Conference PaperPDF Available

What is Full Employment-and Why the Definition Matters

September 2016

September 2016

Conference: International Post Keynesian Conference
At: University of Missouri at Kansas City

Authors:

Rutgers, The State University of New Jersey

Growing numbers of progressive economists have started referring to 4% unemployment as "full employment" or close to it. This article challenges that claim. The origins of the full-employment concept-first as a theoretical category in Keynes's General Theory, and then as a progressive public policy goal-is reviewed, and the question is asked whether 4% unemployment corresponds to either the Keynesian concept or the meaning accorded full employment by the public and in the law. The conclusion drawn is that it does not, but uncertainty is expressed as to whether progressive economists who have adopted the usage contested here recognize the disparity. Whether they do or not, it is argued their de facto revision of the full-employment concept has decidedly negative consequences. It stigmatizes unsuccessful job seekers, muddies the public's understanding of actual labor market conditions, inhibits the development of effective full employment policies; and undermines the success of progressive reform efforts.

Content uploaded by Philip Harvey

Content may be subject to copyright.

Page 1 of 22

What is Full Employment—and Why the Definition Matters

Philip Harvey*

Paper Presented at the International Post Keynesian Conference

University of Missouri at Kansas City

Sept. 17, 2016

ABSTRACT

Growing numbers of progressive economists have started referring to 4% unemployment as

“full employment” or close to it. This article challenges that claim. The origins of the full-

employment concept—first as a theoretical category in Keynes’s General Theory, and then as a

progressive public policy goal—is reviewed, and the question is asked whether 4%

unemployment corresponds to either the Keynesian concept or the meaning accorded full

employment by the public and in the law. The conclusion drawn is that it does not, but

uncertainty is expressed as to whether progressive economists who have adopted the usage

contested here recognize the disparity. Whether they do or not, it is argued their de facto revision

of the full-employment concept has decidedly negative consequences. It stigmatizes unsuccessful

job seekers, muddies the public’s understanding of actual labor market conditions, inhibits the

development of effective full employment policies; and undermines the success of progressive

reform efforts.

JEL Codes: E12, E24, E61, J23, J68, K38

* Professor of Law and Economics

Rutgers Law School

philip.harvey@rutgers.edu

Page 2 of 22

What is Full Employment—and Why the Definition Matters

In the State of the Union message last year I set forth what I

considered to be an American economic bill of rights.

I said then, and I say now, that these economic truths

represent a second bill of rights under which a new basis of

security and prosperity can be established for all- regardless

of station, race, or creed.

Of these rights the most fundamental, and one on which the

fulfillment of the others in large degree depends, is the right

to a useful and remunerative job . . .

Franklin Delano Roosevelt

State of the Union Message

Jan. 6, 1945 (his last)

INTRODUCTION

During the so-called dotcom boom of the late 1990s the U.S. unemployment rate stayed below

5% for 50 straight months. Towards the end of the boom it averaged 4% for over a year. It was

the U.S. economy’s best overall labor market performance since the Vietnam war years of the

late 1960s, and the best peacetime labor market performance since the federal government began

reporting unemployment data on a systematic basis in the early 1940s. In short, it was a big deal.

Moreover, as progressive economists have pointed out, it was an especially big deal for low

wage workers and disadvantaged population groups who benefited disproportionately from the

economy’s relatively low levels of unemployment (Bernstein and Baker 2003; Wilson 2015).

However, progressive economists haven’t just trumpeted these benefits. A growing number

have begun referring to the 4% unemployment achieved at the end of the dot com boom as full

employment or close to it.1 That is the claim addressed in this article. Does 4% unemployment

deserve to be described as full-employment—and is the answer to that question substantively

important?

I begin with a review of the origins of the full-employment concept—first as a theoretical

category in Keynes’s General Theory (1936), and then as a progressive public policy goal. I then

consider whether 4% unemployment corresponds to either of these conceptions of the term’s

meaning. I conclude that it does not, but express uncertainty as to whether the progressive

economists who equate the two recognize the disparity. Whether they do or not, I argue that their

de facto revision of the full-employment concept has decidedly negative consequences. It

stigmatizes unsuccessful job seekers at the top of the business cycle; muddies the public’s

understanding of actual economic conditions, inhibits the development of economic strategies

that actually could achieve full employment; and thereby undermines the success of progressive

reform efforts. This judgment is harsh, but I believe it is deserved.

Page 3 of 22

What Is “Full Employment”?

Prior to the publication of the General Theory, economists rarely used the term full

employment, and when they did, their usage was generic. They referred to individual workers

finding full employment, to the full employment of a business firm’s credit, to the full

employment of a firm or industry’s capital stock, to the full employment of the factors of

production (including labor), to a group of workers or a particular occupation enjoying full

employment and so on.2 Only in the work of economic reformers was the term occasionally

used in its modern sense to refer to the policy goal of furnishing “full employment” for all

workers (A Practical Man, 1828; Maxwell, 1891). However, these precursors of Keynes’s usage

were too isolated and idiosyncratic to give the term any special meaning prior to the publication

of Keynes’s General Theory in 1936.

However, the fact that Keynes “invented” the concept of full employment doesn’t mean he

and his followers retained control over its meaning as it entered the lexicon of public policy

discourse in the decade following the General Theory’s publication. In fact, given the purely

theoretical nature of Keynes explanation of the concept in the General Theory, a translation of its

meaning into a more easily understood set of characteristics was crucial to its popularization.

Keynes made clear that the General Theory was a theoretical treatise whose intended

audience consisted of other economists rather than the general public.

I hope that it will be intelligible to others. But its main purpose is to deal with

difficult questions of theory, and only in the second place with the applications of

this theory to practice. (Keynes 1936: v)

In keeping with this purpose, the descriptions and definitions of full employment Keynes

included in the General Theory (of which there are several) are designed for theoretical rather

than practical purposes. They include

• the “the maximum quantity of employment . . . compatible with a given real wage”

(Keynes 1936: 12)

• the elimination of “involuntary unemployment”—but not of “frictional” or “voluntary”

unemployment (ibid.: 6, 15-16)

• the point at which “a further increase in the value of effective demand will not be

accompanied by any increase in output” (ibid.: 26)

• the point at which the “aggregate demand function” intersects with the “aggregate supply

function” (ibid.: 30)

• the point at which “money-wages have to rise, in response to an increasing effective

demand in terms of money, fully in proportion to the rise in the prices of wage-goods”

(ibid. 301)

Keynes fashioned each of these definitions to fit a particular theoretical context. The

definitions were not designed to provide a description of full employment that would be

intelligible to non-economists. Nor were they designed to provide a practical means of measuring

the achievement of full employment. The predictable result was disagreement among economists

Page 4 of 22

as to how the achievement of full employment should be measured in the real world (Reese

1957).

Because the General Theory did not answer this questions and because the economists

among Keynes’s popularizes also failed to answer them. the promotion of full employment as a

practical public policy goal has generally relied on the public to form its own impression of what

the goal means in practical terms—without regard to the various ways in which economists

(beginning with Keynes) have conceived and defined it.

This was especially true when the goal was first popularized during World War II. The term

was still brand new; it had no prior associations attached to it; it possessed unusual salience due

to memories of the Great Depression; and professional economists were principally concerned

with the promotion of Keynes’s policy advice rather than with painting a picture of life under

conditions of full employment. In that context progressives had both the freedom and motivation

to invest the full employment goal with a full measure of their aspirations.

The result was the emergence of significant differences in the way economists and the

public conceived of the full employment. Policy-oriented economists such as Keynes and his

followers tend to view themselves as physicians of the economy. Their ultimate motivation may

be to help people or promote a particular vision of the good life; but the economy is the

immediate object of their attentions. It is the “patient” their policy prescriptions are designed to

“treat.”

For economists, therefore, full employment is likely to be conceived as something the

economy achieves or fails to achieve, a condition characterized by the relationship of certain

economic parameters, generally conceived in abstract terms, to one another—the aggregate

supply of and demand for labor, the aggregate level of employment, aggregate demand, the price

level, and so on.

Non-economists tend to conceive of full employment in very different terms—as something

that individual workers experience. The “ready availability of decent jobs at decent wages” are

its markers rather than an abstractly conceived equilibrium between the aggregate supply of

labor and the aggregate demand for it.

Keeping this in mind is important for a proper understanding of the term’s meaning in

different contexts. It doesn’t mean the two ways of thinking about full employment will

necessarily conflict with one another. Indeed, the reason the progressive embrace of full

employment as a public policy goal occurred so quickly and went so smoothly in the 1940s is

because the economist’s vision of the goal was in sync with the public’s vision of the goal. In

other words, economists were confident they could deliver what the public expected full

employment to achieve, so there was no reason to ramp back the public’s expectations. “We’re

not talking about pie in the sky. You can have pie in this life as soon as we win control over the

levers of fiscal policy.”

On the other hand, the different ways in which economists and the public tend think of full

employment can also result in a divergence in expectations. Non-economists might think of full

employment as the ready availability of decent jobs at decent wages for everyone who wants to

Page 5 of 22

work, while professional economists (even progressive ones) think of it as the lowest level of

unemployment consistent with the maintenance of reasonable price stability. But more on that

later.

To return to our story, progressives not only enjoyed wide latitude in imagining what full

employment meant when they embraced the goal in the later part of World War II. In the United

States, at least, they had two highly salient reference points to guide their thinking. The first was

the U.S. economy’s 1.7 percent average unemployment rate during the last three years of the

war, a condition that Keynesian economists were quick to cite as a practical demonstration of

what Keynes’s economic policies could do (Galbraith 2000). It was hardly surprising that

American progressives perceived the labor market conditions that existed during that period as

full employment in practice.

The second reference point that influenced the public’s conception of full employment at the

time was President Roosevelt’s advocacy of the right to work as an entitlement that governments

had a duty to strive to secure for all members of society (Harvey 2013b). The President’s 1944

State of the Union Address was particularly noteworthy in this regard, but he had been saying the

same thing since before he was first elected President in 1932. In his 1944 address, Roosevelt

called on Congress to enact legislation securing nine economic rights that he argued were needed

to realize all of the human rights proclaimed in the nation’s Declaration of Independence—i.e.,

“life, liberty and the pursuit of happiness” The first item on Roosevelt’s list was “the right to a

useful and remunerative job.” The second was “the right to earn enough to provide adequate food

and clothing and recreation” (Roosevelt 1944).

Progressives in Congress responded to Roosevelt’s call by drafting legislation designed to

achieve full employment via the Keynesian strategy of using government spending to maintain

aggregate demand at the full employment level (U.S. Congress 1945). In fact, the right to work

was mentioned only once in this proposed legislation—just enough to make it clear that the bill

was designed to secure that right. Yet no one, including President Roosevelt, thought there was

any problem in this substitution of the full employment goal for that of securing the right to

work. Why? Because the two goals were viewed as synonymous.

The effort to enact this legislation failed (Bailey 1950). Its importance for our inquiry lies in

what it tells us about the correspondence between progressive conceptions of full employment

and the right to work at the time the former goal was embraced as a progressive public policy

goal.

A similar account of the meaning of full employment (if we ignore its gendered language)

was articulated at the same time in the United Kingdom by William Beveridge, a highly

respected labor economists and social welfare reformer who served as the Director of the London

School of Economics from 1919 to 1937. Beveridge described full-employment in the following

terms in a highly influential 1944 book entitled Full Employment in a Free Society.

It means having always more vacant jobs than unemployed men, not slightly

fewer jobs. It means that the jobs are at fair wages, of such a kind, and so located

that the unemployed men can reasonably be expected to take them; it means, by

Page 6 of 22

consequence, that the normal lag between losing one job and finding another will

be very short. (Beveridge 1944: 18).

This was the generally accepted conception of full employment at the time, and as such it

was accorded formal recognition in international law along with FDR’s claim that access to a

decent job was a human right that governments have a duty to strive to secure (U.N. 1945: arts.

55 & 56; 1948: art. 23; Harvey 2013).

It is important for progressive economists to recognize that this history gives the public

understanding of the full employment goal a status in discussions of public policy that they do

not control—notwithstanding the fact that it was an economist who invented the concept in the

first place. The economics profession lost ownership of the concept by virtue of their very

success in promoting it. The non-economists who embraced the full employment goal neither

understood nor cared how Keynes and other economists conceived of it in theoretical terms.

They invested the goal with a meaning based on their own sense of what full employment meant

for individual workers—the availability of decent work for everyone who wants to work.

It is also important for progressive economists to recognize that this conception of full

employment suffered no erosion over time, even if its political fortunes did. Its durability three

decades later is readily apparent in the advocacy the right to work by Coretta Scott King and

other progressives in the 1970s (Stein 2016) and in the presidential election platforms the U.S.

Democratic Party adopted that decade. The party’s 1972 platform described full employment as

“a guaranteed job for all” and called its achievement “the primary economic objective of the

Democratic Party” (Democratic Party 1972). The party’s 1976 platform reaffirmed the party’s

commitment to “the right of all adult Americans willing, able and seeking work to have

opportunities for useful jobs at living wages (Democratic Party 1976).

The legislation promoted by the Democratic Party in the 1970s to achieve these goals—the

Humphrey-Hawkins Full Employment Act—suffered the same fate as the full employment bill

progressives crafted in response to FDR’s 1944 State of the Union Address. It was stripped of its

operative sections designed to achieve its goals. Nevertheless, the enacted version (The Full

Employment and Balanced Growth Act of 1978) does include a strong policy commitment to the

achievement of the right-to-work conception of full employment.

The Congress declares and establishes as a national goal the fulfillment of the

right to full opportunities for useful paid employment at fair rates of

compensation of all individuals able, willing and seeking to work (15 U.S.C. §

1021(b)).

This commitment still resides in the U.S. Code as the statutorily-mandated goal of the U.S.

government with respect to the achievement of full employment.

Contrary to what a number of commentators have suggested (see, e.g., Bernstein 2006: 90;

Palley 2007: 7) the Humphrey-Hawkins Act does not link this goal to a targeted unemployment

rate. The only definition of full employment included in the act is the one expressed in the policy

commitment quoted above. What the act did do was direct the President to develop

programmatic and budgetary proposals designed to achieve the interim goal of reducing the

Page 7 of 22

unemployment rate of persons 20 years of age and older to 3% and of persons 16 years of age

and older to 4% within 5 years; and “[u]pon achievement of the 3 and 4 per centum goals” to

pursue “the goal of achieving as soon as practicable and maintaining thereafter full employment

and a balanced budget” (15 U.S.C. § 1022a(b)(1) and (c)(1)).

The only year in which either of the interim goals established by the Humphrey-Hawkins

Act has been reached was 2000. That year, as noted above, the nation’s overall unemployment

rate (i.e., the rate for persons 16 years of age and older) was 4%. However, the rate for persons

20 years of age and older was 3.4%, a few ticks above the 3% interim goal for “adults.” It goes

without saying that the act’s ultimate goal—the achievement of full employment conceived in

the right-to-work sense of the term—has never been achieved.

Indeed, the passage of the Humphrey-Hawkins Act turned out to be the Democratic Party’s

swan song in the pursuit of full employment. In 1980 the party’s presidential campaign platform

“reaffirmed” its commitment “to achieve all the goals of the Humphrey-Hawkins Full

Employment Act,” but the commitment appeared in a laundry list of other economic goals

(Democratic Party 1980). In 1984 there was no reference to the Humphrey-Hawkins Act and the

party’s “commitment to full employment” was treated on a par with its “commitment to housing”

and “rebuilding the infrastructure of America” (Democratic Party 1984). In 1988 the term

appeared only once in the party’s platform, in a sentence expressing the belief that “as a first-rate

world power moving into the 21st century, we can have a first-rate full employment economy”

(Democratic Party 1988). Through the next 6 presidential election cycles, the term did not appear

at all in the party’s platform. Then, this past summer, it finally reappeared in the party’s 2016

platform, which commits the party to “doing everything we can to create a full-employment

economy, where everyone has a job that pays enough to raise a family and live in dignity with a

sense of purpose” (Democratic Party 2016). What’s important about this commitment—other

than the fact of its reappearance after a 28-year absence from the party’s platform—is that it

demonstrates the enduring vitality of the right-to-work conception of the full-employment goal.

Nor is it necessary to rely on this indirect evidence of the durability of that conception of the

goal. When directly queried, progressives still describe full employment the same way they did

in the 1970s (Goldberg, Harvey and Ginsburg 2007).

It's clear that the precipitous decline in progressive support for the pursuit of full-

employment at the end of 1970s had nothing to do with any dissatisfaction or disaffection with

the conventional (among non-economists at least) conception of the goal itself. It was entirely

attributable to a loss of confidence across the political spectrum in the ability of the Keynesian

full-employment strategy to actually achieve full employment. It is also clear that this loss of

confidence was itself attributable to the decade’s so-called “stagflation” crises, which laid bare

the Keynesian strategy’s vulnerability to the economy’s growing inflationary tendencies.

The difference now seems to be that the confidence of progressive economists in the

efficacy of the Keynesian strategy—at least for some purposes—has been restored, with

adjustments here and there in the way the strategy is implemented. What’s not clear is the degree

to which this renewed confidence in Keynesian macroeconomic policy is attributable to changed

circumstances—i.e. a dramatic moderation in the economy’s inflationary tendencies—as

Page 8 of 22

opposed to a change in the way progressive economists conceive and define the achievement of

full employment. The reason it’s hard to tell is because most of the economists who refer to 4%

unemployment as full employment don’t explain what they mean by it.

To their credit, Jared Bernstein and Dean Baker are an exception to this rule. In their 2013

book Getting Back to Full Employment: A Better Bargain for Working People, they provide the

following definition of full employment on the first page of the book’s first chapter.

Full employment can be defined as the level of employment at which additional

demand in the economy will not create more employment. All workers who seek

a job have one, they are working for as many hours as they want to or can, and

they are receiving a wage that is broadly consistent with their productivity. The

only people in the labor market not working are the ones who do not have the skill

or ability to work (the structurally unemployed) and those who are between jobs

(the frictionally unemployed). (Bernstein and Baker 2013: 1)

Although it doesn’t match Keynes’s definition(s) of full employment perfectly, Bernstein’s

and Baker’s definition is clearly Keynesian in both substance and methodological style. What’s

most interesting about it, though, is how limited a role it plays in their book. Instead of

estimating the full-employment rate of unemployment (FERU) by assessing the correspondence

of real world conditions to the requirements of their definition, Bernstein and Baker instead rely

on the assumption that the FERU corresponds to the true value of the “non-accelerating-inflation

rate of unemployment” (the NAIRU). For that purpose, they opine, the NAIRU estimates

produced by the Congressional Budget Office are “generally going to be close to the actual full

employment rate (with the exception of the mid-1990s, when CBO’s estimate turned out to be

well above the rate consistent with full employment)” (Bernstein and Baker 2013: 2, note 1).

Why are the FERU and the NAIRU approximately equal to one another? Bernstein and

Baker offer the following explanation.

When demand in the economy can no longer create more employment, where

does the pressure find an outlet? The obvious answer is higher prices, as

purchasers bid up the price of goods and employers bid up the price of workers

(Bernstein and Baker 2013: 2).

The problem with this explanation is not that its premise is false. It’s undoubtedly true that

prices will rise if an increase in aggregate demand can no longer generate an increase in

employment. But that doesn’t mean the converse is true—that an increase in prices necessarily

means that employment can no longer expand in response to increased demand.3 Keynes

explains why very clearly. “Since resources are not interchangeable, some commodities will

reach a condition of inelastic supply whilst there are still unemployed resources available for the

production of other commodities.” Consequently, wages and prices will tend to rise “before full

employment has been reached” (emphasis added). Accordingly, once unit production costs start

to rise in some sectors of the economy, any further increase in effective demand will “generally

speaking, spend itself partly in increasing the quantity of employment and partly in raising the

level of prices. Thus instead of constant prices in conditions of unemployment, and of prices

rising in proportion to the quantity of money in conditions of full employment, we have in fact a

Page 9 of 22

condition of prices rising gradually as employment increases” (Keynes 1936: 296). In other

words, what economists today refer to as the NAIRU is simply the point at which production

costs start their asymmetric rise—before full employment is reached. There is no basis for

believing the FERU and NAIRU equal one another except in a “blackboard economy” whose

features have been carefully specified to achieve that effect.

My purpose here is not to single out Bernstein and Baker for criticism. As noted above, they

should be commended for offering a definition of full employment to accompany their claim that

it corresponds to 4% unemployment. Other progressives who have referred to 4% unemployment

as full employment have been less forthcoming. Why this has been the case is for them to

explain, but the result is confusion as to what they mean when they use the term. The most

reasonable assumption is that they too are conflating it with the NAIRU, in disregard not only of

the commonly understood and legally mandated meaning of the term, but also of Keynes’s

theoretically-grounded definition(s).

The Economy’s Job Gap

What can be learned about the likely level of the FERU from an empirical perspective?

Figure 1 illustrates what I call the economy’s job gap. The bottom line in Figure 1 shows the

number of job openings employers in the United States were seeking to fill on a monthly basis

from December 2000 (the date this data series began) through June 2016. The line immediately

above it, labeled " Official Unemployment," shows the number of jobless individuals who were

actively seeking work in the United States on a month to month basis over the same period of

time. The next line adds "Involuntary Part-Time Workers" to official unemployment. This group

consists of individuals who are working part time but want to work full-time. Finally, the top line

adds to these two groups jobless individuals who say they want a job even though they have not

“actively” searched for one during the preceding four weeks.4 The distance between the bottom

and top lines in this figure provides a rough estimate of the number of additional jobs needed to

provide paid employment for everyone who wants to work. I shall refer to this as the economy’s

“job gap.”

[FIGURE 1 HERE]

The principal source of the “roughness” of this estimate is the fact that it does not

distinguish between full and part time jobs. Involuntary part-time workers have some of the

employment they want, so expressed in terms of full-time-equivalent (FTE) jobs, each of them

needs only some fraction of an FTE job to satisfy their desire for more work. Similarly, not all

unemployed workers are seeking full-time jobs, and the same is true of job wanters who are not

counted as unemployed. On the other hand, the BLS’s Job Openings and Labor Turnover Survey

(JOLTS) does not include information concerning the full or part-time character of reported job

openings, so the number of FTE job openings in the economy is also smaller than Figure 1

suggests. Still, the importance of the figure consists in what it tells us concerning the persistence

of the economy’s shortage of jobs—its failure to satisfy even the most basic pre-requisite for the

achievement of full employment: the availability of enough jobs to fully employ the economy’s

labor force, rather than the exact dimensions of that shortage.

Page 10 of 22

As one would expect, Figure 1 shows that the size of the economy’s job gap mushroomed

during the so-called Great Recession, reaching a peak shortage of 27.8 million jobs in May

2009—the month before the recession officially ended.5 More remarkable is the fact that the job

gap remains sizeable even at the top of the business cycle. In November 2007 (the month before

the official start of the so-called Great Recession) the U.S. unemployment rate was 4.7%, but

there were still about 11.8 million more job wanters in the country than there were job openings.

More importantly for our inquiry, in December 2000, with the national unemployment rate at

3.9%, there were still approximately 8.2 million more job wanters in the country than job

openings.

The existence of this job gap is hard to reconcile with the claim that the economy was at or

approaching full employment in 2000—irrespective of whether that goal is defined in Keynesian

terms or according to the right-to-work/jobs-for-all conception of the goal’s meaning. To cite

once again the Keynesian portion of Bernstein’s and Baker’s proffered definition of full

employment, how is it possible to maintain that the U.S. economy in 2000 was at or approaching

the “level of employment at which additional demand in the economy will not create more

employment”? Or, to cite the right-to-work/jobs-for-all portion of their definition, how is it

possible to maintain that labor market conditions in 2000 came close to satisfying the

requirement that “[a]ll workers who seek a job have one” and “they are working for as many

hours as they want to or can”? The only way the U.S. economy can be described as at or close to

full employment in 2000 is if that goal is equated with the NAIRU, but as explained above, that

assumption is no more consistent with the Keynesian definition of full employment than it is

with the right-to-work/jobs-for-all definition of the goal.

Unemployment in a Job Short Economy

In a job short economy unemployment is as inevitable as the fact that one child is always left

standing at the end of a round of musical chairs. No matter how skilled the children are at

playing the game and no matter how hard they try to win a seat, there will still be one child left

standing after the music stops. “Chairlessness” is built into the game, and the same is true of

unemployment in a job short economy.

There are two important conclusions concerning the achievement of full employment that

flow from this observation. The first concerns the central role played by job availability in

determining the true level of the FERU. The second concerns the difference between

employment policies that affect the level of unemployment and those that affect its distribution.

Job Availability: To achieve full employment in the Keynesian sense of the term the existence

of an adequate number of jobs to employ all willing workers is both a necessary and sufficient

condition for the achievement of full employment. This is so because any unemployment that

remains if enough jobs are available to employee everyone who is not “voluntarily unemployed”

can be attributed to “frictional” factors—a term Keynes used to refer to structural imbalances in

resource allocation as well as the time it takes job seekers and employers to find one another and

conclude a hiring (Keynes 1936: ___). More than that is required to achieve full employment in

the right-to-work/jobs-for-all sense of the term, but adequate job availability is the first

Page 11 of 22

requirement—as William Beveridge’s account of the meaning of full employment quoted above

illustrates.

Unfortunately, the job-openings time series reported in Figure 1 is available only as far back

as December 2000. However, a job-vacancy time series going back to 1923 has been estimated

based on help-wanted advertising data (Zagorsky 1998). Those combined time series suggest that

there have been only 3 years in the past 92 when there were more job openings in the United

States than unemployed workers—from 1943 through 1945 when an average of 11 million prime

age workers were “employed” by the U.S. armed forces, the economy was operating at full hilt

in support of the war effort, and the unemployment rate averaged 1.7 percent.

A number of countries other than the United States have experienced unemployment in the

1% to 2% range for extended periods in the late 1940s through the early 1970s—with some

experiencing it as late as the early 1990s (Martin 1994). In New Zealand the unemployment rate

averaged 0.9 percent throughout the 1950s and 1960s (Denman and McDonald 1996), so it’s

possible the level of unavoidable frictional and structural unemployment can settle down to less

than 1% in economies that manage to achieve sustained full employment and provide effective

support for public labor exchanges.

No country has achieved unemployment that low on a national basis in recent years, but it

does occasionally occur on a subnational level. In 2000, for example, when the U.S.

unemployment rate averaged 4.0 percent on an annual basis, there were 39 counties in the

country that had average annual unemployment rates between 1 and 2 percent. In other words,

that’s still where unemployment rates tend to go when jobs are truly plentiful—even if it is only

on a local level.

This gives us good reason to believe that the FERU is between 1% and 2% and that it has

not risen in recent decades—only that it has become more difficult to achieve because the

NAIRU is higher now than it was in the immediate post-World War II era.

The Level and Distribution of Unemployment: It is generally accepted that cyclical

unemployment is caused by a shortage of jobs. In contrast, the problem tends to be attributed to

other causes at the top of the business cycle. Conservatives tend to blame it on behavioral defects

on the part of jobless individuals who say they want to work. Progressives tend to blame it on the

unequal distribution of employment opportunities. Both conservatives and progressives find

support for their point of view in the experience of individuals and groups who overcome

disadvantages to achieve success, and this has led to broad agreement across the political

spectrum that education and training constitute a highly effective strategy for combating

unemployment.

But Figure 1 suggests that the problem of unemployment has the same cause at the top of the

business cycle that it does during a recession. There simply aren’t enough jobs to go around, so a

certain number of people are destined to suffer unemployment irrespective of how much

collective energy is devoted to pursuing either conservative or progressive solutions to the

problem. Nor is this suggestion inconsistent with the success of those solutions in individual

cases. The problem is that the logic underlying both conservative and progressive policy

responses to the problem of unemployment at the top of the business cycle is deeply flawed by

Page 12 of 22

the fallacy of composition. Because people with more education suffer less unemployment than

people with less education, it is assumed that if everyone got more education, then everyone

would suffer less unemployment. Not so!

This point is immediately comprehensible if we consider whether the problem of

“chairlessness” in the game of musical chairs can be solved by furnishing every child who is left

standing at the end of a round of the game with a package of assistance that combines all the

strategies conservatives and progressives prescribe to combat unemployment at the top of the

business cycle. Motivational encouragement, special instruction in seat-grabbing strategies,

athletic training in moving quickly when the music stops, strict enforcement of anti-

discrimination and anti-bullying rules in the conduct of the game, and so forth. Even if everyone

who is helped in this way succeeds in winning a seat the next time they play the game, it will

have no effect whatsoever on the number of children left standing at the end of a round of the

game.

This does not mean that the strategies commonly proposed for overcoming the problem of

unemployment at the top of the business cycle are a waste of time. It’s just that their effect is

different than we imagine. What Figure 1 suggests is that the level of unemployment in a market

economy is determined by the size of the economy’s job gap across all phases of the business

cycle, not just during recessions. However, the distribution of unemployment among the

individuals and groups comprising the economy’s labor force is determined by all the factors that

conservatives and progressives target via the strategies they advocate for combating

unemployment at the top of the business cycle—differences in education and training, individual

motivation and behavior, geographic disparities in the availability of jobs or public

transportation, employment discrimination, individual needs for employment-enabling services

like child care, and so forth.

If you want every child to have a seat after the music stops in a game of musical chairs, you

have to make sure there are chairs enough for everyone who is playing the game. If you want to

end the problem of unemployment, you have to make sure there are enough jobs to provide work

for everyone who wants it.

Why the Definition of Full Employment Matters?

Some readers may wonder whether it really matters how full employment is defined. Isn’t it

just a semantic dispute? Achieving 4% unemployment is rare enough, and on the rare occasions

it is achieved its salutary effects are significant. Wages rise, working conditions improve,

poverty and inequality decline. Why aim for a target we apparently have failed to achieve in

peacetime in the last 95 years, and likely much longer?

I believe there are four reasons the definition of full employment matters a great deal. First,

claiming that 4% unemployment constitutes full employment tends to stigmatize the millions of

workers who suffer unemployment at the top of the business cycle simply because there aren’t

enough jobs to go around—the very same reason workers suffer unemployment during a

recession. They aren’t unemployed because of behavioral shortcomings, as conservatives claim,

but neither are they unemployed because of the structural barriers to equal employment

opportunity and local economic development that progressives blame. Even the slowest or most

Page 13 of 22

bullied child in a game of musical chairs will find a seat if the number of chairs in the circle

equal or exceed the number of children playing the game, and the same is true of the least skilled

job seeker in an economy with enough job opportunities to provide work for everyone who

wants a job. Referring to 4% unemployment as full employment obscures this enormously

important but rarely recognized fact. And if progressive economists don’t recognize it, who will?

Conservatives will feel affirmed in their jaundiced view of the unemployed not just by the failure

of progressive interventions to succeed in individual cases, but also by the fact that the number

of people needing such assistance tends to stay the same even when the interventions succeed in

individual cases.

The second reason the definition of full unemployment matters is because the use of any

term in a way that differs from its commonly understood meaning is likely to foster

misunderstanding. That’s a serious shortcoming when you’re in the business of explaining

economic conditions to people who rely on your judgment. Economists may not like the fact that

they no longer “own” a term they created, but that’s what has happened to the concept of full

employment. It’s meaning has moved into the public domain and is no longer controlled by what

economists say it means. That doesn’t mean economists are barred from defining and redefining

the concept of full employment however they want. It simply means they should be cognizant of

the term’s commonly understood meaning when referencing the full employment goal, and if

their conception of the term differs from its commonly understood meaning, they should

acknowledge and explain the difference.

The third reason the definition of full employment matters is because the recent tendency

among progressive economists to equate the FERU and the NAIRU inhibits the development of

policies capable of pushing past the latter barrier. When the stagflation crisis shattered the

confidence of progressive economists in the ability of the Keynesian full-employment strategy to

achieve full employment, they should have immediately gone back to the drawing boards to

figure out an alternative or supplemental strategy that could. Their failure to tackle that task had

the practical effect of consigning the study of alternative means of achieving full employment to

the margins of heterodox economic inquiry. The concentrated attention the task deserved—from

progressive think tanks and foundations as well as university-based economists—never

materialized.

Nor was work on alternative strategies for achieving full employment encouraged by the

renewed attention progressive economists accorded the goal in the wake of the dotcom boom and

the Great Recession. Why? Because the progressive economists cited in this article’s first

footnote pronounced the problem solved. If full employment is equated to the NAIRU, we don’t

need new strategies to achieve it. The 4% unemployment achieved during the dotcom boom is

within our reach using conventional Keynesian measures, and the Great Recession showed that

those measures are still useful as an anti-cyclical policy as well. So efforts to develop an

alternative full employment strategy are rendered unnecessary.

The fourth reason the definition of full employment matters is the one suggested by FDR in

this article’s epigraph. It’s because securing the right to work (i.e., the commonly understood

meaning of full employment) is essential for the success of the rest of the progressive reform

Page 14 of 22

agenda. As FDR explained it, of all the rights included in his proposed second bill of rights, the

“one on which the fulfillment of the others in large degree depends, is the right to a useful and

remunerative job.”

The reason for this is not because access to a decent job is more important than other

economic and social rights in and of itself. It’s because of the instrumental role its achievement

would play in facilitating the realization of other economic and social rights. Just as securing free

speech rights plays a crucial facilitative role in securing other civil and political rights, so to

would the realization of the right to work facilitate the realization of other economic and social

entitlements. On an economic level, securing the right to work would reduce the cost to

government of fulfilling the basic needs of all members of society while increasing the resources

available to it government to accomplish that task. Politically, insuring the ready availability of

decent work for everyone who wants it would dramatically reduce the controversy that has

swirled around the provision of aid to the “able-bodied poor” since at least the mid-1300s

(Harvey 1989). Simply stated, securing the right to work would make it both economically and

politically easier to win and sustain public support for the rest of the progressive economic and

social reform agenda.

For all of these reasons, the development of a strategy capable of achieving full employment

in the commonly understood, right-to-work sense of the term deserves the concerted attention of

progressive economists. Those who think the achievement of 4% unemployment is a good

enough macroeconomic goal need only reference the late 1960s, when urban riots, widely

recognized as rooted in the economy’s failure to provide “jobs for all,” swept the country during

a four-year period of under 4% unemployment. Both the limited success of progressive efforts to

achieve equal employment opportunity for all members of society and the erosion of support for

the Democratic Party among white working-class voters since the 1970s, illustrates the

difficulties involved in promoting a progressive reform agenda that is not supported by a credible

strategy for insuring the availability of decent work for all members of society.

In short, FDR had good reason to view the achievement of full employment as the key to

success in advancing the progressive reform agenda. The failure of progressive efforts to enact

full employment legislation at the end of World War II marked the end of the New Deal era. The

civil rights movement succeeded in jump-starting a new era of progressive reform in the 1960s,

but the failure of that effort to include an effective full-employment strategy severely

circumscribed its achievements, and when progressives finally did concentrate on securing the

right to work in the 1970s, their campaign foundered because the stagflation crisis exposed

inadequacies in the Keynesian full employment strategy that they were unable to repair or

ignore.

And so the 1960s and early 70s reform era ended. The neo-liberal era of conservative

retrenchment, persistently high levels of unemployment, and rising inequality took hold; and four

decades later, progressives are still struggling to restart the engines of progressive reform. The

resilience of the neo-liberal era—despite its harsh social consequences—is its most

distinguishing characteristic. It represents the longest period of conservative ascendency in

American politics since the beginning of the industrial age. Progressives tend to blame the

Page 15 of 22

staying power of neo-liberalism on the wealth and power of its principal beneficiaries; but

there’s another candidate for the spoiler’s role—the failure of progressives to develop a credible

full employment strategy—the foundation (to paraphrase FDR) on which the success of the next

period of concentrated progressive reform arguably depends?

Is an Alternative Strategy Possible?

But what if we can’t do any better than 4% unemployment? If driving the rate of

unemployment down to the NAIRU level is as good as we can do, shouldn’t progressives focus

on realizing that goal rather than wishing for what we can’t achieve? Maybe, but what’s certain

is that we are unlikely to figure out if we can achieve genuine full employment as long as

progressive economists continue to muddy the issue by referring to 4% unemployment as full

employment.

It is also important to note that strategies for achieving full employment that take into

account the inflationary tendencies of Keynesian macroeconomic measures have been suggested.

Based on his experience administering price controls during World War II, John Kenneth

Galbraith always believed some form of price administration was necessary to permit Keynesian

policies to achieve sustained full employment (Laguerodiea and Vergarab 2008). Gösta Rehn

and Rudolf Meidner famously advocated a coordinated strategy of restrictive macroeconomic

policies, solidaristic wage policies and active labor market measures to achieve full employment

without sacrificing either price stability or high rates of economic growth (Erixon 2010). More

recently, modern-money theorists working in the post-Keynesian tradition have developed policy

proposals based on Hyman Minsky’s suggestion that a direct job creation program could perform

a price-stabilizing buffer stock function in the labor market while simultaneously achieving full

employment (Mitchell and Watts 1997; Wray 1999). My own contribution to this literature (see,

e.g., Harvey 1989; 2011; 2013a) also relies on direct job creation, but with greater attention to

securing all aspects of the right to work. A recent iteration of this strategy can be seen in full-

employment legislation introduced in the last three sessions of Congress by Rep. John Conyers

(U.S. Congress 2015).

This is not the place to discuss the merits of any of these proposals. My point is simply that

alternative strategies for achieving full employment have been proposed. At the very least they

provide a starting point for investigations into the question of whether and how that goal can be

achieved. It is my hope that the progressive economists criticized in this article can be persuaded

to join in that effort and use their influence to alert progressives to the importance of the task.

Footnotes

1. See, e.g., Bernstein and Baker (2013) (describing “4 percent — the average

unemployment rate for 2000, the last time we were at full employment” as “a reasonable

target, one worth shooting for”); Bivens (2014) (describing the U.S. economy’s

performance in 2000 as an “approximation of genuinely full employment); Boushey,

Mishel and Bernstein (2002) (noting that “The full-employment economy of the late

1990s made a large and positive difference in the growth of real income for low- and

Page 16 of 22

middle-income families”); Krugman (2013) (commenting that that “[t]he US economy in

2000 had really, really full employment — it was an era when labor was so scarce that

McDonald’s was actively trying to recruit senior citizens, when the joke was that you

could get a job as long as your breath would fog a mirror, that is, as long as you were

actually alive”); Kuttner (2012) (noting that “In the late 1990s, when we had full

employment, in one three-year period Social Security’s Year of Reckoning was set back

by eight years, from 2029 to 2037); Madrick (1999) (noting that “little has been better for

the country than a dose of full employment”); Mishel (2015) (stating that “we should

strive to reach genuine full employment” described as “roughly 4 percent

unemployment”); Palley (2007) (describing the distinguishing features of full

employment as “wages systematically rising with productivity” as in the “late 1970s, the

late 1980s, and the late 1990s, when wages started rising with productivity at the tail end

of booms”); Pollin (2012) (“[W]hat do we mean by full employment at decent jobs? Full

employment, in my view, a realistic definition is below 4 percent as officially measured

by the government.”); Stiglitz (2002) (commenting, when asked to describe the benefits

of trade, that “in the United States and in advanced industrial countries, we often say that

trade is good, but imports are bad. We like to export, but we don't like to import. Why?

Because we recognize that sometimes imports lead to job destruction. We recognize that

it is the responsibility, though, of those involved in macromanagement, the Central Bank

-- the Fed in the United States -- to ensure the economy remains close to full

employment. But even when we had unemployment of 3.9 percent, 4 percent, we

worried. We worried about the unemployment that a flood of imports could create, even

though these people could get jobs eventually elsewhere in the economy.”).

2. The term “full employment” does not appear at all in Adam Smith’s Wealth of Nations.

Ricardo used it just once in his Principles of Political Economy and Taxation when he

described laborers being “without full employment” following a war (Ricardo 1912, p.

176). John Stuart Mill used the term three times in his Principles of Political Economy,

always, like Ricardo, with an individual connotation. He described a worker employed to

perform only one task in the division of labor as having “full employment in that

occupation,” economic conditions tending to promote an increase in marriages as

“seasons of cheap food and full employment,” and the low wages workers earned in

certain agricultural districts as deplorable even when they were “in full employment”

(Mill 1871, pp. 81, 210, & 216). Alfred Marshall (1890) did not use the term at all in his

Principles of Economics, and only once in Industry and Trade, referring to the “full

employment” of “a machine” rather than a worker (Marshall 1919, p. 194). A.C. Pigou,

Keynes’s foil in the General Theory used the term only twice in his 1914 treatise on

unemployment, referring to “persons other than those for whom full employment can be

found” and to workers employed in a field of industry in which “full employment could

be accorded to them” (Pigou 1914, pp. 57 and 194). In his Theory of Unemployment

(1933) he used it just once, in referring to a general equilibrium condition in which there

is “no unemployment and no unfilled vacancies anywhere.” Hobson used the term more

frequently (a handful of times in several of his books) but always generically, although in

Page 17 of 22

the Economics of Unemployment (1922: 6, 29, 40, 69 & 149) this generic usage included

references to macroeconomic conditions that are similar to those Keynes later described

as constituting full employment.

3. The logical error committed by Bernstein and Baker in this instance is categorizable as an

example of the propositional fallacy of affirming the consequent.

4. For a jobless individual to be counted as unemployed in BLS statistics, they have to

satisfy the following criteria.

People are classified as unemployed if they do not have a job, have

actively looked for work in the prior 4 weeks, and are currently available

for work. Actively looking for work may consist of any of the following

activities:

Contacting:

An employer directly or having a job interview

A public or private employment agency

Friends or relatives

A school or university employment center

Submitting resumes or filling out applications

Placing or answering job advertisements

Checking union or professional registers

Some other means of active job search

Passive methods of job search do not have the potential to connect job

seekers with potential employers and therefore do not qualify as active job

search methods. Examples of passive methods include attending a job

training program or course, or merely reading about job openings that are

posted in newspapers or on the Internet.

(U.S. Bureau of Labor Statistics 2016). Jobless individuals who say they want a job but

do not satisfy these criteria are deemed to be outside the labor force, but the BLS does

estimate the number of such persons in the economy along with the number who are

either employed or unemployed. And, as Figure 1 illustrates, their number grows and

shrinks over the course of the business cycle in the same way unemployment does.

5. One of the interesting characteristics of the job gap estimates reported in Figure 1 is that

the gap’s peak and trough correspond more closely to the trough and peak of the business

cycle than the unemployment rate. While it is well known that unemployment peaks

sometime after the trough of the business cycle passes, the size of the economy’s job gap,

as measured by the data reported in Figure 1, provides a contemporaneous indicator of

the business cycle’s turns.

Page 18 of 22

References

Bailey, Stephen K. Congress Makes a Law: The Story behind the Employment Act of 1946. New

York: Columbia University Press. 1950.

Bernstein, Jared. “Where Have All the Jobs Gone” New York Times. 3 May 2013. Op. Ed.

Bernstein, Jared, and Dean Baker. The Benefits of Full Employment. Washington DC: Economic

Policy Institute. 2003.

__________. “The Unemployment Rate at Full Employment: How Low Can You Go?” New

York Times 20 Nov. 2013. Economix Blog.

Beveridge, William H. Full Employment in a Free Society. London: Allen and Unwin. 1944.

Bivens, Josh. Nowhere Close: The Long March from Here to Full Employment. Washington DC:

Economic Policy Institute. March 11, 2014.

Clark, John Maurice, et al. National and International Measures for Full Employment. Lake

Success, NY: United Nations. 1949.

Denman, James and Paul McDonald. “Unemployment Statistics from 1881 to the Present.”

Labour Market Trends 104, 1 (1996): 5-17.

Democratic Party. “1972 Democratic Party Platform.” The American Presidency Project.

Archived by John T.Wooley & Gerhard Peters. http://www.presidency.ucsb.edu/platforms.php

________. “1976 Democratic Party Platform.” The American Presidency Project. Archived by