This preview shows page 1. Sign up to view the full content.
Unformatted text preview: Unemployment I
Lecture 11, March 27, 2012 Announcements
• Final Exam:
– A-LI TUE 17 APR AM 9-11 SEEL
– LL-Z TUE 17 APR AM 9-11 SHER • Final Exam Structure: Three Questions
– “Define and explain the significance”
– Two theoretical/quantitative problems • Tutorials continue this Friday
– Alfia will be taking up selected problem set questions
– Last tutorial is this Friday
– Alfia’s Office Hours Posted • Q&A Session (by me)
– Monday, April 16th, 3:00 to 5:00, SS2106 1 Outline
• Introduction to unemployment: excess supply of labour
• Measuring unemployment:
– What would we like to measure?
– How does Statistics Canada measure unemployment using the Labour Force
– The historical experience of unemployment in Canada
– Unemployment as a summary of “waste”
– Unemployment as a summary of “misery” • Labour Market Dynamics
– Stocks and flows of labour
– Incidence and duration of unemployment • International Comparisons
– Canada versus the United States
2 Unemployment: Introduction
• Unemployment is an important topic in labour economics, though the
degree of interest varies with the business cycle.
– Policy relevant, especially for long-term unemployed, and links to income
– Design of social insurance programs (like Employment Insurance, links to
Social Assistance, Public Pensions, etc.)
– Policies that affect labour markets more generally (payroll taxes, income taxes,
regulations, trade, industrial relations). • To what extent does the existence of unemployment inform us about the
relevance of the competitive, market-clearing “supply and demand” model
of labour markets?
– Does it need to be “tweaked” to accommodate unemployment as a market
failure? 3 Unemployment as excess supply
• Unemployment as excess supply of labour, at a “wrong” or rigid market
LD L* LS L
4 Overview of Questions
• In this setting, there is disequilibrium in the labour market, the wage fails
to adjust downward, and supply exceeds demand at the prevailing wage.
– Why is the wage “stuck”? • Our focus over the next two weeks:
• How is unemployment measured?
– What does it mean?
– What does it hide, or rather, what is hidden?
– Does Statistics Canada use this diagram to (at least indirectly) measure
unemployment? • What do theories of unemployment say?
– Does unemployment occur in equilibrium, or because of wage rigidity?
– What may explain wage rigidity? 5 Who are the unemployed?
• In order to be classified as unemployed, there are three criteria:
– Available for work
• Not full-time students, “housewives”, disabled, retired. Must be in the labour force. – Actively seeing work. Cannot just be wishing for work, even if discouraged.
– Not working • “Unemployment” is intended to be a distinct economic state
• Thought experiment; Assume that after graduation, you could obtain the
following job offers:
– Tim Horton’s ($20,000 per year)
– Stats Canada Phone Surveyor (say, $35,000 per year)
– Marketing department of Proctor & Gamble (say, $65,000 per year) • Who would accept one of these job offers, without of any guarantee of
receiving another offer?
– Are you unemployed if you turn down the Tim Horton’s job? 6 Excess supply at the going wage
• In the data, it will be impossible to distinguish between people with
different search objectives.
• Clearly, however, there are differences of reservation wages at which
people will be willing to work (accept a job).
• Implicit in the definition of unemployment is “no jobs” at the prevailing
wage (as opposed to any old wage), or at the individual’s reservation wage.
• So unemployment does not mean:
– There are no jobs;
– A person is underemployed
– No job improvement is desired. 7 Unemployment in the LFS
• The primary source of data for measuring unemployment in Canada is the
monthly Labour Force Survey (56,000 households)
• We can look at the key questions.
• To be considered unemployed by Statistics Canada, an unemployed person
must satisfy the above conditions more precisely:
– Available for work (exclude students, etc.)
– Actively searching (during the past four weeks)
– Not working (even one hour) • The fraction of the labour force that is actively seeking, but not working, is
the unemployment rate.
• Also add:
– Temporary layoffs;
– Waiting to start a new job in 4 weeks or less
8 The key LFS Questions
• Key Questions:
• LFI_Q101 — Last week, did ... have a job or business from which he/she
• LFI_Q190 — Could he/she have worked last week [if he/she had been
recalled/if a suitable job had been offered]?
• LFI_Q170 — In the 4 weeks ending last Saturday, [date of last day of
reference week], did ... do anything to find work?
• LFI_Q171 — What did he/she do to find work in those 4 weeks? Did he/
she do anything else to find work?
– [Public employment agency, private employment agency, union, employers
directly, friends or relatives, placed or answered ads, looked at job ads, “other”] • LFI_Q172 — As of last week, how many weeks had he/she been looking
for work? [since the date last worked]
9 495 CHAPTER 16: Unemployment: Meaning, Measurement, and Canada’s Experience Unemployment in Canada, 1921-2010
F IGURE 16.1 Unemployment in Canada, 1921–2010 20 Unemployment rate (%) 15 The national unemployment
rate is plotted by year over
the period 1921–2010.
dramatically during the Great
Depression, and fell to very
low levels during World War
II. During the postwar period,
unemployment has fluctuated cyclically, initially around
a generally rising trend. In
the past three decades,
this upward trend has been
reversed. 10 5 0
1921 1927 1933 1939 1945 1951 1957 1963 1969 1975 1981 1987 1993 1999 2005 2010
SOURCE: Historical Statistics of Canada (1921–1945), Statistics Canada CANSIM database. How well does “Unemployment” capture
• There are alternative measures of employment that may better reflect what
is going on in the labour market.
– Labour Force Participation
Employment Growth • The division of the population into three distinct states (E, U, N), while
carefully done, has an arbitrary edge to it.
• There may employed individuals who work very few hours, or there may
be people who are not in the LF, but who want to work (“discouraged
– These “marginally attached” individuals may be part of the “hidden
unemployed.” 11 PART 6: Unemployment Labour Force Participation, Employment,
and Unemployment,a Canada 1946-2010
Labour Force Participation, Employment, and Unemployment,
Canada, 1946–2010 Year Labour Force
Participation Rate Employment
Rate 1946 55.0 53.1 3.4 1956 53.5 51.7 3.4 1966 55.1 53.1 3.6 1973 59.7 56.4 5.6 1981 65.0 60.0 7.6 1989 67.2 62.1 7.5 2000 65.9 61.4 6.8 2008 67.7 63.5 6.1 2010 67.0 61.6 8.0 Rate of Growth
a All statistics are based on authors’ calculations from data on the civilian labour force (annual averages) derived from the
Civilian employment; growth rates are averages of compound annual increases from the level in the year in the row one
line above to the level in the year one line below.
SOURCE: Authors’ calculations based on data from the Statistics Canada CANSIM database. Unemployment as a “summary” of misery
• The simple unemployment rate may fail to capture some of the degree of
hardship occurring in the economy, or it may overstate it.
– Does the unemployment rate mean the same thing now, as it did in 1930? • One must be very careful in attaching a welfare interpretation to this
– Does an unemployment rate of 5% always imply the same degree of hardship? • For example, incidence and duration of unemployment will have different
• For this reason, Statistics Canada has several other unemployment
measures that address some of the possible shortcomings of the official
• While these measures change the level, they do not change the time-series
13 Supplementary Measures of Unemployment,
Measure Definition Rate R1 Counting only those unemployed one year or more 0.9% R2 Counting only those unemployed three months or more 3.1% R3 Made comparable to the U.S. Official Rate 7.1% R4 Official Rate 8.0% R5 Official Rate plus discouraged workers 8.2% R6 Official Rate plus those waiting for recall, waiting for replies,
and long-term future starts 8.7% R7 A measure of both unemployment and underemployment
(involuntary part-time employment) expressed in full-time
equivalents 10.6% R8 Official Rate plus those in R5, R6, and R7 11.3% 14 Labour Market Dynamics
• Even more than hiding nuances of hardship, the unemployment rate at a
point in time hides the dynamic, churning nature of the labour market.
• In an important (though not exciting!) paper, Stephen Jones of McMaster
used linked Labour Force Surveys to follow individuals for the 6 months
that they are in the LFS.
• This allows us to explore dynamics:
– How long are people unemployed?
– What are the flows between the major states, U, E, N? • Summarized in the following figure:
– The absolute levels of flows between states are remarkably high;
– The probabilities of movement between states are also high;
– So while the stocks of unemployment change only marginally month to month,
there is actually a lot of change. • Also note, the many paths to “unemployment.”
15 501 CHAPTER 16: Unemployment: Meaning, Measurement, and Canada’s Experience Labour Market Stocks and Flows,
F IGURE 16.3 Labour Market Stocks and Flows, 1976–1991
(0.04) Not in the labour
(0.03) This figure illustrates the
large flows that occur each
month between the labour
force states of employment
(E), unemployment (U), and
not in the labour force (O).
On average, every month
workers obtained jobs (i.e.,
moved from U to E), and
190,000 workers lost or left
jobs and joined the pool of
job searchers (i.e., moved
from E to U). The average
net monthly flow was thus
235,000 2 190,000
5 45,000 from U to E. The
probability of an unemployed
worker becoming employed
the following month equals
0.22, and the probability
that an employed worker
equals 0.02. Unemployed
1084 NOTES: 1. All numbers are in thousands.
2. All stocks and flows are averages of monthly values from 1976 to 1991.
SOURCE: Stephen R.G. Jones, “Cyclical and seasonal properties of Canadian gross flows of labour.” This article first 16 PART 6: Unemployment Decomposition of Unemployment by
Decomposition of Unemployment by Reason, 2000–2010
Reason for Separation
Rate (%) Job Losers (%)
(%) New Entrants
(%) 2000 6.8 29 7 12 38 2001 7.2 29 8 11 39 2002 7.7 29 7 10 41 2003 7.6 30 7 10 40 2004 7.2 28 7 11 41 2005 6.8 28 7 10 42 2006 6.3 26 8 11 42 2007 6.0 27 7 11 43 2008 6.1 29 7 11 44 2009 8.3 33 8 8 40 2010 8.0 30 6 6 42 NOTES: 17 1. All statistics are based on authors’ calculations from data on the civilian labour force (annual averages).
2.. Components do not sum to the total because reasons for separation “Reason unknown”and “Future starts” are omitted.
SOURCE: Statistics Canada, Labour Force Historical Review, Catalogue 71F0004, 2010. Incidence and Duration
• The distinction between incidence and duration of unemployment is also
very important to understanding the nature of unemployment.
• Incidence of unemployment: The proportion of individuals who become
unemployed in a given period (e.g., a month);
• Duration of unemployment: The length of time spent in the unemployed
state, before either obtaining a job, or dropping out of the labour force (e.g.,
the number of months a person is unemployed).
• In a steady state: UR = I × D
• Note that the units of measurement must be compatible.
– Incidence (I) is usually measured as the proportion (or percentage) of the LF
that becomes unemployed in a given month;
– Duration (D) is usually measured as the average number of months of time
spent unemployed (completed spells).
18 Incidence and duration: Example 1
• For a given month, if the unemployment rate is 10%, and the incidence of
unemployment is 2%, then in a steady state, the average duration is
10/2 = 5 months.
– Individuals who are unemployed are unemployed for an average of 5 months. • Of course, there will be hidden variation.
– e.g., for 5 unemployed individuals (4 each could be unemployed for 1 month,
and 1 person for 26 months)
– This would lead to an average duration of 30/5 = 6 months. 19 Incidence and duration: Example 2
• As another illustration of a steady state, consider an economy with 100
Duration: 1 2 3 1 10 10 10 2 5 5 5 3 0 0 0 Total: 15 15 15 • In any given period, there are 15 unemployed people, and the
unemployment rate is 15%
– In any month, the probability of unemployment (Incidence) is 10%
– The average duration is ( 5 × 1) + ( 5 × 2 ) 10 = 1.5 ( ) 20 Variation in incidence and duration
• When the media report a change in the unemployment rate, it could be due
to a change in incidence or duration.
• There are two interesting dimensions in which to study variation of
incidence and duration.
• In a cross-section, there are big differences in the pattern of unemployment
– Young people have much higher incidence, but lower duration;
– For older workers, the pattern is reversed. • In a time-series, there are strong cyclical patterns to incidence, and
especially, duration. 21 CHAPTER 16: Unemployment: Meaning, Measurement, and Canada’s Experience Incidence and Duration of Unemployment,
by sex, Canada 2010
Incidence and Duration of Unemployment by Age and Sex,
Canada, 2010 Unemployed
> 6 Months
(%) Sex Age
(Months) Men 15–24 17.1 7.4 2.3 9.1 25–44 7.6 2.0 3.7 23.2 45 1 6.9 1.5 4.5 33.0 All 8.7 2.6 3.3 22.3 15–24 12.4 5.9 2.1 7.6 25–44 6.9 2.1 3.2 21.2 45 1 5.5 1.4 4.0 28.3 All 7.2 2.4 3.0 19.7 Women NOTES: The incidence rate is calculated as the percentage of the labour force that has been unemployed between 1 and 4 weeks. The average duration is calculated from the formula UR 5 I 3 D, given the incidence and unemployment rates in the
first two columns.
SOURCE: This analysis is based on the Statistics Canada CANSIM database http://cansim2.statcan.gc.ca, Tables
2820002 and 2820048, 2011. All computations, use and interpretation of these data are entirely that of the authors.
22 2.3 percent. More commonly, incidence of unemployment is measured as a fraction of the
labour force, so that it measures the percentage of the labour force that is newly unemployed. Incidence and Duration of Unemployment,
60 Unem. Rate / Months of Duration 12 50 10 40 8
20 4 10 2
1975 1980 1985
Duration 2000 Percentage Unemployed > 14 Weeks 14 0
2005 > 14 Weeks 23 Summary of key “facts”
• The labour force is highly dynamic, with large flows into and out of
unemployment each period;
• Less than half the flow into unemployment in an average year is due to
individuals losing their jobs; the remainder is associated with job leavers,
new entrants, and re-entrants.
• Even in 2010 (a time of above-normal unemployment), the average
duration of unemployment was approximately three months, with only one
fifth of all unemployment spells lasting more than six months.
• The age groups with the highest unemployment rates have the shortest
average unemployment durations, but the highest incidence of
• Not all unemployment spells end in employment.
• Heterogeneous experience with unemployment.
24 International Comparisons
• Understanding the causes of, and the potential policy responses to,
unemployment can benefit by international comparisons.
• Especially given the clear importance of labour market dynamics, we can
see the importance of “long term” unemployment.
– The short-term fluctuations of employment are less of a problem. • What causes long-term unemployment?
– What models explain long term unemployment?
– What role is played by labour market institutions? 25 International Differences in Long-Term
Unemployment PART 6: Unemployment TABLE 16.4 International Differences in Long-Term Unemployment
as a Percentage of Total Unemployment
2005 2009 6 Months
and Over 12 Months
and Over 6 Months
and Over 12 Months
and Over Canada 17.2 9.6 18.0 7.8 France 61.2 42.5 55.3 34.7 Germany 71.0 54.0 61.8 45.5 Japan 49.1 33.3 46.3 28.5 Sweden 37.3 18.9 29.4 12.8 United Kingdom 38.2 22.4 44.8 24.6 United States 19.6 11.8 31.5 16.3 SOURCES: OECD, Employment Outlook 2006 and Employment Outlook 2010 (Paris: OECD, 2006 and 2010). to Europe, unemployment rates in the United States during most of the 1990s and 2000s
were very similar to those experienced in the 1960s and 1970s (see Exhibit 16.1). Relative to
the United States, Europe had gone from being a low-unemployment region to being a high- Unemployment and Canada and the U.S.
• One specific comparison of particular interest is Canada versus the United
• Until about 1980, the U.S. and Canadian unemployment rates were
essentially identical (see Figure 17.3).
• After 1980, the rates diverged, with the Canadian rate usually much higher
than the U.S.
Was is it due to differing industrial sectors?
Or a less vibrant Canadian labour market?
Government policies in Canada or the U.S.? • Consider more recent data on the next slide:
– LFS (Canada) and CPS (U.S.), men from 1975-2010. 27 .04 Unemployment Rate
.12 .14 Unemployment Rates (Men): Canada vs. U.S. 1975m1 1980m1 1985m1 1990m1 1995m1
U.S. 2000m1 2005m1 2010m1 Canada 28 Differences in Definitions of Unemployment
• To convert the Canadian data to a comparable U.S. definition of
– Remove 15 year olds.
– Remove people who looked for work only by using job ads. The U.S. does not
include such ’passive job-searchers among the unemployed.
– Remove people who did not look for work, but who had a job to start in the
next four weeks. In Canada, these ’future starts’ are counted as unemployed.
– Remove those unavailable to take a job because of per- sonal or family
responsibilities. In Canada, they are considered among the unemployed; in the
U.S., no such exception is made.
– Add full-time students looking for full-time work. In Canada, they are not
included among the unemployed; in the U.S., they are included. • In any given month, these adjustments normally shave almost one full
percentage point from the Canadian unemployment rate.
29 Employment-Population Ratio
.72 .74 Employment-Population Ratios (Men): Canada vs. U.S. 1975m1 1980m1 1985m1 1990m1 1995m1
U.S. 2000m1 2005m1 2010m1 Canada 30 Other explanations of the gap
• Differences in measurement account for a significant part of the gap
• Other explanations:
– For any non-employed worker, a Canadian is more likely to report being
• Perhaps due to unemployment insurance? – Duration of unemployment in Canada is longer.
• Perhaps due to unemployment insurance? – Incidence of unemployment declined in the U.S. (through 2005) • Another interesting comparison:
– Men versus women: “Mancession”
– But note that the recovery is also concentrated among men. 31 .04 Unemployment Rate
.1 .12 Unemployment Rates (U.S.): Men versus Women 1975m1 1980m1 1985m1 1990m1 1995m1
Male 2000m1 2005m1 2010m1 Female 32 .06 Unemployment Rate
.12 .14 Unemployment Rates (Canada): Men versus Women 1975m1 1980m1 1985m1 1990m1 1995m1
Male 2000m1 2005m1 2010m1 Female 33 Next Week
• While measurement can yield many important insights, a deeper
understanding of unemployment, and the potential role of policy, entails
exploring theoretical explanations.
• Chapter 18
– Except the “macro” stuff (pp. 565-573) • Problems from Chapter 17:
– None. 34 ...
View Full Document
- Spring '11