Is it Really that Bad to be Irrational?

Several years ago I heard a joke about two macroeconomists who went hunting with a non-economist colleague.  As they tracked a deer up a slope and over the crest of a hill, they noticed it edge into a clearing, making a perfect silhouette on the horizon.  Within heartbeats of each other the two economists fired.  One missed by ten feet to the right and the other missed by ten feet to the left.  The deer ran off immediately, unharmed and the two men yelled in joy and began to hive five and slap each other on the back.  Their colleague asked how they could be so excited when they had both missed by a substantial margin.  They replied. "Yes, we missed.  But on average we hit him right between the eyes!"

Modern macroeconomic thinking incorporates the concept of rational expectations.  Rational expectations is the notion that while individuals in the economy may not always be able to accurately predict the future, they at least do not consistently underestimate or overestimate.  On average people hit their forecast targets right on.

To continue with the hunting analogy, suppose you have a gun with a bad sight.  When you fire it into a target you consistently hit it too low and to the right.  If you could not adjust the sight and were forced to take the gun hunting, you would mentally adjust by aiming a bit high and to the left of the intended target, in an attempt to correct the bias.  Households and firms do something similar when they forecast the future of the economy.  Household savings decisions, for example, are based on expectations about what interest rates will be earned in the future and what wages the household can expect in future years.  Often household rely on analysis from professional financial planners and/or economy-watchers, but they make the same kinds of adjustments.

Rational expectations as an economic theory says that agents in the economy will take all relevant information into account when forming a forecast or expectation of the future.  Failure to do so leads to mistakes that the agents will regret later.  Just as failure to adjust for the bad sight on a gun leads failure in hunting, failure to adequately process all relevant information leads to bad economic outcomes.  Incorporating this insight into economic models has led to a vast improvement in the quality of macroeconomic models over the past several decades.

But are people really rational?  Not all economic researchers are convinced that rational expectations is a good approximation of actual behavior.  Behavioral economists often prefer an alternative model of behavior that stresses time-inconsistency.  Agents that behave this way make plans for the future with the intent of following those plans, but when the future rolls around they often change their mind.  A time-inconsistent planner might eat burger and fries for lunch today with the intent of starting a healthy diet of salad and fruit tomorrow.  However, when lunch time rolls around again the next day, he eats the burger again and the salad gets postponed .  The end result is someone who is heavier and less healthy than he initially planned.

From a macroeconomic perspective, this behavior is important only if it affects the aggregate behavior of the economy as a whole.  One area where it does have an effect is in economic policy.  Time-inconsistent households, for example, are argued to be less likely to save for the future.  They will tend to put off savings and consume instead.  Hence, they arrive at retirement with less savings than they need.  This means there is a role for the government to play in saving for the household or forcing the household to save.  One of the reasons - among many - that the US Social Security system was set up was to alleviate this problem.

In this light, a recent working paper by two macroeconomists at Utah State University's Huntsman School of Business is quite enlightening.  Frank Caliendo and Scott Findley show that when one accounts for the age-of-retirement decision, things do not look so bad.  They compare a model economy where all the agents are rational with one where all the agents have a time-inconsistency problem.  In both models, agents retire at the same age and earn an exactly equal amount of income over their lifetimes.  The rational individuals choose to retire at age 65, make a consistent savings plan, and stick to it over the course of their lifetime.  The time-inconsistent individuals choose to retire earlier when making plans at young ages.  But, they tend to over-consume and under-save and end up pushing their planned retirement age back each year.  In the end, they too retire at age 65.

The interesting result is that both sets of agents end up retiring with roughly the same stock of wealth set aside.  Time-inconsistent people under-save relative to their planned early retirement, but end up saving the appropriate amount for the age they actually retire.

The policy implication of this work is that the need for a government correction to the savings problem is not nearly as strong as previously thought.  If Caliendo and Findley's work is representative of the actual economy, then even if all the households in the economy are poor savers, they still end up saving the right amount.