.  With such a utility function, insofar as Barbara

gets a nominal raise and Ann gets none, Barbara may indeed be happier.

52

  Furthermore this

interpretation of the behavior of the utility function is exactly consistent with Shiller’s finding. 

If people think that they should receive a nominal raise and they do receive it, then— just as they

say—they will be happier if their wages go up and prices go up as much. 

The same can be said about the anomalies regarding indexation of contracts.  Let’s

consider the absence of indexation up to the inflation trigger.  If the worker has only concerns

about her real wage and wants to minimize her risk, there will be a preferable contract with a

lower nominal wage but complete indexation.   But failure to index can be an optimal solution up

to a trigger, if the worker thinks that she should be receiving a nominal wage increase (and will

lose utility insofar as it fails to occur).  Such a utility function will also be consistent with the last

of the six observations, in which wage bargains tend to ignore inflationary expectations when

inflation is low, but add it on when inflation is high.  

Consequences of Nominal Considerations.  What are the consequences of such nominal

considerations in utility functions?   It is well-known that downward nominal wage rigidity will

induce a long-run trade-off between inflation and unemployment, contrary to natural rate theory. 

---

53

This proposition is known from Schultze (1959) and Tobin (1972).  For a simulation and estimation of the

trade-off, see Akerlof, Dickens and Perry (1996).

54

Of course workers may be making cognitive errors. They may fail imperfectly distinguish nominal

changes in wages from real changes in wages.

55

The intuition comes straightforwardly from considering the consequences of an increase in the nominal

rate of inflation.  At the higher inflation rate, if unemployment were constant the bargain between firms and workers

would result in lower real wages—since the typical employees, like Ann or Barbara, can be satisfied more easily. 

But such an increase in the real wage will depress the price that the representative firm wants to charge relative to

other firms.  It will fall below one.  The rate of inflation will be decreasing if that desired price is below one;

constant if it is equal to one; and increasing if its above one.   It will take an increase in aggregate demand (with

higher employment) then to keep inflation stable at the new higher level.  At this new higher level of employment,

the representative firm will want to charge a relative price of one.  The rise in that price will be the resultant of three

different factors: first, at the higher level of demand and of employment the labor market will be tighter, and

therefore the bargained real wage will be higher; second, the marginal product of labor will have declined; and also

the firm’s mark-up of prices over wages may have changed.

The logic is more complicated than is usually understood.  Most economists think that it is the increase in

demand for labor at the lower real wage that will cause more labor to be hired at the higher steady-state inflation. 

With perfect competition—where the desired price that the firm wants to charge is the real wage divided by the

marginal product of labor—the real wage will decline.  A rise in the marginal product then can only occur if there is

a fall in the real wage.  But that logic is not at all general.  In a standard model with monopolistic competition where

firms have a constant mark-up of wages over costs, if production is proportional to labor input, the equilibrium real

wage will be constant.  What allows unemployment to fall with steady inflation in such a model is that at each

unemployment rate the real wage that will come from labor bargains will be lower.  Only at lower unemployment

will that real wage be sufficiently high that the firm’s desired real price be restored to one.

56

There are three possible reasons for the violation of the neutralities.  In this lecture we have explored

preferences that include norms, which have been omitted from the preferences that yield the neutralities.  Previous

economists have emphasized the role of frictions.  But there is yet a third way in which the neutralities may be

violated: because of the wrong equilibrium concept.  This is demonstrated in a set of experiments by Fehr and Tyran

39

In that case higher long-run inflation will be associated with lower long-run unemployment.

53

  In

addition, if workers derive satisfaction from nominal wage increases, there will also be such a

trade-off.  If workers like Ann and Barbara feel happier the higher their nominal wage increase,

then higher steady-state inflation will be associated with lower steady-state unemployment.

54

 

The logic, as explained in the footnote, is straightforward.

55

  Since triggering in indexed wage

contracts suggests similarly that workers want nominal wage increases, this finding also suggests

that there is a trade-off between inflation and unemployment, at least when inflation is low. 

And, also, if there is only a low coefficient on inflationary expectations in Phillips Curves when

inflation is low, there is, once  again, a long-run trade-off between inflation and unemployment.

56

 

---