The Effects of Graduate-Student Unionization

The impact unionization has on remunerations, including wages and health benefits, has on teaching and research assistants in the United States.

Transcript

1. The Effects of Graduate-Student Unionization Tom Schenk Jr. Department of Economics Iowa State University
2. History of Unionization
3. Graduate-student unions include teaching assistants and, sometimes, research assistants. Begun in 1969, grew quickly in the 1990s and early 2000s. Approximately 32 unions in 2006. Not permitted at private universities. HISTORY
4. Approximately 0 unions in 1968. Graduate-student unions include teaching assistants and, sometimes, research assistants. Begun in 1969, grew quickly in the 1990s and early 2000s. Not permitted at private universities. HISTORY
5. Graduate-student unions include teaching assistants and, sometimes, research assistants. Begun in 1969, grew quickly in the 1990s and early 2000s. Approximately 10 unions in 1995. Not permitted at private universities. HISTORY
6. Graduate-student unions include teaching assistants and, sometimes, research assistants. Begun in 1969, grew quickly in the 1990s and early 2000s. Approximately 18 unions in 2000. Not permitted at private universities. HISTORY
7. Graduate-student unions include teaching assistants and, sometimes, research assistants. Begun in 1969, grew quickly in the 1990s and early 2000s. Approximately 32 unions in 2006. Not permitted at private universities. HISTORY
10. Causes of Unionization
11. “ graduate students have been taking longer to complete their degrees and spending more time as graduate assistants...Meanwhile, outside monies from the federal government has decreased...Graduate students, particularly in the humanities, are facing a tighter job market…universities are cutting back on the number of tenured faculty positions being offered, and some real wages have fallen. (p. 7) ”
12. graduates in the humanities, and to a lesser extent, social sciences, have been hit hardest… (p. 11) “ ”
16. Longer time-to-degree Higher debt burden More work Less faculty pay Fewer tenure positions CAUSES
17. Opposition to Unionization
18. Graduate assistants are not employees. Unions interfere with faculty-student relationships. Higher stipends will increase costs for others (e.g., undergraduate tuition). May decrease the number of available assistantships.
19. Graduate assistants are not employees. Unions interfere with faculty-student relationships. Higher stipends will increase costs for others (e.g., undergraduate tuition). May decrease the number of available assistantships.
20. Graduate assistants are not employees. Unions interfere with faculty-student relationships. Higher stipends will increase costs for others (e.g., undergraduate tuition). May decrease the number of available assistantships.
21. Graduate assistants are not employees. Unions interfere with faculty-student relationships. Higher stipends will increase costs for others (e.g., undergraduate tuition). May decrease the number of available assistantships.
22. What Do Unions Do?
23. The standard industrial model suggests unions withhold labor supply for better wages, lower distribution of income, and new policies. Unions can improve productivity by creating a better work atmosphere.
24. Economic Effects
25. Hypotheses Data Set Econometric Model Results: Stipends Results: Health Care Results: Wage Distribution EFFECTS
26. Unions will increase the union-nonunion wage gap. Wage increases will be disproportionately higher for humanities and social sciences. Therefore, the intra-university wage distribution will decrease. Unions will increase the probability of receiving health benefits
27. Unions will increase the union-nonunion wage gap. Wage increases will be disproportionately higher for humanities and social sciences. Therefore, the intra-university wage distribution will decrease. Unions will increase the probability of receiving health benefits
28. Unions will increase the union-nonunion wage gap. Wage increases will be disproportionately higher for humanities and social sciences. Therefore, the intra-university wage distribution will decrease. Unions will increase the probability of receiving health benefits
29. Unions will increase the union-nonunion wage gap. Wage increases will be disproportionately higher for humanities and social sciences. Therefore, the intra-university wage distribution will decrease. Unions will increase the probability of receiving health benefits
30. Hypotheses Data Set Econometric Model Results: Stipends Results: Health Care Results: Wage Distribution EFFECTS
31. DATA The Chronicle of Higher Education TA and RA Stipends for 2000-01; 2001-02; and 2003-04 were collected. Forty-five universities in 2000-01 and 2001-02 surveys from the Association of American Universities…eighty-three from “leading research” universities in 2003-04. In total, 101 universities, 25 over all three years. Data on health-care coverage for students and dependents. Averages for biology, economics, English, history, mechanical engineering, and sociology
32. DATA Integrated Postsecondary Education Data System Institution characteristics such at enrollment size; endowment; and cost-of-living. National Science Foundation Department ranks from 1995 Coalition of Graduate Employees & newspapers Union status
33. University Biology Economics English History Mechanical Engineering Sociology Wealth Ratio Cost-of-Living Undergraduate Tuition Cost Private or Public Union Status Average Stipends Department Rank Health-care coverage
34. University Biology Economics English History Mechanical Engineering Sociology 2000-01 2001-02 2003-04
35. Hypotheses Data Set Econometric Model Results: Stipends Results: Health Care Results: Wage Distribution EFFECTS
36. University A Biology Economics English History Mechanical Engineering Sociology 22% ) corr(  A A Soc , , Bio  Observations within universities are not independent
37. Ordinary least-squares (OLS) will increase Type I errors. Artificially high number of observations lead to a higher critical value. Standard errors will be higher.
38. CV n , * i i t t ˆ 1 SE      Assume the t-statistic is exactly equal to the critical value But, the standard error and critical value are higher, so, CV n , CV n , i i i i t t ˆ ˆ 1 1 SE SE          Thus, CV n , i i t ˆ 1 SE    
39. Alternative econometric models are OLS with robust errors and hierarchical models. OLS w/ robust errors corrects the symptoms, but does not model the issue. Multilevel models are more ideal since they explicitly model two “levels” of data.
40. University Biology Economics English History Mechanical Engineering Sociology Level 2 Level 1
41. ij k k m k ij x y        1 Start with a standard OLS model: for universities i in department j, but levels may vary with the university. So let the intercept vary: ij k k m k i ij x u y         1 ) (
43. Dependant variables of interest: Stipends Health-care coverage Wage variance (OLS)
44. Hypotheses Data Set Econometric Model Results: Stipends Results: Health Care Results: Wage Distribution EFFECTS
45. Teaching Assistants, Log of Stipends
46. Teaching Assistants, Log of Stipends
47. Years Organized? Contract Union 0.23 * 0.09 * Noncontract Union 0.07 0.06 Years Organized -0.04 * Years Org. Sq. 0.001 *
50. Variable Coefficients Std. Error Economics -0.034 0.024 English -0.046 * 0.025 History -0.069 ** 0.025 Engineering -0.013 0.024 Sociology -0.054 *** 0.026 Annual 0.338 ** 0.031 Contract Union 0.086 *** 0.037 Non-contract Union 0.061 0.051 Department Rank 0.000028 0.00037 Private 0.11 0.11 COLA Log 0.094 0.078 Tuition Log 0.005 0.058 Wealth Ratio 5.7E-10 5.7E-7 Intercept 8.421 * 0.881 Random Intercept 0.106 * 0.013 10% Sig: * 5% Sig: ** 1% Sig: *
51. Variable Coefficient Std. Error Economics -0.033 0.024 English -0.045 * 0.025 History -0.07 ** 0.025 Engineering 0.013 0.024 Sociology 0.053 *** 0.025 Annual 0.338 * 0.031 TA Union 0.122 *** 0.048 TA+RA Union 0.08 *** 0.037 Non-contract Union 0.063 0.05 Rank 0.000038 0.00038 Private 0.105 0.11 COLA 0.096 0.078 Tuition Log 0.008 0.058 Wealth Ratio 5.4E-10 5.76E-08 Intercept 8.367 ** 0.885 Random Intercept 0.107 0.014 10% Sig: * 5% Sig: ** 1% Sig: *
52. Research Assistants, Log of Stipends Research
53. Variable Coefficient Std. Error Economics -0.168 ** -0.04 English -0.291 ** -0.043 History -0.33 ** -0.043 Engineering -0.156 ** -0.036 Sociology -0.188 ** -0.04 Annual 0.258 ** -0.037 Contract Union -0.028 -0.058 Non-contract Union 0.032 -0.076 Rank -2.20E-04 -0.001 Private 0.236 -0.175 COLA -0.073 -0.132 Tuition Log -0.042 -0.089 Wealth Ratio -1.7E-08 *** -8.79E-08 Intercept 10.54 * -1.457 Random Intercept 0.154 -0.009 10% Sig: * 5% Sig: ** 1% Sig: *
54. Research Assistants, Log of Stipends
55. Variable Coefficient Std. Error Economics -0.168 *** 0.04 English -0.292 *** 0.043 History -0.331 *** 0.043 Engineering -0.156 *** 0.036 Sociology -0.188 *** 0.04 Annual 0.257 *** 0.038 TA Union 0.013 0.08 TA+RA Union -0.047 0.065 Non-contract Union 0.033 0.077 Rank -2.10E-04 0.001 Private 0.219 0.177 COLA -0.065 0.132 Tuition Log -0.033 0.09 Wealth Ratio -1.7E-07 8.80E-08 Intercept 10.38 ** 1.505 Random Intercept 0.156 ** 0.014 10% Sig: * 5% Sig: ** 1% Sig: *
56. Hypotheses Data Set Econometric Model Results: Stipends Results: Health Care Results: Wage Distribution EFFECTS
57. Teaching Assistants Research Assistants +0.02 (0.576) N.A. +0.21 (0.788) N.A. Contract Union Non-contract Union Student Health Benefits, Logit (p-value)
58. Variable Coefficient (TAs) p-value Economics -0.035 0.29 English -0.096 ** 0.05 History -0.047 0.172 Engineering -0.034 0.344 Sociology -0.029 0.423 Annual -0.033 0.336 Contract Union 0.024 0.576 Rank -0.002 * 0.009 Private -0.364 0.610 COLA 0.060 0.307 Tuition Log 0.043 0.672 Wealth Ratio 0.000 0.438 10% Sig: * 5% Sig: ** 1% Sig: *
59. Variable Coefficient (RAs) p-value Economics 0.015 0.620 English -0.113 * 0.053 History -0.074 0.183 Engineering -0.021 0.512 Sociology 0.000 0.991 Annual -0.022 0.561 Contract Union 0.021 0.788 Rank -0.002 *** 0.010 Private -0.483 0.604 COLA 0.124 0.310 Tuition Log 0.074 0.654 Wealth Ratio 0.000 0.192 10% Sig: * 5% Sig: ** 1% Sig: *
60. Teaching Assistants Research Assistants +0.16 (0.341) 0.44 (0.036) +0.043 (0.794) 0.520 (0.025) Contract Union Non-contract Union Dependent’s Health Benefits, Logit (p-value)
61. Variable Coefficient (TAs) p-value Economics 0.011 0.816 English -0.016 0.701 History 0.021 0.618 Engineering -0.059 0.244 Sociology -0.008 0.881 Annual -0.127 0.300 Contract Union 0.166 0.341 Non-contract Union 0.441 ** 0.036 Rank -0.002 0.412 Private -0.036 0.929 COLA 0.305 0.471 Tuition Log -0.127 0.542 Wealth Ratio 0.000 0.938 10% Sig: * 5% Sig: ** 1% Sig: *
62. Variable Coefficient (RAs) p-value Economics 0.211 0.018 English 0.211 ** 0.037 History 0.160 * 0.086 Engineering 0.085 0.222 Sociology 0.147 0.110 Annual 0.177 0.156 Contract Union 0.043 0.794 Non-contract Union 0.520 ** 0.025 Rank -0.002 0.347 Private 0.148 0.784 COLA 0.261 0.479 Tuition Log -0.170 0.453 Wealth Ratio 0.000 0.330 10% Sig: * 5% Sig: ** 1% Sig: *
63. Hypotheses Data Set Econometric Model Results: Stipends Results: Health Care Results: Wage Distribution EFFECTS
64. Measures of wage variance: Standard deviation of wages Low-to-High ratio Coefficient of variation (st. dev/mean)
65. Contract Union Non-contract Union -87.96 (0.045) -0.028 (0.048) +1112.7** (483.5) +0.079** (0.038) Standard Deviation Low-to-High Ratio Dependent’s Health Benefits, OLS (std. error) +0.077 (0.087) -0.082 (0.069) Coefficient of Variation
66. Tallying the Effects
67. Economic Effects Distribution of Wages Social Organization Monopoly Face Collective Voice union-nonunion wage gap Does not harm relationships (Gordon 2000) ? ? ?/ intra-university wage gap wages health-care
69. Around 6% returns to unionization for TA’s. No increase in health benefits for students or dependents. No effect on distribution of wages. Unclear effects on the cost of the university
70. Around 6% returns to unionization for TA’s. No increase in health benefits for students or dependents. No effect on distribution of wages. Unclear effects on the cost of the university
71. Around 6% returns to unionization for TA’s. No increase in health benefits for students or dependents. No effect on distribution of wages. Unclear effects on the cost of the university
72. Around 6% returns to unionization for TA’s. No increase in health benefits for students or dependents. No effect on distribution of wages. Unclear effects on the cost of the university.