Citation
Determinants of the Labor Force Participation Rates in the United States

Material Information

Title:
Determinants of the Labor Force Participation Rates in the United States
Creator:
Shapiro, Ashley T.
Publication Date:
Language:
English

Notes

Abstract:
My thesis researched the true factors that determined the labor force participation rate. I researched the all the different types of people in the labor force and found the largest groups. From there, I created an OLS model to find the significance of each group. After seeing my results, I knew there had to be some omitted variable bias. I decided to run a fixed regression model in order to achieve the most accurate results. The fixed effects model increased my R^2 from .75 to .97, meaning the bias existed. I performed a robustness check of my fixed effect regression by allowing for robust standard errors. My final regression helped me determine the three significant factors that determine the labor force; unemployment rate, never married and disability rate. ( en )
General Note:
Awarded Bachelor of Arts, summa cum laude, on May 8, 2018. Major: Economics
General Note:
College or School: College of Liberal Arts and Sciences
General Note:
Advisor: Michelle Phillips. Advisor Department or School: Economics

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Ashley T. Shapiro. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.

UFDC Membership

Aggregations:
UF Undergraduate Honors Theses

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

Ashley Shapiro Thesis Page 1 Determinants of the Labor Force Participation Rates in the United States I. Introduction: The labor force participation rate (LFPR) is measured monthly by the U.S. Bureau of Labor Statistics as a fraction of the number employed and the number of unemployed over the total adult population. The labor force participation rate is one of the main driving factors of the economy: if the labor participation rate decreases, then so does the generation of goods and services (DiCecio). Therefore, the labor force participation rate is a direct reflection of a country, state, or world economys well being. The labor force participation rate can also reflect certain societal changes. For example, since 1950, there has been a dramatic increase in womens labor force participation rate, as well as a less significant decrease in mens labor participation rate. This rising frequency of women in the labor force allows for more diversity and creates mo re job competition. Additionally, from the 1950s to the 1970s, there was an increase in labor force participation across both genders that directly correlates with the baby boomer generation entering the workforce. However, more recent changes could be an effect of the shifting and developing social norms in the modern world, which is closing the gap between men and woman labor force participation rates (Fullerton). The purpose of this study serves to reveal the determinants of labor force participation rat es using regressions from more recent years, 20032015. This study will be demonstrating three models to test the determinants of the labor force participation rate. Model 1: (Total) LRPR = + (Median Age) + (Education Total) + Unemployment Rate) + (Never Married)+ (Disability Rates) + (Minimum Wage) + (Percent of Foreign Born) + (Manufacturing Percentage) + (Construction Percentage) + (Percent Hispanic) + error Model 2: (Men)

PAGE 2

Ashley Shapiro Thesis Page 2 MLRPR = + (Age) + Percent with Bachelors Degree or Higher) + (Marital Status) + (Unemployment Rate) + error Model 3: (Women) FLRPR = + (Age) + ( Percent with Bachelors Degree or Higher) + (Marital Status) + (Unemployment Rate) + error II. Sample: The sample will comprise all fifty states and include the District of Columbia from 20032016. There will be 714 observations total. III. Dependent Variable Labor Force Participation Rate The dependent variable in this study is the labor force participa tion rate for both men and women. As shown in the model, there are many factors that control the decision to enter the labor force. The data was collected from the Bureau of Labor Statistics from the years 2003 to 2016. The model being used for this study contains three different parts. The first is the Total Labor Force Participation Rate, the second is Male Labor Force Participation Rate, and the third Female Labor Force Participation Rate. It is important to consider each variable individually because t he rates differ and yield significant, independent results. IV. Independent Variables: Age The age variable I am using is the median age per state to effectively compare the age differences. This variable will most likely reflect the direct development of gender roles. Older women will presumably be l ess involved in the labor force as they were not given the opportunity when they were younger due to social norms. However, in younger generations, you will see the gender gap closing and more people enteri ng the labor force. That is why looking at the median is important. If the median age is older, we can infer that there are less people in the labor force. Conversely, if the median age is younger, there is a higher likelihood that there will be

PAGE 3

Ashley Shapiro Thesis Page 3 a higher L FPR. Younger males are staying in school longer and older males are retiring earlier and living longer. This is also supported in the Journal of Political Economy by Donald O. Parsons (p.118). In 1948, the rates of male nonparticipation in the labor force from ages 3544 was 2.1% and in 1976 it increased to 4.6%. This change in younger people in the labor force was less significant than the changes in the rates of male nonparticipation in the labor force from ages 5564. In 1948, there was 10.5% male nonparticipation and by 1976 it was up to 25.5%. Education As women begin to achieve degrees in higher education, they are empowered to compete for jobs that were filled solely by men in the past. The introduction of more accessible birth control methods ha s enabled higher educated women to handle a full time career and a family size of choice. 1 Additionally, education will effect the labor force participation rate in a positive way. Educated people are going to have a higher chance of joining the labor for ce and staying in the labor force. The lesser frequency of job turnover of more educated workers, which creates fewer episodes of unemployment, is in large part attributable to more onthe job training (Mincer, 232). Those who are more educated are able to maintain more skilled jobs rather than bounce around minimum wage jobs trying to find a satisfying job. People without college degrees may also end up working in jobs that are hard on their bodies and have to stop working earlier. They may also get lai d off or have a harder time maintaining their daily operations. I will be measuring this variable in Model 1 (Total) using the total amount of people per state who have obtained their bachelors degree or higher and for Model 2 (Men) and Model 3 (Woman) using the percent value of those who have obtained their bachelors degree or higher 1 In the past, household chores were more time consuming, but as our society has become more technologically advanced, household chores have become more efficient. As gender roles begin to change, men could even become stay at home dads, while women are able to enter the workforce. This will be seen by using the two separate male and female LFPR models.

PAGE 4

Ashley Shapiro Thesis Page 4 Marital Status Single men and women are more likely to be in the labor force because they have the need to support themselve s. This variable will measure the amount of singles per state in order to see if there is significance in the relationship between marital status and LFPR. I have chosen to look at statistics on women who have never married. Women who have never been marr ied have to provide for themselves. Furthermore, they may have kids or others to provide for, which means they will continue to stay in the labor force. Those who are divorced, widowed or separated may still be receiving money from other sources. However, women who have never been married still can receive child support. Unemployment Rate The labor force is calculated by adding those employed to those unemployed. If the amount of people unemployed is larger than the amount employed then there will be a higher unemployment rate and a lower labor force participation rate. According to the Labor Force Statistics from the Current Population Survey, unemployment rates have decreased drastically since the 2010 financial crisis, when the unemployment rate was an alltime high. Today, the unemployment rate is at 4.1% and has been decreasing by about .08% per year since 2010. Minimum Wage When the minimum wage increases, it will have a positive effect on the LFPR because more people will be drawn to the market In an article Alan Krueger wrote in the New York Times, he explains that increasing salary does not necessarily mean that they will have to cut jobs in order to increase profits, but companies will be able to fill vacancies (Krueger). As the minimum wag e increases, we expect the LFPR to increase as well. Disability Rates The disability rates will have a negative impact on the labor force. When people are disabled there is a higher chance they will leave the labor force completely or they will retire at an earlier age. According to an Economic News Release from the Burea u of Labor Statistics, in 2016, unemployed people with a disability remained constant from 20152016 and

PAGE 5

Ashley Shapiro Thesis Page 5 those who were not disabled decreased the unemployment rate by 4.6% from 20152016 (Persons). For those who are disabled, it is more difficult to find a job, which can lead to discouraged workers and ultimately leaving the labor force. There are also insurance policies for people with disabilities, which provide them with a sufficient amount of income to work a shorter period of time than those without a disability. Percent of Foreign Born in State The data I am using is from the Percent Foreign Born: ACS American FactFinder, Sex by Age by Nativity and Citizenship Status Community Survey 1year Status. Measure of Industry Composition The effects of industry composition will be measured using the joint effects of manufacturing and construction. It is important to include manufacturing because there seems to be decreasing demand for lower skilled workers in manufacturing due to technolo gical improvements. Manufacturing and construction are the largest sectors of industry composition and will help provide a measure of the fluctuations in the work force. Another finding to note is that after the 2008 financial crisis a large amount of people left jobs from declining manufacturing and construction sectors and have since not reentered the labor force. As confidence levels decreased in the economy, companies were producing less output and therefore did not need as many employees. There is an e xpected correlation between the industry compositions and the LFPR because if people are losing jobs in these sectors, the LFPR is decreasing.

PAGE 6

Ashley Shapiro Thesis Page 6 V. Summary Statistics Table 1 ( Regression Results are from year 2010 and later) Table 2 (Summary Statistics)

PAGE 7

Ashley Shapiro Thesis Page 7 Table 2 (Summary Statistics) (contd) R Squared: According to the summary statistics in table 1, the data were able to explain 75% of the variance in the regression. There were many variables that could have been added to use in order to expl ain the labor force participation rate in the economy. Education: After running my regression, I found that th e pvalue for education was .004. This result was 0.01, meaning it is statistically significant. However, what is interesting about this variable is that is has a negative coefficient, meaning that for an increase in people with an education, there was actually a .003% decrease in the labor force participation rate. This was not what I had originally hypothesized. I assumed it was going to have a significant relationship, but I indeed thought the coefficient would be positive. When people become educated, they demand higher paying skilled jobs. The result is unexpected and it may be due to endogeneity ( possibly omitted variable bias ).

PAGE 8

Ashley Shapiro Thesis Page 8 Median Age: The P value for the median age variable was .986. Becaus e .9 86 0.10 it is not statistically significant and we fail to reject the null and can conclude that there is no linear relationship between the median age and the labor force participation rate. Unemployment: This variable was another variable that came out to be ins ignificant. The P value was .586, which is greater than .10 so we cannot reject the null hypothesis for unemployment, as the results were insignificant. This variable may be influenced by some variation of bias because it would be assumed that there would be a correlation between unemployment and the labor force participation rate. Never Married: Unsurprisingly, the pvalue of the never married variable was 0.000, which means it is high ly significant and we reject the null hypothesis. This means for every one percentage point of woman that are not married, the labor force participation rate increases by .15 percentage points. Disability: This variable was also si gnificant with a P value of 0.000. For every one percentage point increase in those disabled there is a negative coefficient that explains that the labor force participation rate will decrease by 1.7 percentage points. Minimum wage: As expected from my hypothesis there was a significant relationship between minimum wage and LFPR because 0.046 0.10. This means the differences in minimum wage between the states does reflect a difference in the behavior of employees when they are looking for a job. Percent Foreign Born: There is a 0.00 p value for the percent foreign born. As hypothesized, this variable has significance and with every one percent percentage point increase of people who are foreign born in a state, there is a .22 percentage point decrease in the labo r force participation rate.

PAGE 9

Ashley Shapiro Thesis Page 9 Industry Composition: The variables I use to measure for this were construction and manufacturing. For both, we reject the null hypothesis because for manufacturing, I found that .00 0.01 and for construction I found that .004 0.01. However, manufacturing had a lower coefficient than construction did. For every one percentage point increase in construction there is a .18 percentage point increase in the labor force participation rate. Conversely in manufacturing, for every one person increase there is a .91 percentage point increase in the labor force participation rate. Percent Hispanic (Percentage ~ C) : The results for percentage of Hispanics per state had been found to be significant with a pvalue of .003. As hypothesized, there was a negative coefficient related to this variable. As the percentage of Hispanics increase, there is a .05 percentage point decrease in the labor force participation rate. Table 3 (Table Statistics for Men) Table 4 (Table Statistics for Women )

PAGE 10

Ashley Shapiro Thesis Page 10 Model 2: (Men) MLRPR = + (Age) + Percent with Bachelors Degree or Higher) + (UnemploymentRate) + error Model 3: (Women) FLRPR = + (Age) + ( Percent with Bachelors Degree or Higher) + (UnemploymentRate) + error I used two separate models, as seen in table 2, in order to compare sex to labor force participation rate. The mean male labor force participation rate was 71.78 and t he mean for females was 59.82. This can be interpreted as 12% more men than women are par ticipants in the labor force overall. The mean age of men was 36 years old and the mean age for women was almost 39 years old. The most surprising statistic when comparing men and woman was the mean education. The mean number of educated males was 599,404.3 per state, and the mean number of educated females was 630,810.9 per state. That is over a 30,000 person difference. This result was interesting because it would be assumed that since there are more men in the labor force, there would be more educated me n. Finally, the unemployment rate for men was 6.4% and the unemployment rate for women was 5.7%. This variable supports my hypothesis because I assumed there was a smaller population of women in the labor force because there are more women who are housewives and there are also more women who freelance and work from home. It is unclear which variable is responsible for the difference in labor force participation rate between men and women, since the mean education for women was higher and there was a lower unemployment rate. 2 2 I had run separate regressions using bot h the male and female data and in both only the unemployment rate was significant. The r squared for each model wer e about .3, so I dont feel as if the results were necessary to include because there wasnt enough variables in those separate regressions to make the models significant for my research.

PAGE 11

Ashley Shapiro Thesis Page 11 Fixed Effects Model This model will control for the unobservable differences between states and any omitted variables. For example, according to a Forbes list of the most and least technologically advanced states in America, it would be unfair to compare Washington D.C., which is not technically a state but has the most science, technology, engineering and math jobs, to West Virginia where there is the lowest share of technology companies (Bloom). 3I was able to find that there were not as many significant variables as I had anticipated and narrow down my results in order to include only the variables that have the greatest effect on the labor force participation rate. The unemployment rate now had a pvalue that is 0.10 and the coeffi cient is significant at the 90% significance level. Minimum wage was also significan t because its pvalue was 0.049 0.05. The never married variable was significant at the 95th percentile and disabilities was also still significant, but at the 95th pe rcentile. Table 5Fixed Effects Model 1 3 The industry composition variables, manufacturing and constructio n, may have been overseen in the fixed regression model because these factors vary with sophistication across each state. There are certain variables that I did not include in my model because they are automatically a part of the fixed effects model. The i ndustry composition may have been one of these variables that should have not been included in my final model. P E R C E N T A G E H I S P A N I C 3 4 6 2 2 0 1 2 3 5 0 2 6 2 1 4 7 0 1 4 2 8 0 8 9 6 8 4 1 1 6 5 2 8 2 C O N S T R U C T I O N P E R C E N T A G E 4 9 5 7 7 9 9 3 5 4 7 6 6 1 1 4 0 0 1 6 3 2 0 2 7 2 6 9 1 1 9 4 2 8 7 M A N U F A C T U R I N G P E R C E N T A G E 1 5 7 6 7 8 2 3 3 4 5 5 5 5 0 4 7 0 6 3 8 5 0 1 0 3 5 5 8 1 6 3 9 1 9 P E R C E N T F O R E I G N B O R N 2 7 5 8 8 9 8 2 1 0 8 6 1 9 1 3 1 0 1 9 2 6 9 1 0 6 0 5 1 3 9 2 8 0 8 M I N I M U M W A G E 2 3 9 5 9 4 7 1 2 1 2 6 7 7 1 9 8 0 0 4 9 0 0 0 8 2 8 1 4 7 8 3 6 1 3 D I S A B I L I T Y R A T E S P E R C E N T A G E S 5 3 5 6 7 1 2 1 5 0 0 4 8 1 3 5 7 0 0 0 0 8 3 1 1 0 4 3 2 4 0 2 3 8 2 N E V E R M A R R I E D 3 2 2 6 5 4 5 1 1 2 3 3 8 3 2 8 7 0 0 0 4 5 4 3 8 3 9 8 1 0 1 4 6 9 2 U N E M P L O Y M E N T R A T E T O T A L 1 4 0 6 9 0 1 0 7 6 0 4 5 1 1 8 5 0 0 6 5 0 0 9 0 3 6 8 2 9 0 4 1 7 E D U A C T I O N T O T A L 0 0 0 1 6 2 2 0 0 0 3 7 7 5 0 4 3 0 6 6 8 0 0 0 5 8 1 1 0 0 0 9 0 5 5 M E D I A N A G E 0 3 7 5 8 7 5 2 0 4 2 8 3 1 0 1 8 0 8 5 4 3 6 4 6 2 9 9 4 3 9 8 0 5 L F P R C o e f S t d E r r t P > | t | [ 9 5 % C o n f I n t e r v a l ] T o t a l 5 0 2 9 4 9 0 2 2 3 2 1 1 5 6 6 8 1 9 3 8 R o o t M S E = 8 0 9 6 A d j R s q u a r e d = 0 9 5 8 2 R e s i d u a l 1 7 4 3 5 0 3 5 1 2 6 6 6 5 5 4 5 2 4 4 8 R s q u a r e d = 0 9 6 5 3 M o d e l 4 8 5 5 1 3 9 8 7 5 5 8 8 2 7 5 2 7 0 3 P r o b > F = 0 0 0 0 0 F ( 5 5 2 6 6 ) = 1 3 4 6 8 S o u r c e S S d f M S N u m b e r o f o b s = 3 2 2

PAGE 12

Ashley Shapiro Thesis Page 12 Fixed Effects Model (additional robustness check) To further my results, I ran one last regression with robust errors (to correct for heteroskedasticity and autocorrelation). By doing so, I was able to narrow my results down to three variables that had the greatest effects on the labor force participation rate. The unemployment rate is 0.096, which is less than 0.10, and significant at the 90th percentile level. Never married, had a pvalue of .028, which was les s then or equal to .05, so it is significant at the 95th percentile. Lastly, Disabilities was also significant a t the 95th percentile with 0.015 0.05. Table 6 Fixed Effect Model 2 VI. Conclusion The results from the fixed effect model with the additional robustness check highlight three main variables that supported the determinants of labor force participation rate; unemployment rate, never married and disability rate. Unemployment rate was one o f the independent variables that I P E R C E N T A G E H I S P A N I C 3 4 6 2 2 0 1 2 0 9 3 4 7 7 1 6 5 0 0 9 9 7 5 8 4 0 9 5 0 6 5 9 6 9 3 C O N S T R U C T I O N P E R C E N T A G E 4 9 5 7 7 9 9 3 6 9 0 1 4 1 3 4 0 1 8 0 2 3 0 7 8 0 1 1 2 2 2 3 4 M A N U F A C T U R I N G P E R C E N T A G E 1 5 7 6 7 8 2 3 4 3 8 9 0 8 0 4 6 0 6 4 7 5 1 9 4 1 6 1 8 3 4 7 7 2 4 P E R C E N T F O R E I G N B O R N 2 7 5 8 8 9 8 2 1 6 0 4 4 6 1 2 8 0 2 0 3 7 0 1 2 6 4 8 1 4 9 4 8 5 2 M I N I M U M W A G E 2 3 9 5 9 4 7 1 1 8 1 4 1 9 2 0 3 0 0 4 4 0 0 6 9 8 2 5 4 7 2 2 0 6 9 D I S A B I L I T Y R A T E S P E R C E N T A G E S 5 3 5 6 7 1 2 1 6 0 6 7 4 2 3 3 3 0 0 0 1 8 5 2 0 2 6 2 2 1 9 3 1 6 3 N E V E R M A R R I E D 3 2 2 6 5 4 5 1 1 5 8 5 2 6 2 7 9 0 0 0 6 5 5 0 7 5 9 3 0 9 4 5 4 9 7 U N E M P L O Y M E N T R A T E T O T A L 1 4 0 6 9 0 1 0 7 6 6 9 9 3 1 8 3 0 0 6 8 0 1 0 3 2 4 8 2 9 1 7 0 5 E D U A C T I O N T O T A L 0 0 0 1 6 2 2 0 0 0 2 6 4 2 0 6 1 0 5 4 0 0 0 0 3 5 8 0 0 0 6 8 2 4 M E D I A N A G E 0 3 7 5 8 7 5 2 0 9 7 9 9 9 0 1 8 0 8 5 8 3 7 5 4 9 2 1 4 5 0 6 6 7 2 L F P R C o e f S t d E r r t P > | t | [ 9 5 % C o n f I n t e r v a l ] R o b u s t R o o t M S E = 8 0 9 6 R s q u a r e d = 0 9 6 5 3 P r o b > F = 0 0 0 0 0 F ( 5 5 2 6 6 ) = 2 9 9 5 2 L i n e a r r e g r e s s i o n N u m b e r o f o b s = 3 2 2

PAGE 13

Ashley Shapiro Thesis Page 13 had expected to have an effect on the labor force participation rate. The OLS model had the unemployment rate as insignificant. It was necessary to rerun the model in order to correct this error even if some of my significant variables then became insignificant. In both fixed effects models the unemployment rate was significant. The coefficients and p values are slightly different than the FE regression with robust standard errors shown immediately above Thus the R squa red values are slightly different too. T he R squared value from this model (OLS with Dummies FE model and Robust Errors) was 0.965 compared to the Standard OLS R squared value 0.749 and the OLS with Dummies FE model R squared value 0.965. The R squared of both FE models is 0.965 as compared to the original OLS R squared value of 0.749.

PAGE 14

Ashley Shapiro Thesis Page 14 VIII. Sources Bloom, Laura Begley. The 10 Most (And 10 Least) Innovative States In The U.S. Forbes, Forbes Magazine, 3 Oct. 2017, www.forbes.com/sites/laurabegleybloom/2017/03/28/the 10most and 10least innovative states in the us/#81fb7fb10a64. Christofides, Louis N, et al. The Impact of Immigration on Unemployment, Labour Force Par ticipation and Part Time Employment in Cyp rus. Cyprus Economic Policy Review, vol. 3, no. 1, pp. 5184., doi:10.1787/662360140770. DiCecio, Riccardo, et al. "Changing trends in the labor force: A survey." REVIEW FEDERAL RESERVE BANK OF SAINT LOUIS 90.1 (2008): 47. Fullerton Jr, Howard N. "La bor force participation: 75 years of change, 195098 and 19982025." Monthly Lab. Rev. 122 (1999): 3. Krueger, Alan B. The Minimum Wage: How Much Is Too Much? The New York Times, The New York Times, 9 Oct. 2015, www.nytimes.com/2015/10/11/opinion/s unday/the minimum wage how much is toomuch.html. Mankiw, G. N. (2014). Principles of macroeconomics cengage learning write experience 2.0 powered by myaccess. Place of publication not identified: SouthWestern. Mincer, Jacob. Education and Unemployment. NBER Working Papers Series, National Bureau of Economic Research Sept. 1991. Parsons, Donald O. "The decline in male labor force participation." Journal of Political Economy 88.1 (1980): 117134. Persons with a Disability: Labor Force Characteristics Summary. U.S. Bureau of Labor Sta tistics, U.S. Bureau of Labor Statistics, 21 June 2017, www.bls.gov/news.release/dis abl.nr0.htm. U.S Census Bureau, Hi spanic or Latino Origin 20102016, American Community Survey 1Year Estimates, Using American Fact Finder; Table B03003. U.S Census Bureau, Place of Birth for the Foreign Born Population 20102016, American Com munity Survey 1Year Estimates, Using Amer ican Fact Finder; Table C05006. U.S Census Bureau, School Enrollment 20052013 American Community Survey 1 Year Esti mates, Using American Fact Finder; Table S1401. U.S Census Bureau, Selected Social Characteristics in the United States 20102016, Ameri can Community Survey 1 Year Estimates, Using American Fact Finder; Table DP02.

PAGE 15

Ashley Shapiro Thesis Page 15 U.S. Department of Labor, Expanded State Employment Status Demographic Data, 20032016, Last Modified: 10 Oct 2017. Local Area Unemployment Statistics, NE Washington, Cur rent Population Survey. U.S Census Bureau, Total Full Time and Part Time Employment by NAICS Industry 20032016, American Community Survey 1Year Estimates, Using American Fact Finder; Table S1810. U.S Department of Commerce, Bureau of Economic Analy sis, Total Full Time and Part Time Employment by NAICS Industry, Table SA25N U.S Department of Labor, Office of Communications, Wages and Hour Devision (WHD), Changes in Basic Minimum Wages in Non Farm Employment Under State Law: Selected Years 1968 to 2016, Modified: Dec. 2016.