Keynesians, Monetarists, and Unemployment

The old debate between Keynesians and Monetarists is often considered to have been about the monetary transmission mechanism or the Phillips curve. However, an important element of the debate was in regards to unemployment and its effect on real output.

The typical Keynesian story about unemployment is that an increase in unemployment reduces income, which reduces consumption, and reduces aggregate output. Individuals base decisions on current income receipts. Unemployment benefits thus serve to compensate the job loser from the fall in income.

The monetarist view of unemployment, however, is framed within the context of Milton Friedman’s (1957) permanent income hypothesis. According to the PIH, permanent income is known, but there is uncertainty surrounding current receipts. A job loser whose behavior is characterized by the PIH will increase measured unemployment, but whether or not the job loss impacts real output depends on whether the change of circumstance results in a permanent downward revision of anticipated income. An increase in unemployment that is expected to only cause a change in receipts will not cause a reduction in consumption or real output.

Consider the implications of each theory. Suppose that the demand for labor falls. Employment and current income receipts decline. In the Keynesian theory, unemployment could be reduced if money wages were to decline. However, money wages are fixed by institutional datum. Unemployment compensation is necessary to prevent further reductions in real output because consumption is based on current income receipts.

In contrast, in the Monetarist theory, the reduction in employment and current receipts only affects output to the extent that anticipated income declines. If individuals fail to revise down expectations of permanent income, there is no effect on output. What’s more, money wages in the Monetarist theory remain sticky. However, unlike the Keynesian story, this is not the result of institutional datum. Rather, money wages are sticky because the supply of labor is a function of the real wage AND permanent income. Thus, unemployed workers only accept lower money wages if they revise down their estimates of anticipated future income. Unemployment compensation thus serves to smooth out receipts, but it does not affect consumption.

I bring up this distinction because much of the talk surrounding extensions to unemployment benefits has been framed in the context of the Keynesian theory or the labor-wedge-real-business-cycle-type theory. As a result, the debate has focused on the disincentive to work created by unemployment benefits. While I have little doubt that unemployment benefits and expansions thereof provide a disincentive to work, I am not convinced that this effect is large in magnitude.

In contrast, I think that looking at this debate in the context of the Keynesian v. Monetarist debate provides much more insight. According to the Keynesian theory, prolonged unemployment causes a reduction in income receipts and therefore a decline in aggregate demand and a reduction in real output. Increases in the duration of unemployment benefits thus serve to prevent (at least in part) the reduction in income receipts thereby preventing reductions in aggregate demand and real output. Meanwhile, the Monetarist theory suggests that extending unemployment benefits smooths current receipts, but has little effect on aggregate output.

This distinction between Keynesian and Monetarist theories is important as the former implies that increases in unemployment compensation have a stimulative effect on the economy. Meanwhile, the Monetarist theory suggests that such extensions will have little, if any, effect.

UPDATE: Blog friend Nick Rowe suggests in the comments that I am arguing that output does not fall at all in the Monetarist theory. I hope that is not the takeaway from this post. The decline in unemployment should be understood as the result of a real or nominal shock that pushes output below potential. The difference between the Keynesian and the Monetarist theories is whether consumption declines following the increase in unemployment and whether the extension of unemployment benefits is “stimulative”. For more, see my thoughts in the comments.

21 responses to “Keynesians, Monetarists, and Unemployment

  1. Wait, they are both wrong!

    What people and other mammals do is arrange their affairs such current inventory variance matches a biological uncertainty constant that is fixed within each mammal species. When we go to work, when we are at home, we are guided by a biological expected arrival and consumption of inventory, the arrival rate fixed in our brain.

    The wealthy man and the poor man both contrive to keep the inventory of food, housing, clothing at the same variance; as does the factory worker, the factory owner and manager.

    When the inventory of some good, say money, accumulates, they adjust all other inventories to put money and other goods back to the same relative balance, even if it means moving to a rich neighborhood.

    The effect of constant uncertainty, or constant variance is to force us into production lines such that we meet that criteria. It is the biological cause of Zipf’s law., it is the reason that skyscrapers height, language, and business hierarchies all follow the power law pattern. Mathematically, the correct way to handle this process is something called a Shannon Channel reaching maximum throughput in a channel modeled by fixed variance Gaussian noise. One also needs the concept of an entropy norm in economic theory.

  2. Josh: I disagree on this.

    For both Keynesians and Monetarists, there is a link between unemployment and output via the production function: fewer people working means less output produced, given capital stock, technology, etc. Or, the less output that firms can sell, the less they need to produce, and so the fewer employed workers they need to produce it. Allow for a bit of labour hoarding during temporary recessions, and you get Okun’s Law.

    That’s the same for both Keynesians and Monetarists.

    The relevance of the Permanent Income Hypothesis is that it determines the marginal propensity to consume out of current income, and thus affects the magnitude of the multiplier.

    Suppose there’s an initial fall in aggregate Demand, output, income, and employment. If you believe the current income hypothesis, the MPC is large, so the initial fall in AD and income causes a large subsequent fall in AD and income. Under the PIH, since current income doesn’t affect permanent income much (if the recession is expected to be short), the mpc is smaller, so the subsequent fall in AD is smaller. Smaller multiplier under PIH.

    An exogenous shocj shifts the IS curve left by less under PIH, and the IS curve is also steeper.

  3. Nick,

    I think that you have misunderstood me. I am taking the initial decline in output as given. In other words, I am assuming the increase in unemployment is as a result of a decline in output.

    This is the sequence of events:

    1. AD declines, aggregate output falls.

    2. Current income receipts decline.
    a. In the Keynesian theory, the reduction in current receipts causes a reduction in consumption and therefore a further reduction in AD.
    b. In the Monetarist theory, consumption only declines if the fall in income receipts is expected to be permanent. If not, there is no further reduction in AD.

    Thus, if unemployment benefits are extended, the Keynesian theory implies that this prevents (at least in part) the second shift in AD. In the Monetarist theory, the extension smooths income receipts, but will not affect consumption. Consumption will only decline if permanent income is revised downward.

  4. Nick,

    Here is Allan Meltzer (1978):

    “Every day of lay off is counted as part of measured unemployment, but, as long as income anticipations are not revised, permanent income remains unchanged and the worker is not unemployed in an economic sense. There is no loss of aggregate output.”

    Meltzer again is taking the following condition as given:

    “Both [real and nominal] shocks change current output, y, relative to anticipated output…If the observed fluctuations in current output are drawings from an anticipated distribution, the deviation of the current output does not cause of revision of plans in the aggregate.”

    In other words, AD falls, output falls, measured unemployment increases. The question is therefore how the decline in income receipts and increase in measured unemployment influences consumption.

  5. Josh: Sorry. I did misunderstand you.

    One point though, on Allan Meltzer.

    If I unexpectedly find a $100 bill, my permanent income does rise, by $5 per year, assuming a 5% interest rate., even if I never expect to find another.

  6. You are not looking at the big picture of full social macroeconomics. The unemployed people are a drain on the tax payer, but when they are employed they pay taxes from what they earn and spend. “Creating work” by the government means taking more money from taxation for this purpose. Otherwise the government can borrow money from the banks and public (for which the tax payer covers the interest) or print up some fresh currency (for which inflation makes the savers loose but the owers (particularly the government) gain.

    In order to reduce unemployment we need to raise the demand for goods by causing their production costs to reduce and the best and obvious way to achieve this is by taxing land values instead of incomes and purchases.

    14 ASPECTS of LAND-VALUE TAXATION affecting Government, Land Owners, Community and Ethics

    3 aspects for Government

    1. Most of the ground-rent being collected as LVT, adds to the national income. It allows the other taxes on earnings, purchases and family/corporate ownership of buildings to be reduced and eventually to be eliminated.

    2. The ownership of each land parcel is registered. Then the cost of collecting the LVT is much smaller than for income tax and other production-related taxes. Using regularly updated maps, the rental value of each site (as if without buildings) is public knowledge. Then the LVT is simple to understand, the amount of tax is easily found and its payment by the land owner is impossible to avoid.

    3. With LVT, the national economy stabilizes and no longer experiences the 18 year housing boom and bust cycle, which was due to the changing prices that arose from speculation in land-values during town expansion.

    6 aspects affecting Land Owners

    4. LVT is progressive, the owners of the most potentially productive sites pay the most tax. None is paid on marginally productive sites, since their owners cannot claim ground-rent from possible tenants.

    5. The land owner pays his LVT regardless of how the land is used. When the land is leased to tenants most or all of the resulting ground-rent is the tax.

    6. LVT stops the speculation in land prices because any withholding of land from proper use is too costly.

    7. The introduction of LVT reduces the sales price of sites even though their value (or potential usefullness) may continue to grow.

    8. With LVT, land owners are unable to pass the tax on to their tenant renters, due to the competition for land use. The users of (untaxed) marginal sites price their produce according to the costs of their labour, the use of the durable capital and the added transport needs. Owners/occupiers who access more productive land pay LVT/ground-rent and compete in their production, so this tax cannot be added to what buyers willingly pay.

    9. With the introduction of LVT, land prices will drop. Speculators in land values will tend to foreclose on their mortgages and to withdraw their money for reinvestment. Depending on the rate of these changes, bankrupcies can result. Then LVT should be introduced gradually to allow the investors sufficient time to transfer money to company-shares in durable capital goods, where their greater use will meet the increased demand for produce (see below).

    3 aspects regarding our Community

    10. With LVT, there is an incentive to use land for production, rather than it laying idle or being partly used. An optimum amount of urban land is brought into use, which reduces the spread of suburbs onto rural land and avoids vacant city centers.

    11. With LVT, greater working opportunities exist due to cheaper land and a greater number of available sites. Consumer goods become cheaper because entrepreneurs have less difficulty in starting-up and running their businesses. Demand grows, unemployment decreases and with it a reduction in the polarization of our class-society and its degree of poverty.

    12. As LVT is introduced, investment money is withdrawn from land and placed in durable capital goods. The investors in company shares tend to be wage-earners (as well as banks and monopolists). Their decisions favour more competition and cheaper local production without heavy transport costs, whilst the monopolists have less control of prices and the unavailability of alternative goods. This is a natural trend of our free-marketing social system.

    2 aspects about Ethics

    13. The collection of taxes directly from productive effort and commerce is socially unjust. The associated philosophy favours coercive robbery and is “Robin Hood” in style. LVT replaces this form of extortion by gathering the surplus rental income which comes without exertion. Consequently LVT is a natural system of money-gathering, which avoids the present-day distortion of business economics.

    14. Bribery and corruption cease with LVT. Before, this was due to the leaking of news of municipal plans for housing development. However, the speculation in land values is no longer worthwhile after LVT is in place.

  7. David Beckworth


    With all issues, the devil is in the details. If someone is unemployed and sees no change in his/her permanent income then it works. But, if this person continues to stay unemployed for some time there is the chance of losing important human capital and lowering one’s permanent income.

    On another note, aren’t the principles of PIH implicit in New Keynesian models?

  8. paul davidson

    Keynes (as oposed to “Keynesians”) did NOT believe that lowering the money wage would increase employment. If you read Chapter 19 of THE GENERAL THEORY entitled “changes in money wages” especially page 257 you will see that Keynes specifically stated that classical theory assued that a rigidity in money wages prevented the system from automatically getting back to full eployment but he said “my difference from this theory is primarily a difference of analysis”.. If one reads my JOHN MAYNARD KEYNES book, [2007, Macmillan] or my 2009 book THE KEYNES SOLUTION [Palgrave/Macmillan], you will see why KEYNES does not believe that flexible wages per se would cure unemployment.

    PLEASE also read why Paul Samuelson never understood the General Theory — as explained in Samuelson’s own words in either book!!!

  9. paul davidson

    Regarding the permenant income hypothesis — If you read page 43 of my book POST KEYNESIAN MACROECONOMIC THEORY [elgar, 1994] you will note that Friedman boasts that he does not define consumption in the usual way. Consumption is defined as “the value of services[utility] consumed” during the period. Accordingly for all durable consumer goods purchased, the value of consumption of such goods is the depreciation (or wearing out) during the accounting period [ See Friedman, p. 11 of his book THE THEORY OF PERMANENT iNCOME]. Friedman boasts that his taxonomy is superior to others because “much that one classified as consumption is reclassified as savings” [p.28] Thus transitory income is almost always saved — since it most likely to buy durables (or as an earlier comment noted if one found $100 bill it is equivalent of $5 per year in icom — or if the $100 buys a durable, only $5 of utility of the durable is used up (consumed) each year!

    Such uncommon use of common language is misleading for it encourages those who do not understand the strange taxnomy of Milton F.that savings (the buying of a producible durable) creates jobs just as much as consumption (defined commonly). But SAVINGS creates jobs — while in Keynes’s terminology– Savings is the decision not to spend today’s income on any producible goods –it is a nonemployment decision . Instead savings out of current income is used to buy liquid assets that has certain “essential properites” [see Chapter 17 of the general theory] such that the elasticity of production and the elasticity of sbstitution are both zero! Elasticity of production means that money (and all other liquid assets) do not grow on trees and therefore when the demand for liquid assets out of current income is diverted to savings — jobs are, ceteris paribus, lost. Moreoever if the demand for savings increases, then although the demand for liquid assets including money may rise and push up the price of all lliquid assets, this increase in liquid asset prices relative to the prices of producibles do not enourage savers to substitute producibles for nonproducible liquid assets.

  10. paul davidson

    Did both of my comments get through? If so I hope they made sense

  11. David,

    You wrote, “With all issues, the devil is in the details. If someone is unemployed and sees no change in his/her permanent income then it works. But, if this person continues to stay unemployed for some time there is the chance of losing important human capital and lowering one’s permanent income.”

    Yes, this is correct. However, I don’t see how it applies to the analysis of unemployment compensation. If individuals revise down their estimate of permanent income, this means that they will be willing to accept a lower wage. Since this doesn’t take place immediately, money wages are sticky. Nonetheless, a downward revision of permanent income does not imply that an extension of unemployment benefits is now warranted. Even under this scenario, the extension of unemployment benefits would only increase transitory income and therefore would not affect consumption. Thus, regardless of whether permanent income is revised downward, an extension of unemployment benefits does not have a stimulative effect.


    I wholeheartedly agree with you that the position that I attributed to “Keynesians” was not the position of Keynes.

  12. Nick,

    You wrote, “One point though, on Allan Meltzer.

    If I unexpectedly find a $100 bill, my permanent income does rise, by $5 per year, assuming a 5% interest rate., even if I never expect to find another.”

    You are correct, but earlier in the piece Meltzer assumes (somewhere in the middle of those two quotes), “as long as anticipations are firmly held, the present value of every negative deviation is offset by anticipated positive deviations.”

    He should be forgiven for my omission.

  13. It’s Nice Post, keep posting and have a nice day… 08:54

  14. Pingback: Unemployment Benefits | The Everyday Economist

  15. thank you, i needed this badly.

  16. thank for then contributions…….

  17. just summarize

  18. It’s fantastic that you are getting ideas from this paragraph as well
    as from our discussion made here.

  19. I support the keynesian view whose state that, an increase in unemployment lead to a reduction in income, which also lead to a reduction in consumption culminating in reduction in real output. Because in a generical sense without employment, it would lead to shortage in permanent income, which make you underconsumption, that lead to products or goods unsold in the market.

  20. this should be the beat answer to my question. love it……….

  21. fully convinced

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s