Structural Changes in Male/Female Differences in Unemployment Rates Over Time
and is replicated here as part of the SERC Pedagogic Service.
Summary
In this case, students will learn how shifting time horizons can change the expectations for gaps in the data. They will look at gender differences in the unemployment rate first over the last 30 years and then for the entire post-war period. Students will find that whereas over the shorter time span women have experienced significantly lower unemployment rates than men, the reverse was true prior to 1980. Through peer instruction, students will address each of the following issues:
1. Why is there a divergence in the unemployment rates following recessions?
2. What may have triggered the shift in 1980?
3. Why is it important to specify your time horizon when describing results?
Note that a more advanced statistics class may use this to segue into structural breaks and how to use dummy variables and/or a Chow Test to identify them. On the other hand a principles or labor economics class will probably focus more on the first two issues through a discussion on the structural causes of unemployment variability across demographic sub-groups.
Learning Goals
After the completion of this module students will be able to:
1. Use FRED graphs to perform a visual data comparison.
2. Identify historical gender differences in unemployment rates.
3. Understand the importance of specifying time horizons.
Context for Use
Ideally this module should be directed to an introductory statistics class; however, a more qualitative version can also be used in an economics principles course as well. It is ideally suited as an applied exercise included in a larger lecture. Between the presentation of the data and the ensuing discussion, instructors should allocate 20-30 minutes of classroom time. The discussion could focus on what comprises the unemployment rate, and why the rate is so different among various groups of people.
Prerequisite skills: Graph literacy
Prerequisite concepts: Unemployment Rate; Gender Differentials
Description and Teaching Materials
1. The Statistical (Advanced) Version
Within a lecture introducing time series data graphs, an instructor will use this module to help students understand, "How important is it to specify time horizons?"
The instructor will first present the following graph:
The instructor should point out that the two series included on this graph represent the unemployment rate of men (in green) and of women (in red). Both sets are drawn from the Bureau of Labor Statistics (BLS).
In groups, students can discuss how the graph can help them identify which gender typically recovers faster following a recession (the gray bars).
The instructor should then post the second graph and ask if those conclusions still hold:
Students will find that what they found in the first graph, i.e. that women have experienced faster recoveries than men lately, will be the opposite of what historically has been the case.
The instructor may wish to take the initiative during the peer instruction and point out how recoveries in the 1960s were much steeper for men than for women. That can lead to a question of, "What does the slope measure here?"
The instructor can follow this up with a new discussion question, "What happened in 1980?" or "Why did it flip?" and ask the students to put together an hypothesis for the cause.
Finally, the instructor will be able to point out at the conclusion of the discussion how important it is to specify time horizons up-front because oftentimes what holds true in one period (t) may not necessarily hold true for the entire period (T).
The following is a list of questions that could be asked:
- How can we test for a structural break?
- Why is it important to use as much data as possible?
- Why is it important for statisticians to indicate their time horizon?
- What is the implication of the change in the gap in the unemployment rate?
- When we examine the change in 1980, did women improve their standing or did men lose their standing? Or was it a combination of both?
2. The Principles-Level Version
Within a lecture introducing unemployment, an instructor will use this module to help students understand, "How is the unemployment rate different for men and women?"
The instructor will first present the following graph:
The instructor should point out that the two series included on this graph represent the unemployment rate of men (in green) and of women (in red). Both sets are drawn from the Bureau of Labor Statistics (BLS).
In groups, students can discuss how the graph can help them identify which gender typically recovers faster following a recession (the gray bars).
The instructor should then post the second graph and ask if those conclusions still hold:
Students will find that what they found in the first graph, i.e. that women have experienced faster recoveries than men lately, will be the opposite of what historically has been the case. A follow-up question for the class here at this point would be, "In 1980 we can see that there was a structural change – what do you think caused that shift? And how does that affect men and women today?"
The instructor may wish to take the initiative during the peer instruction and point out how recoveries in the 1960s were much steeper for men than for women but since then, the slope for men has decreased. That can lead to a question of, "Why?" – the intent here is to address the impact of globalization and other factors that impacted male-dominated jobs like manufacturing more so than female-dominated jobs in the service sector.
The following is a list of questions that could be asked:
- Why is it important to use as much data as possible?
- What is the implication of the change in the gap in the unemployment rate?
- When we examine the change in 1980, did women improve their standing or did men lose their standing? Or was it a combination of both?
- Why was there so much divergence before 1980?
- Suppose we compared another two groups of people – college-educated and non-college-educated people. What differences do you think we would see? Do you think there would be a similar structural change? Why?
Teaching Notes and Tips
Be sure that students understand how unemployment rates are calculated. Perhaps give them a brief overview of who is (isn't) counted as being unemployed. An instructor may want to reinforce to students that one of the causes of a decline in unemployment may not actually be economic growth but may simply be people leaving the labor force. Additionally, it will be beneficial to discuss how the rates are different for men and women, college-educated and non-college educated, as well as between whites and minorities.
The following resources should help the instructor develop a bit more background on the issue. Note, however that these are probably too difficult to be used in class and should serve purely as background theory for the instructor:
Using Okun's Law to explain how unemployment varies over a business cycle:
http://research.stlouisfed.org/publications/review/12/09/399-418Owyang_rev.pdf
(Due to an error with the St. Louis Fed website, the link must be copied and pasted into the URL bar)
Racial differences in unemployment during business cycles:
The gender unemployment gap:
http://www.princeton.edu/economics/seminar-schedule-by-prog/macro-s13/Draft-January-2013_4.pdf (link broken)
Assessment
The recommended method for assessment for a statistics class would be to have students write up a short memo where they verify their hypothesis by drawing on outside sources. This should be a take-home assignment.
For a principles class, the instructor should collect the answers to the peer-instruction questions.
References and Resources
FRED Graphs:
1. Full Period: http://research.stlouisfed.org/fredgraph.png?g=coC
2. Truncated period: http://research.stlouisfed.org/fredgraph.png?g=cob
Economic research on the topic:
1. http://www.princeton.edu/economics/seminar-schedule-by-prog/macro-s13/Draft-January-2013_4.pdf (link broken)
2. https://www.newyorkfed.org/research/staff_reports/sr613.html
3. Unemployment Inequalities (Three takes on another study):
- https://themaneater.com/mu-study-illustrates-unemployment-inequalities-bet/
- http://phys.org/news/2012-08-unemployment-gap-persists-women-minorities.html
- http://munews.missouri.edu/news-releases/2012/0807-unemployment-gap-persists-among-women-minorities-white-men-according-to-mu-research/
News articles:
1. It's not just here in the USA: http://www.bbc.co.uk/news/business-18074790
2. Unemployment rates converging for men and women: http://www.huffingtonpost.com/2012/02/03/jobs-report-women_n_1253092.html
3. "Where have all the women's jobs gone?" http://money.cnn.com/2012/04/19/news/economy/women_jobs/index.htm