Milton Friedman’s Permanent Income Hypothesis

Milton Friedman’s Permanent Income Hypothesis Revisited

Milton Friedman is most certainly the most influential economist of the last fifty years and, perhaps with John Maynard Keynes, the most influential of the entire 20th century.  His theories and research continue to shape public policy debates even to today.

One of Friedman’s most influential and revolutionary theories was his challenge to the traditional Keynesian consumption function, which includes simple after-tax income as a variable in the consumption.  Friedman countered, however, that those who consume today take future taxes, price increases, salary increases, and other factors into account.  This is summarized in his Permanent Income Hypothesis.  More specifically, this counters that people consume based off of their overall estimation of future income as well as opposed to only the current after-tax income.

This seems to be a very sound hypothesis at first glance.  Consider the college student or the worker nearing the end of his career.  The college student borrows a lot of money (consuming it on tuition and other expenses) to go into a moderate amount of debt.  This is done in the expectation that future salary based off of the new college degree will more than make up for the increased debt now.  In this way, consumption is “smoothed” so that you are consuming only slightly less during college than you are after, even though you have a huge difference in income.

Milton Friedman - from Wikipedia
Milton Friedman – from Wikipedia

Another way to look at it is that a person is going to earn a certain amount of money in his lifetime.  The idea is to smooth the spending over a career based off of these expectations, as opposed to bouncing wildly around as raises and salary increases come.  This is more like the case of the near-retiree.  Near-retirees are usually at their highest or near-highest income in their entire lifetime.  At this point, however, they are usually saving the most in the expectation that their future income will decrease significantly (down to the amount of Social Security payments).  Thus, precautions are made so that the near-retiree’s lifestyle does not change too much the day he retires.  All of these anecdotal examples are pretty strong examples of Milton Friedman’s hypothesis.

Some clarifying remarks are necessary before moving on to more modern research on the topic.  All of the consumption smoothing decisions are based only on the information that is readily available to the individual when the consumption decisions are made.  If a person is going to win the lottery tomorrow, they are not going to start to increase their lifestyle dramatically today, simply because it cannot be known (legally) that one will win the lottery.  Similar examples can be made in stock market windfalls, larger than expected inheritances and other changes to lifetime income expectations, such as a major health problem which will severely damper the ability to earn wages.

Now, the real and most important reason that this is relevant to all policy decisions is that this has significant implications for government entitlement programs. It is especially relevant for those who can expect to be working for an extra ten to fifteen years.  So, for those under the age of 45, if an increase in spending today is expected to be equally matched with an increase in taxes during your working years, then there should be a negligible change in consumption and spending practices.  This means that the government spending will not in fact spur short term spending, it will instead incur more saving (there has been some evidence of an increase in saving rates lately) leading to lower interest rates and more investment which will counteract the crowd-out effect the government lending will have.  Milton Friedman is known for this counter-example and counter-hypothesis to Keynes’ deficit-fueled recession-fighting formula.  Friedman believes, instead, that monetary policy is the appropriate tool to fight recessions with and that short term lower rates lead to an increase in short-term output at the cost of long-term inflation.

A major problem for Friedman’s hypothesis is how much information can be known.  If you are 35, can you actually know to what extent an increase in spending now will lead to an increase in taxes for yourself later in your life?  Even if you do realize there will be a tax increase, will you trade out exactly 1 for 1?  My initial gut reaction would be ‘no’, but that there could be a tradeoff, so that perhaps the true way citizens react to spending increases is a blend between Keynes’ and Friedman’s hypotheses.

One of the best papers that directly tests this hypothesis is one by David Wilcox entitled “Social Security Benefits, Consumption Expenditure, and the Life Cycle Hypothesis” printed in the Journal of Political Economy in April, 1989.  The paper looks at Social Security benefit increases in between 1974 and 1982.  Before 1974, increases in Social Security benefits for recipients were random and varied wildly.  In 1974, however, a law was put into effect that linked the SS payment increases to the increase in the CPI.  Over the next eight years (which even continues to today), the increase in Social Security benefits were announced about two and a half months before the effect took place.  Now, this clearly directly tests the hypothesis.  These are known increases in a Social Security recipient’s income.  The hypothesis would predict that before the effect went into place the retiree would then increase their spending (borrowing against their future income) so that their consumption would smooth.  Thus, there would be a negligible, statistically insignificant change in spending habits before and after July 1st of the year (the date that the change goes into effect).  Even though there are nuances in the paper that would be too long of a read to detail, the paper found that there was a rather large percentage increase in spending amongst retirees found immediately after the date of effect took place.  This approach and paper provides strong evidence that Milton’s hypothesis does not hold too much water.  For the paper itself, please see:

Obviously there is much work to be done on determining the appropriate measure to help spur the economy during difficult times and a lot more light to be shed on the Keynes versus Friedman debate of Monetary versus Fiscal policy.  What is clear, however, is that the debate may not go away for a long time and that the results of which will determine how future generations of politicians shape their economic policy.

For the full explanation of Milton Friedman’s hypothesis and research on the topic:



Cameron Daniels


  1. says

    The problem with the savings rate right now is that a major fraction of U.S debt is in mortgages, and obviously with the foreclosure rate increasing and housing prices (and thus Debt to Income ratios for those who have jobs and buy homes have decreased) so the savings rate will go up simply due to that technical issue.

    As for the corporations, if there is an incentive (for corporations there are often reasons to do many crazy ideas) to save, then I would expect them to either:
    a.) invest in their own employees to get them to stay, in programs such as education and other benefits programs
    b.) hand out bonuses so that their end-of-year net income decreases, which is done so that they will lower their tax liability
    c.) or, what the government hopes: hire more people/lower prices.

    As of now, I would say the majority of it would be in the case of c: lower prices, which is counteracted with raise freezes and may not be sustainable without the stimulus dollars. Otherwise, it is tough to say simply because corporate money today has a very low chance of strictly correlating with an increase in corporate taxes tomorrow.