In his proposal to include American Indians in our workforce initiatives, Rep. Steve Hickey says, "South Dakota celebrates a low unemployment rate only because tribal unemployment statistics are excluded. The real unemployment rate in South Dakota which include tribal unemployment are much more dire."

Logically, a commenter asks, "Does anyone know what the unemployment rate would be in SD if the reservations were included?"

Good question. For what I think is the answer, skip straight to the table at the bottom.

But if you want some wonky caveats, plow through the following exposition:

Let us establish that South Dakota does not count unemployed Indians in its job stats because counting Indians is the feds' job. And even the feds don't exert themselves very hard on that task. The Bureau of Indian Affairs issued a report on Indian job data in 2007, then skipped its biennial legal mandate to do so in 2009 and 2011. They caught up last January with data from 2010:

The law mandating the report doesn’t provide funding for it. One full-time employee – an economist – was assigned the task of producing the report....

But it’s a question of priorities, according to [Assistant Secretary for Indian Affairs Kevin] Washburn. Although the report is “not unimportant,” Washburn said he didn’t want to pull human resources from other important aspects of the BIA’s mission “and I’m not sure Congress would like us to...."

So why not hand the task over to the Census Bureau or Labor Department with their armies of researchers with expertise in statistics?

“I’m not authorized to do that,” Washburn said [Gale Courey Toensing, "Finally! Indian Country Gets Its Labor Force Report," Indian County Today, 2014.01.29].

That one poor economist also warns us not to apple-orange the BIA numbers with standard unemployment figures:

It is very important to note that this report does not provide any estimates of "unemployment.” “Unemployment,” as also defined above, refers to the proportion of people in “the labor force” who are not working, in accordance with the official Federal definition of “the labor force” as set by the Department of Labor. For example, people may be “available for work” in the sense of wanting a job, but would not be “in the labor force” if they do not have a job and are not actively looking for one. This measurement standard has long been established in the Federal Statistical System.

The above-mentioned standard definition of unemployment does not reflect any judgment as to whether people who are not working, and who want a job, should be actively looking for work. Indeed, it may well be the case that there are no job opportunities in certain areas of Indian Country, thereby making it pointless, or perhaps even irrational, for anyone without a job under these circumstances to be continually and actively looking for one in the same, economically depressed, geographic area.

The Act does not require this report to provide estimates of unemployment, and such estimates are not possible because data on the size of the labor force within tribal service populations are not available. Therefore, none of the statistics provided in this report should be seen as reflecting or representing estimates of unemployment. Anyone who uses the statistics provided in this report to infer any statement about unemployment with regard to any individual tribe, with regard to the tribes in a state or region, or with regard to Indian Country overall, will be misinterpreting and misrepresenting the findings of this report [2013 American Indian Population and labor Force Report, Bureau of Indian Affairs, January 2014, p. 9].

Oh, for criminy pete jeeper's sakes, can we just have a spreadsheet?

No, but we can creep slowly toward the most responsible answer possible to my commenter's question.

First, the 2014 BIA report indicates that South Dakota is doing worse at putting the Indian workforce to work than almost anyone else. Between 25.7% and 28.5% of tribal service populations age 16 and over who were available for work in 2010 were not working [pp. 33–34]. That's one of the worst rates in the nation, matched only by Arizona (27.0%–28.3%) and Utah (22.5%–29.0%). The national rate is 17.6%–17.9%. Evidently, it's harder for Indians to find work in South Dakota than most other states, and we could use proposals like Rep. Hickey's to catch us up.

Alas, the 2014 report does not provide the same direct numbers (or even arrive at them by the same methodology) as its last report, from 2007, based on 2005 data. To keep my extrapolations low, I want to look at the 2005 data.

Now let's address this problematic distinction between "in the labor force" (the category used for traditional unemployment calculations) and "available for work." Discussing nationwide data [p. 35], the 2014 BIA report finds that about 96% of those "available for work" were in the "labor force" either holding jobs or actively seeking jobs. In 2005, that percentage was closer to 97% (see 2005 BLS labor force data and count of those available to work but not in the labor force). As suggested by the report [see above, p. 9], that greater lack of jobs on in Indian Country may depress that percentage much further.

I want to combine BIA report numbers with South Dakota employment data from 2005. But before we do any calculations, we face a choice:

  1. Do we damn the torpedoes, mash the numbers together as is, and leave it to Governor Daugaard to hire his own economist to better rectify the numbers?
  2. Do we multiply and BIA "available for work" numbers by 97% to estimate "labor force"?
  3. Do we assume a much lower Indian labor force participation rate and multiply by a guessed lower percentage?

To strike a balance between restraint and extrapolation error, I will take the middle route and multiply BIA "available for work" numbers by the nationally, statistically attested 96%

Now, finally, let's crunch some 2005 numbers. That year, South Dakota reported an unemployment rate of 3.65%. Let's add the 2005 data South Dakota's nine tribes reported to the BIA for its 2007 report and see what comes out:

2005 avail to work labor force* employed unemployed unemp rate
SD 429,485 413,825 15,660 3.65%
Cheyenne River 11,205 10,843 1,312 9,531 87.90%
Crow Creek 1,317 1,274 558 716 56.22%
Flandreau Santee 1,247 1,206 775 431 35.78%
Lower Brule 731 707 351 356 50.38%
Oglala 29,539 28,584 3,131 25,453 89.05%
Rosebud 14,428 13,961 2,519 11,442 81.96%
Sisseton-Wahpeton 8,599 8,321 2,576 5,745 69.04%
Standing Rock 3,565 3,449 491 2,958 85.77%
Yankton 719 695 631 64 9.31%
SD + Tribal 498,530 426,169 72,361 14.51%
*Reservation labor force calculated by multiplying reported "available for work" by 2005 national labor force/available to work of 0.96769

According to this tribal data, there were more Oglala Sioux unable to find work on the Pine Ridge reservation than there were not working in South Dakota's official count of the rest of the state. If South Dakota is not including 57,000 out-of-work Indians in its unemployment data, then in 2005, instead of 3.65%, the statewide unemployment rate was 14.51%.

South Dakota currently (as of July 2014) brags about an official unemployment rate of 3.7%, tying with Vermont for fourth-lowest in the nation. That's half of California's 7.4%. That's well below the worst rate in the nation, Mississippi's 8.0%.

The Indian population has been growing faster than the white population in South Dakota. I haven't heard anyone saying that jobs have been proliferating on the reservations since 2005. And I don't think Northern Beef Packers or any of Mike Rounds's other magical job creation schemes created 57,000 jobs on the reservations.

So if the above 2005 data are not riddled with errors, we could safely speculate that South Dakota's all-inclusive unemployment rate is over 14%.

Whether that 14% would make us worst in the nation depends on how badly other states would be affected by counting their Indian labor data. But 14%