The Canadian Employment Research Forum (CERF) has a mission of fostering discussion of labour market issues between government and researchers in the academic and other communities. In the past, the principal vehicle has been the sponsoring of research and conferences on a specific topic. This has worked well, but requires substantial resources and time.
The CERF Board discussed and decided to undertake a much simpler activity, namely a workshop of invited participants, who would share their current knowledge and research.
This is a small Workshop for which no "call for papers" is issued and a small group of people with expertise to share in the area are invited to discuss an issue. The Workshop has the very simple purpose of determining if there are research gaps in the area or if the "profession" is comfortable with its knowledge base about the issue.
The Workshop Report can be a synthesis or summary of the findings, with the intent of making it available to interested CERF members and others. If the Workshop concludes that there are major research gaps, then CERF may consider sponsoring a Conference at some later date in this area.
The topic for the first CERF Workshop was "Participation Rates and Employment Ratios: What Do We Know?" The Workshop was held on December 17, 1997 from 1 PM to 5 PM in Ottawa. It was organized by two members of the CERF Board - Mike McCracken and Andrew Sharpe.
The behavior of the participation rates have been somewhat of a puzzle in the last seven years, with substantial declines as in the 1981-82 recession, but without a return to pre-recession levels as occurred at that time. Is it because the economy is still cyclically weak? Or have structural changes occurred that will permanently depress participation rates henceforth?
Pierre Fortin has also commented on the declines in the employment ratio in Canada, with no bounce-back evident in this recession. The behavior is quite distinct from recent US experience. What is happening?
Indeed, these two characteristics should be linked, since the participation rate and the employment ratio differ only by the "unemployment ratio". In other words, is part of the explanation for the lower participation rate, traceable to the lower employment ratio? Is the employment ratio only a cyclical phenomenon? Or, are there structural elements as well? Is the causal flow from EMPLOYMENT RATIOS (ER) to PARTICIPATION RATES (PR) or are the two influenced by common factors?
Introduction of Participants and Current Work
1. WHAT ARE THE FACTS?
There has been a decline in the employment ratios for most age-sex groups. The
ratios in the US have bounced back - not here.
There has been a decline in participation rates since 1990 and little recovery
to date.
2. WHAT ARE SOME OF THE EXPLANATIONS?
3. WHAT RESEARCH HAS BEEN DONE?
4. DO WE "KNOW" WHAT HAS HAPPENED?
5. IMPLICATIONS FOR THE OUTLOOK AND POLICY.
6. FUTURE STEPS
The list of participants and
their coordinates are provided in an Appendix to this Report.
Information that was provided ahead of time or at the meeting is noted in an
Appendix.
With the limited time available, the approach used to gather information from the group was a technique described as a "Rice Storm" (from Michael Machalko, Thinkertoys (Ten Speed Press, Berkeley) 1991, p. 308 ff.). In the first iteration, participants identified various dimensions of the "problem" by recording a brief observation or question (challenge) on 3x5 cards, one per card. These were then collected, recycled, and grouped into logical headings by the participants. ( The outcomes of this cycle are detailed below in Section 3 and Section 4.)
In the second iteration, solutions based on current knowledge or areas for further research were identified on a second set of cards, and the classification exercise repeated. These findings are summarized in Section 5 and Section 6.)
The session closed with a discussion of the possible specification of participation rate equations and employment ratio equations (Section 7 and 8).
A draft of this Report was circulated to all participants, who were encouraged to check the draft for accuracy and to add their views in Section 9 - Principal Conclusions. As well they were encouraged to respond to the questions in Section 10.
Is the distinction between cyclical and structural components useful? Or do we find a "RESIDUAL" after these calculations?
Any analysis using the term "cyclical" usually has embodied in it some notion of "potential output", which in turn assumes some level of full employment or full capacity or both. What is the best measure of the output gap? How far should monetary policy go before freezing the economy again? Does this depend on the fiscal policy settings?
What about the structural problems? What are these influences and what can be done about such structural problems?
Are provincial differences compatible with the identified structural differences? Can these major structural trends be identified and explained at the provincial level?
What about the age-sex dimensions of these structural factors? Are there different influences on young, old, etc.
What role has the minimum wage rate played in the shortage of student jobs and non-student jobs?
Should wages of youth be rising relatively more rapidly, given that their supply (the source population) has dropped in the 1980s by 20%? Employment and participation rates rose rapidly for youth during the 1980s. Now youth population is rising and immigration is up. Are the youth participation rates of the 1980s the peaks that we will ever see?
One role of the unemployment rate is as an explanatory variable in a wage equation. What might be a better variable or set of variables in light of recent behaviors?
Several of the studies tabled
examined the trends in the data over time, using various smoothing techniques.
The emerging patterns raised some questions.
Why is the structural trend for youth different in Canada and the US?
Why has the trend increase in female part rates stopped in the 1990s? Will it
resume in the future? Since US core part rates have risen, shouldn't we see the
same thing here once demand for labour (that is relevant for females) rises
further?
What is driving the decline in male employment rates, and to a lesser extent
part rates for males 25-54? This seems to be happening in many countries.
What do the different trends between the US and Canada tell us?
Do different regions (provinces) behave differently? Why?
Can cross-sectional information enrich the understanding of changes? Should
age-specific groups be analyzed in greater depth? Do different cohorts behave
differently?
What factors may cause a
reversal in the decline of participation rates (PR) and employment ratios (ER)
for those 55 and over? And what might cause the female, 55+ to shift up over
the net decade?
Why have the declines occurred mainly in the "young" and
"old" groups? Are these the "flexible" groups? Or is this
the way our institutions work?
What role do private pensions play? If pension surpluses (that resulted from
high real rates in 1980s) are being "given away" to today's seniors
in order to allow them to be replaced with younger workers or to avoid layoffs
of younger workers, then what happens when the surpluses disappear (if, for
example, real interest rates are finally reduced to low levels)?
More generally, what are the effects of pension and other non-labour income on
the participation rates of older people (55+)?
Will the "baby-boom" cohort retire at the same age as the current 60+
generation? Earlier? Later?
What impact does medicare have on PR and ER? Effect of move to US-type system?
We may need a broader framework
for analysis that starts with a view that people may be engaged in a variety of
ways - of which "paid work" or "looking for paid work" are
only two of many possible categories.
Can the term "not in labour force" (NILF) be treated as the
"residual" anymore?
What is the role of second-best considerations in the decision to do other
things rather than seek work?
· Child rearing
· Voluntary Activities
· Social Action
· Illegal or Underground Activities
What are the long-term effects
of the recent Canadian "catch-up" with the US in education levels for
young workers? Although education participation may lower youth PR initially,
it probably raises their PR and ER when they emerge from the educational
system. Any evidence yet?
Are youth developing their capital or are they engaged in
"consumption"? That is, are they "parked" because they are
"discouraged"? Is this a sustainable shift or one-time effect? If
consumption activity, then income and price elasticities may matter more. Can
the individual distinguish between "consumption" and
"investment"?
What will be the effect of increases in the costs of education being borne by
students? Or are these offset by rising rates of return to education in a high
unemployment world?
Are there long-term adverse
effects associated with delayed labour market experience among the young (i.e.,
scarring)?
Are there interactions with the unemployment insurance system (UI) or welfare
(social assistance of SA) that may be affecting these rates? Are SA and UI
substitutes or complements?
Is the decline in the PR a transitory adjustment to the advent of the knowledge-based economy?
Net immigration is contributing half of source population growth. Age-gender distribution is close to "host" population distribution. But human resource qualities, propensities to participate, education, fertility, etc. may be different. Also what are the time lags for adoption of "domestic" behavior by immigrants? Has this changed over time as immigrant group mix has changed? This suggests that immigration flow (or stock?) should be part of equations seeking to explain aggregate or detailed PR.
Should employment ratios be
analyzed as is, or adjusted for hours worked? In particular, should part-time
and over-time be used to decrease or increase these ratios? (Alternatively,
shares of part-time and over-time workers could be added as explanatory
variables. Differential productivity levels might also be included.)
What is the marginal contribution of employment rate changes to changes in the
participation rate?
How should changes over time be decomposed? By fixing source population shares
or participation rates at a historical value?
Differentials in education between Canada and the US may explain some of the PR differences, particularly for older people. In Canada educational attainment is much lower for 55+ than in the US, as are the PR. As younger group ages this difference should shrink.
Different cyclical performance should allow a test of this factor between Canada and the US. (Also could use different regions in both countries.) Note that in the US, as in Canada, some rural regions have very high unemployment rates.) Indeed, a case study comparing US and Canadian rural characteristics for some region might be useful.
Canada and the US provide different regimes regarding social programs immigration levels, race, and educational differences, in addition to the cyclical differences.
The overall fall in participation rates largely reflects the fall of youth and older male participation rates. (KNOW)
Self-employed tend to be older people. LFS counts them as employed at usual hours, even if those hours are seasonal or periodic in nature. (KNOW)
PR declines are explained in part by discouragement. (MAY KNOW)
How long does it take for PR
transitions?
How does the combination of better educated but less experienced cohort
entering the labour force affect productivity growth and real wage
developments?
What are rates of return faced by students and non-students?
What is the impact of self-employment on PR? On ER?
PR at a point in time is less than PR during the year. What percentage of source population are part-year participants? How meaningful is their withdrawal? Behaviorally, subjectively?
What are the regional aspects of
the decline in youth PR?
Is the decline in Youth PR a good thing or a bad thing? The "good"
presumably is associated with further education and life skills; the
"bad" with lack of experience, income, and frustration.
Is it appropriate to use five-year groupings of 15-19, 20-24 or should
subgroups be distinguished? For example, 15 and 16 year olds are mostly in
school and enrollment rates are high and stable at 95%. But 17-19 may behave
more like 20-23 group. We should re-evaluate the appropriate divisions. Perhaps
we need young core, old core, young seniors and senior seniors.
Apart from minimum wage rates we don't have good data on wage rates. Nor is this well disaggregated by age and sex, region, and educational attainment. Settlement data are too restrictive (base rates for establishments of 500+). Other sources are affected by composition changes. Relative wages are likely important in decisions to seek work.
Average minimum wage in Canada is 42.5% of average industrial wage in Canada and 40.5% in US. How many more jobs would be created (or lost?) in Canada if the minimum average ratio were reduced to the US level? (Would it matter if this was done by raising the average industrial wage or lowering the minimum wage?)
Statistics Canada shows no major increase in the number of discouraged workers in the 1990s despite the high unemployment rates and fall in the PR.
We do not know how much of PR declines are attributable to discouragement. The key research issue is the actual probability of job-finding, given status of NILF, U, Discouraged Workers in previous period.
More work needs to be done on the "discouraged" - their attitudes towards work, their response to economic conditions.
Does macro performance tell all
about the decline in ER? What about the growth of part-time? Extensive
over-time, both paid and unpaid?
Will youth PR and ER recover with the reduction in the gap?
Are structural factors affecting ER different in Canada and the US during the
1990s? If so, why?
How high would PR be if ER returned to 62% in the next three years?
If the Bank of Canada output gap estimate is 1.75% in 1997 QII, then an ER of
59.5% is the equilibrium non-inflationary level. Is this believable?
UI effects - is the income
effect mainly affecting youth? Is household formation affected, with more
"doubling" or "staying in the nest", which in turn can
affect non-youth (e.g., parents, siblings)
With UI restrictions impacting on PR and SA rates, what are the welfare
implications of recent social policy changes.
By how much have UI restrictions affected the equilibrium unemployment rate?
Illiteracy - effect on ER and PR.
"Attitude" Issue - Do we see, in gross entry/exit flows, more
changing of occupations, establishments/firm ties, etc. Is this helping better
allocate resources? Does this "adjustment" occur for all age/sex
groups or differentially?
LFS needs to survey persons not
in the labour force (NILF) more systematically. They cannot be treated as a
residual anymore because they are not all retired, homemakers, or full-time
students.
Should we rethink the concept of labour market attachment? Perhaps a new
definition of "participation" and "employment".
There needs to be a formal
investigation of the determinants of participation.
Disaggregate by (ijk), by age, sex, and region. Also possibly by status prior
to entry (NILF, URATE, in school), education level, urban and rural, gender
roles.
Employment ratio (estimated elasticity), UI effects (benefits, coverage), expected return of being in labour market
It was noted that the number of discouraged workers as measured by the LFS is relatively small. One participant reversed the question as, "Why do we show relatively few discouraged workers in areas of high unemployment and low participation rates (e.g., Atlantic Canada)?". Is the problem the measure we use?
In reviewing the draft session summary, it was pointed out that more emphasis might have been placed on the key issue of distinguishing between cyclical and structural effects and their relevance for potential output, the output gap, and appropriate monetary conditions. It was suggested that there be a simple decomposition approach, with "complexities" added or other approaches compared.
Sins of "omission" included:
· failure to link declines in participation rates to wage developments, particularly in light of the decline in real wages since 1989
· failure to consider wealth effects on participation rates
· failure to analyze immigrant PRATES separately
· failure to consider both the short-term and long-term impacts of the many UI/EI changes on participation rates.
Some people liked the Rice Storm
approach, others "hated" it. It was suggested that more
"classical" forms be used, even if this required a full day rather
than a half-day format.
Other areas that might benefit from a Workshop approach include:
· Effects of wages and wealth on labour markets
· Effects of immigration on labour markets
· Effects of UI/EI changes on labour markets
· Sorting out cyclical and structural dimensions of labour markets.
Andrew Sharpe and Louis Grignon organized a session at the Quebec City meeting of SCESE on May 7 and 8 and two sessions at the Learned Societies in Ottawa on May 29-31. These are a direct result of the Workshop and the papers presented there.
|
Informetrica Limited |
|
HRDC |
· HRDC Applied Research Bulletin, Excerpt on Youth Participation Rates
· Robert Leore, What You Don't Know.... (see particularly p.22-34 and 34 on.)
· Irene Ip, Recent Patterns of Participation Rates: A Canada-US Comparison (Bank of Canada)
· Mary MacGregor and Kurt Mang, Long-term Participation Rate Projections (Finance Canada)
· Jeff Rubin and John Lester, Unemployment in Canada: Understated and Not Falling Fast (CIBC Wood Gundy)
· Informetrica Limited - 2 pages from Briefing Packet
· Informetrica Limited - Historical Background, text and tables re Labour Markets (yellow pages)
· Informetrica Limited - August 1997 National Reference Forecast re Labour Markets (text and tables)
· Benoit Robidoux and Bob Dugan, "What Happened to the Canadian Participation Rate?" (Finance Canada)
· Deborah Sunter, "Trends in Participation Rates: the role of school, retirement, and discouragement" (Statistics Canada)
· Pierre Fortin, Figures and Regression Results
· Richard Archambault and Louis Grignon, "The Decline in the Youth participation Rate: Structural or Cyclical?" (Applied Research Branch, HRDC)
· Richard Archambault and Mario Fortin, "The Beveridge Curve and Unemployment Fluctuations in Canada" (Applied Research Branch, HRDC, W-97-4E))