We loaded the March 2014 U.S. Census Current Population Survey into a MySQL database for our students at MIT. In showing them how painful it is to calculate medians in standard SQL we discovered a larger-than-expected female/male wage gap. We limited our results to young people, age 22-36. They all had the same level of education: a Bachelor’s degree (no more, no less). They all worked at least 30 hours per week. Within the sample, men generally worked 2-6 hours more per week (e.g., 44 hours versus 40). We’re going to make the virtual machine available, probably from the Three-Day RDBMS course page if you want to poke around in the data (you can try right now from the Day 1 problems). Note that we started with a CSV file from the National Bureau of Economic Research. I’ve seen various statistics on male/female wage gaps. Supposedly the trend is that when corrected for working hours, years in the work force, education, etc., there isn’t much of a gap. This is as you’d expect from classical economics. If women were truly cheaper to employ, adjusted for skills and productivity, Target could put Walmart out of business simply by hiring only women. How does the theory translate into practice?
| CPS F | CPS M | Cents/Dollar | |
| Alabama | $40,000 | $47,000 | 85 |
| Alaska | $43,600 | $50,000 | 87 |
| Arizona | $30,000 | $50,000 | 60 |
| Arkansas | $42,000 | $57,000 | 74 |
| California | $40,000 | $55,000 | 73 |
| Colorado | $41,600 | $45,000 | 92 |
| Connecticut | $35,000 | $60,000 | 58 |
| Delaware | $46,000 | $40,000 | 115 |
| District of Columbia | $47,000 | $55,000 | 85 |
| Florida | $40,000 | $50,000 | 80 |
| Georgia | $38,000 | $43,004 | 88 |
| Hawaii | $43,000 | $45,000 | 96 |
| Idaho | $31,000 | $53,000 | 58 |
| Illinois | $40,000 | $50,000 | 80 |
| Indiana | $35,000 | $43,680 | 80 |
| Iowa | $38,000 | $67,000 | 57 |
| Kansas | $46,000 | $50,000 | 92 |
| Kentucky | $40,000 | $45,000 | 89 |
| Louisiana | $32,828 | $63,000 | 52 |
| Maine | $28,000 | $46,500 | 60 |
| Maryland | $37,000 | $51,000 | 73 |
| Massachusetts | $35,000 | $45,000 | 78 |
| Michigan | $37,000 | $60,000 | 62 |
| Minnesota | $45,000 | $58,000 | 78 |
| Mississippi | $25,000 | $38,670 | 65 |
| Missouri | $40,000 | $50,000 | 80 |
| Montana | $28,000 | $40,000 | 70 |
| Nebraska | $36,000 | $50,000 | 72 |
| Nevada | $38,000 | $50,000 | 76 |
| New Hampshire | $42,000 | $60,000 | 70 |
| New Jersey | $43,000 | $53,000 | 81 |
| New Mexico | $35,000 | $40,000 | 88 |
| New York | $47,000 | $53,000 | 89 |
| North Carolina | $37,000 | $36,000 | 103 |
| North Dakota | $35,000 | $39,000 | 90 |
| Ohio | $40,000 | $45,000 | 89 |
| Oklahoma | $30,000 | $40,000 | 75 |
| Oregon | $30,000 | $43,700 | 69 |
| Pennsylvania | $40,000 | $50,000 | 80 |
| Rhode Island | $40,000 | $60,000 | 67 |
| South Carolina | $34,000 | $50,000 | 68 |
| South Dakota | $38,000 | $50,000 | 76 |
| Tennessee | $37,000 | $42,000 | 88 |
| Texas | $41,400 | $50,000 | 83 |
| Utah | $40,000 | $62,500 | 64 |
| Vermont | $38,000 | $45,000 | 84 |
| Virginia | $40,000 | $74,000 | 54 |
| Washington | $41,804 | $55,000 | 76 |
| West Virginia | $28,000 | $60,000 | 47 |
| Wisconsin | $48,000 | $45,000 | 107 |
| Wyoming | $40,000 | $55,000 | 73 |
| average | 77 |
Note that the sample size is fairly small for the CPS. There are 224 women in California in our restricted sample and 210 men, for example. So that could account for a fair amount of variance but I don’t think it can explain the overall pattern.
Related:
- August 2014 posting on whether public companies should be allowed to hire men
- December 2012 posting about the profit opportunity to be exploited if the wage gap is real
- January 2015 posting questioning whether looking at the wage gap makes sense (given that spending ability is what most people care about)
Update: In response to reader comments that some the differences might be due to marital status, I ran the queries again on the 22-36-year-old population of Americans with Bachelor’s degrees. The results were limited to people working at least 30 hours per week and earning at least $5000 per year. The median earnings of women went up from 77 percent of men’s earnings to 86.7 percent. Note that the virtual machine with all of the data loaded is available so readers can do their own poking around (given minimal knowledge of SQL).
Bbbbut Hilary told us that there is a War on Women (TM) and that women make 77 cents on the dollar (and that 1 in 5 college women are raped). Is it possible that she is lying just to gin up female votes? I can’t imagine that any party would tell such cynical lies just to get votes.
Izzie: If you look at the last line of the above data you’ll see that the average across all states for this age/education group is in fact 77 cents on the dollar!
There are lies, damn lies and statistics. Once you adjust for # of working hours, experience, etc. the gap disappears.
Could it be career choice? Are women pursuing degrees/jobs in fields that pay less than degrees/jobs men pursue?
My theory is that women discount the amount that the need to work based on the assumption they will be able marry into an economic relationship that allows them to consume 90-100% of the average man’s income anyways. Based on the declining marriage rate, this may be an increasingly poor assumption. A woman may want to just dedicate her twenties to securing an excellent, high paying job and then going to the sperm bank for some high quality asian genes.
iirc, the main difference isn’t between men and women, it’s between MARRIED women, especially married women with kids – and everyone else, male or female.
Just look at any middle to upper middle class suburban neighborhood and the answer should be pretty clear. It does not make sense for the wife to take on a high paying and stressful job as the husband. With high cost of child care and marginal tax rate at 50%+, the wife has to stress and earn at least $100k+ to make any meaningful impact to the household economics. It’s much better to just relax, stay healthy and care for the children. My wife gave up her promising career in a heartbeat after having our baby and is currently enjoying her “retirement”. Me, not quite as lucky…
Forgot to mention that President Obama is trying to help the situation by giving a $500 tax incentive for both parents to return to work. Intention is good, but missing the mark by one to two orders of magnitude.
I’m with Dan Maas, and not just for the Dutch connection. We know from the other war on women being waged right now (or is it the same one?) that men outnumber women in STEM. For this theory to work we would have to assume of course that STEM pays more on average than non-STEM related work.