Is it the Economy, Stupid?
November 8, 2010 10:38 AM Subscribe
The Economist has created a rather cool interactive US map. The map allows a by state look at economic data (unemployment, GDP, personal income), demographic data, and voting in 2004 and 2008. (single link Economist)
Yes, very cool, but it's all just pretty meaningless numbers. Even TFA states "America's ethnic composition seems to have little consistent economic impact". The same can be said for age-related items; there seems to be no relationship between age and voting, which is commonly accepted as "over 65 = republican".
Also, the Economist also has the same problem everyone else does: the unemployment numbers are what is reported by the government, and with the limits on how long one can be on unemployment, the numbers are believed to be much higher.
posted by Old'n'Busted at 10:49 AM on November 8, 2010 [2 favorites]
Also, the Economist also has the same problem everyone else does: the unemployment numbers are what is reported by the government, and with the limits on how long one can be on unemployment, the numbers are believed to be much higher.
posted by Old'n'Busted at 10:49 AM on November 8, 2010 [2 favorites]
Looking at individual income and GDP in terms of percentage, by state, is, quite simply, a lie. Someone is palming a card, because the difference in population between the states can be an order of magnitude (and if you could DC as a state? Two orders of magnitude.)
It's trivial for SD to move percentage points. They have ~850K people living there. ND? ~640K. The City of Chicago has three times as many alone as South Dakota, and four times as many as North Dakota.
So, you need to hire vastly fewer people in SD or to improve the unemployment rate, and you need to hire/fire *fifteen times as many people* to move the IL unemployment rate the same amount. It's even worse with ND -- *nineteen* times to move IL the same percentage point. 10,000 people out of work in ND probably moves unemployment up a point. 10,000 people out of work in IL is lost in the noise.
Let's see the actual numbers -- number unemployed, actual income. Percentage comparisons on order of magnitude differences are always suspect. Always.
posted by eriko at 10:49 AM on November 8, 2010 [8 favorites]
It's trivial for SD to move percentage points. They have ~850K people living there. ND? ~640K. The City of Chicago has three times as many alone as South Dakota, and four times as many as North Dakota.
So, you need to hire vastly fewer people in SD or to improve the unemployment rate, and you need to hire/fire *fifteen times as many people* to move the IL unemployment rate the same amount. It's even worse with ND -- *nineteen* times to move IL the same percentage point. 10,000 people out of work in ND probably moves unemployment up a point. 10,000 people out of work in IL is lost in the noise.
Let's see the actual numbers -- number unemployed, actual income. Percentage comparisons on order of magnitude differences are always suspect. Always.
posted by eriko at 10:49 AM on November 8, 2010 [8 favorites]
What the heck is going on in North Dakota?
An oil rush, for one thing.
posted by mr_roboto at 10:59 AM on November 8, 2010 [1 favorite]
An oil rush, for one thing.
posted by mr_roboto at 10:59 AM on November 8, 2010 [1 favorite]
Looking at individual income and GDP in terms of percentage, by state, is, quite simply, a lie.
Not a lie, but not a complete truth either. State level breakdowns are useful in determining patterns that might effect the Senate, and the Senate has a greater effect on national policy than the population based House.
posted by kuujjuarapik at 11:00 AM on November 8, 2010 [2 favorites]
Not a lie, but not a complete truth either. State level breakdowns are useful in determining patterns that might effect the Senate, and the Senate has a greater effect on national policy than the population based House.
posted by kuujjuarapik at 11:00 AM on November 8, 2010 [2 favorites]
DC - simultaneously by miles the backest and richest jurisdiction in the nation.
With, you know, almost zero overlap between those two groups.
posted by Navelgazer at 11:10 AM on November 8, 2010
With, you know, almost zero overlap between those two groups.
posted by Navelgazer at 11:10 AM on November 8, 2010
California, Texas, Nevada and New Mexico have high numbers of Hispanics? You don't say. Could that have anything to do with them all having been PART of Mexico in the not too distant past, and explored and settled by the Conquistadors and their followers before the Pilgrims ever landed at Plymouth?
Considering that the entire population of the Mexican Cession in 1848 was in the low tens-of-thousands, and by the time of the gold rush the very next year, hundreds of thousands of eastern US citizens (mostly not Hispanic) flooded the west (San Francisco, one of the largest "cities" in Alta California went from an 1848 population under 1000 to over 25000 between 1848 and 1849 due to the gold rush), I would bet that the large proportion of Hispanic citizens in the population of these states has far more to do with them being border states than their having been ceded to the US from Mexico.
posted by chimaera at 11:15 AM on November 8, 2010 [3 favorites]
Considering that the entire population of the Mexican Cession in 1848 was in the low tens-of-thousands, and by the time of the gold rush the very next year, hundreds of thousands of eastern US citizens (mostly not Hispanic) flooded the west (San Francisco, one of the largest "cities" in Alta California went from an 1848 population under 1000 to over 25000 between 1848 and 1849 due to the gold rush), I would bet that the large proportion of Hispanic citizens in the population of these states has far more to do with them being border states than their having been ceded to the US from Mexico.
posted by chimaera at 11:15 AM on November 8, 2010 [3 favorites]
State level breakdowns are useful in determining patterns that might effect the Senate, and the Senate has a greater effect on national policy than the population based House.
And, technically, small states are overrepresented in the House.
posted by maryr at 11:21 AM on November 8, 2010
And, technically, small states are overrepresented in the House.
posted by maryr at 11:21 AM on November 8, 2010
Not necessarily. Compare Rhode Island with 2 representatives and Texas with 32. Or to be less extreme, NY 28 and Texas 32 representatives.
posted by kuujjuarapik at 11:40 AM on November 8, 2010
posted by kuujjuarapik at 11:40 AM on November 8, 2010
Also, the Economist also has the same problem everyone else does: the unemployment numbers are what is reported by the government, and with the limits on how long one can be on unemployment, the numbers are believed to be much higher.
[morbo]
UNEMPLOYMENT RATES DO NOT WORK THAT WAY!
[/morbo]
Unemployment rates as reported by the government have NOTHING AT ALL to do with being on unemployment or not.
Here is how the government calculates the unemployment rate:
(1) They contact a whopping shitload of Americans., usually around 50,000.
(2) They ask you, "Did you work for pay last week?"
(3) If you said "Yes," congratulations! You are employed.
(4) If you said "No," they ask you a bunch more questions to try to separate people who are just on vacation or retired from people who are looking for work. The first question boils down to "Have you done anything to look for work in the past 4 weeks?"
(5) If you said "Yes," you are unemployed (U-3). This is the "standard" measure.
(6) If you said "No," they ask you whether you'd like to work but have given up looking because there's nothing out there.
(7) If you said "Yes," you are a discouraged worker! You do not count in unemployment U-3, but do in unemployment U-4.
(8) If you said "No," and you would like to work but haven't looked in the past 4 weeks for some other reason, you do not count in U-4 but do in U-5.
(9) At the end, they ask whether you're working part-time or full-time, and whether you want to be working full-time. If you're working part-time but want full-time, you don't count in U-5 but do in U-6.
(10) After all that, they massage the data to get rid of normal seasonal fluctuations.
tl;dr: Unemployment rates are not calculated from unemployment benefits. They are calculated from responses to frequent and very large surveys. The other unemployment measures, U-4 through U-6, are not some deep secret. BLS reports them at the same time as they report the "standard" U-3.
This is not difficult stuff. Remember: This is the INTERNET. Accuracy counts.
posted by ROU_Xenophobe at 11:54 AM on November 8, 2010 [23 favorites]
[morbo]
UNEMPLOYMENT RATES DO NOT WORK THAT WAY!
[/morbo]
Unemployment rates as reported by the government have NOTHING AT ALL to do with being on unemployment or not.
Here is how the government calculates the unemployment rate:
(1) They contact a whopping shitload of Americans., usually around 50,000.
(2) They ask you, "Did you work for pay last week?"
(3) If you said "Yes," congratulations! You are employed.
(4) If you said "No," they ask you a bunch more questions to try to separate people who are just on vacation or retired from people who are looking for work. The first question boils down to "Have you done anything to look for work in the past 4 weeks?"
(5) If you said "Yes," you are unemployed (U-3). This is the "standard" measure.
(6) If you said "No," they ask you whether you'd like to work but have given up looking because there's nothing out there.
(7) If you said "Yes," you are a discouraged worker! You do not count in unemployment U-3, but do in unemployment U-4.
(8) If you said "No," and you would like to work but haven't looked in the past 4 weeks for some other reason, you do not count in U-4 but do in U-5.
(9) At the end, they ask whether you're working part-time or full-time, and whether you want to be working full-time. If you're working part-time but want full-time, you don't count in U-5 but do in U-6.
(10) After all that, they massage the data to get rid of normal seasonal fluctuations.
tl;dr: Unemployment rates are not calculated from unemployment benefits. They are calculated from responses to frequent and very large surveys. The other unemployment measures, U-4 through U-6, are not some deep secret. BLS reports them at the same time as they report the "standard" U-3.
This is not difficult stuff. Remember: This is the INTERNET. Accuracy counts.
posted by ROU_Xenophobe at 11:54 AM on November 8, 2010 [23 favorites]
Not necessarily. Compare Rhode Island with 2 representatives and Texas with 32.
Rhode Island's population: 1,053,209
Texas's population :24,782,302
In other words, Rhode Island with 2 representatives has 1 representative for every approx. 526,604 people. Texas would have 47 representatives at that rate.
posted by applemeat at 11:56 AM on November 8, 2010
Rhode Island's population: 1,053,209
Texas's population :24,782,302
In other words, Rhode Island with 2 representatives has 1 representative for every approx. 526,604 people. Texas would have 47 representatives at that rate.
posted by applemeat at 11:56 AM on November 8, 2010
California, Texas, Nevada and New Mexico have high numbers of Hispanics? You don't say. Could that have anything to do with them all having been PART of Mexico in the not too distant past, and explored and settled by the Conquistadors and their followers before the Pilgrims ever landed at Plymouth?
The Economist is published in London with an international audience in mind - a Cambodian, or a Nigerian reading it do not neccesarily know these things.
posted by atrazine at 12:18 PM on November 8, 2010 [3 favorites]
The Economist is published in London with an international audience in mind - a Cambodian, or a Nigerian reading it do not neccesarily know these things.
posted by atrazine at 12:18 PM on November 8, 2010 [3 favorites]
Dark secret I learned from a Skull & Bones refugee: House representation also includes a fudge factor based on gallons of coffee syrup consumed per capita.
posted by yerfatma at 12:22 PM on November 8, 2010
posted by yerfatma at 12:22 PM on November 8, 2010
ND? ~640K
And we'd like to keep it that way. Nothing to see here, move along.
posted by AElfwine Evenstar at 12:23 PM on November 8, 2010 [1 favorite]
And we'd like to keep it that way. Nothing to see here, move along.
posted by AElfwine Evenstar at 12:23 PM on November 8, 2010 [1 favorite]
applemeat- In my response to maryh, by 'small' I was thinking square mile. There is very little correlation between geographic size and representation, as I meant to show with 'small' NY compared to 'big' TX. If maryh meant 'small' as in population, then I'm wrong.
posted by kuujjuarapik at 12:28 PM on November 8, 2010
posted by kuujjuarapik at 12:28 PM on November 8, 2010
Yes, I meant small in population, not size. For example, Rhode Island has 2 reps while Alaska has 1.
Texas will likely gain a couple reps when districts are next assigned, if I understand correctly.
posted by maryr at 12:31 PM on November 8, 2010
Texas will likely gain a couple reps when districts are next assigned, if I understand correctly.
posted by maryr at 12:31 PM on November 8, 2010
ROU: it's 60,000, not 50, according to the BLS website. The monthly published results are the Employment Situation.
"Because these interviews are the basic source of data for total unemployment, information must be factual and correct. Respondents are never asked specifically if they are unemployed, nor are they given an opportunity to decide their own labor force status. Unless they already know how the Government defines unemployment, many of them may not be sure of their actual classification when the interview is completed." In addition, "the total numbers are 'weighted,' or adjusted to independent population estimates (based on updated decennial census results)". The margin of error is +/- 0.8%.
In addition, for some states, the sample set is too small. Montana for example, only has about 450 in their sample set, and instead use the number of people getting unemployment insurance benefits. It is this benefits number being used that causes most - but not all - of the other reported values in the media to be suspect.
Unless the reporting source clearly identifies where they got their numbers and what they are, it is usually safest to assume that the reported numbers are the benefits values and not surveys; states have a vested interest to have low unemployment rates reported, since the perception is high unemployment == crap place to live and work.
The Economist should have listed and documented all of this information for their pretty chart. I expected better from them. (this is *NOT* a reflection on bearwife).
posted by Old'n'Busted at 12:33 PM on November 8, 2010 [1 favorite]
"Because these interviews are the basic source of data for total unemployment, information must be factual and correct. Respondents are never asked specifically if they are unemployed, nor are they given an opportunity to decide their own labor force status. Unless they already know how the Government defines unemployment, many of them may not be sure of their actual classification when the interview is completed." In addition, "the total numbers are 'weighted,' or adjusted to independent population estimates (based on updated decennial census results)". The margin of error is +/- 0.8%.
In addition, for some states, the sample set is too small. Montana for example, only has about 450 in their sample set, and instead use the number of people getting unemployment insurance benefits. It is this benefits number being used that causes most - but not all - of the other reported values in the media to be suspect.
Unless the reporting source clearly identifies where they got their numbers and what they are, it is usually safest to assume that the reported numbers are the benefits values and not surveys; states have a vested interest to have low unemployment rates reported, since the perception is high unemployment == crap place to live and work.
The Economist should have listed and documented all of this information for their pretty chart. I expected better from them. (this is *NOT* a reflection on bearwife).
posted by Old'n'Busted at 12:33 PM on November 8, 2010 [1 favorite]
The Economist should have listed and documented all of this information for their pretty chart.
Yes, they should have. Thanks to you, Old'n'Busted, for filling in what should have been there.
posted by bearwife at 12:51 PM on November 8, 2010
Yes, they should have. Thanks to you, Old'n'Busted, for filling in what should have been there.
posted by bearwife at 12:51 PM on November 8, 2010
In the Democracy in America blog there is an interesting comparison between the way in which America divides between the city and the country and a number of European countries vote in regional ways without the city/country divide.
posted by sien at 1:45 PM on November 8, 2010 [1 favorite]
posted by sien at 1:45 PM on November 8, 2010 [1 favorite]
So, you need to hire vastly fewer people in SD or to improve the unemployment rate, and you need to hire/fire *fifteen times as many people* to move the IL unemployment rate the same amount. It's even worse with ND -- *nineteen* times to move IL the same percentage point. 10,000 people out of work in ND probably moves unemployment up a point. 10,000 people out of work in IL is lost in the noise.
I'm not sure I take your point here. Why can't you argue it the other way? If there's fifteen times as many people in Illinois, why aren't there fifteen times as many companies, and fifteen times as many hirings and firings? Shouldn't it be easier to create the same number of jobs in Illinois than it is in South Dakota?
Or are you saying that all of states with positive GDP growth in the Great Plains are just experiencing random fluctuations in their GDP that happen to be positive. If so, hey, I'll make you a bet: I'll flip a coin ten times, I'll give you a dollar if it comes up heads ten times, and you give me a dollar if it doesn't.
posted by Johnny Assay at 2:40 PM on November 8, 2010
I'm not sure I take your point here. Why can't you argue it the other way? If there's fifteen times as many people in Illinois, why aren't there fifteen times as many companies, and fifteen times as many hirings and firings? Shouldn't it be easier to create the same number of jobs in Illinois than it is in South Dakota?
Or are you saying that all of states with positive GDP growth in the Great Plains are just experiencing random fluctuations in their GDP that happen to be positive. If so, hey, I'll make you a bet: I'll flip a coin ten times, I'll give you a dollar if it comes up heads ten times, and you give me a dollar if it doesn't.
posted by Johnny Assay at 2:40 PM on November 8, 2010
It's only really interesting to compare states to their peer groups, but that doesn't make a pretty map. That said if I hear another Minnesota politician say that our unemployment is higher than South Dakota's because they have low taxes (and not that they have oil and a tiny population), I'll be very annoyed.
posted by miyabo at 3:00 PM on November 8, 2010 [1 favorite]
posted by miyabo at 3:00 PM on November 8, 2010 [1 favorite]
You'll be very annoyed.
posted by meinvt at 3:33 PM on November 8, 2010 [3 favorites]
posted by meinvt at 3:33 PM on November 8, 2010 [3 favorites]
tl;dr: Unemployment rates are not calculated from unemployment benefits. They are calculated from responses to frequent and very large surveys. The other unemployment measures, U-4 through U-6, are not some deep secret. BLS reports them at the same time as they report the "standard" U-3.
That's the point, though. U-3 is the number that is bandied about almost exclusively, when, in fact, the other U-xs are much more relevant to the problems of unemployment, especially in such a deep recession. You need to forgive the slight inaccuracy of folks whose shorthand for your explanation is "they don't count people who aren't on unemployment" if you want to actually address the issue that was raised.
posted by Mental Wimp at 3:44 PM on November 8, 2010
That's the point, though. U-3 is the number that is bandied about almost exclusively, when, in fact, the other U-xs are much more relevant to the problems of unemployment, especially in such a deep recession. You need to forgive the slight inaccuracy of folks whose shorthand for your explanation is "they don't count people who aren't on unemployment" if you want to actually address the issue that was raised.
posted by Mental Wimp at 3:44 PM on November 8, 2010
Oh, also forgot to say, I wonder what the correlation with the 2010 voting would be.
posted by Mental Wimp at 3:46 PM on November 8, 2010
posted by Mental Wimp at 3:46 PM on November 8, 2010
In the Democracy in America blog there is an interesting comparison between the way in which America divides between the city and the country and a number of European countries vote in regional ways without the city/country divide.
Thanks. That was actually a bit more interesting than the link in the OP. I've long wondered if the real reason Europe is so much more liberal than the US is because it's dramatically more urban and less rural. While the data in that link seem to argue otherwise, I would counter that the "rural" parts of Europe aren't rural at all by American standards. Why there should be such a strong link between rural living and conservative politics in the first place is something I'd certainly like to know.
posted by Marla Singer at 5:28 PM on November 8, 2010
Thanks. That was actually a bit more interesting than the link in the OP. I've long wondered if the real reason Europe is so much more liberal than the US is because it's dramatically more urban and less rural. While the data in that link seem to argue otherwise, I would counter that the "rural" parts of Europe aren't rural at all by American standards. Why there should be such a strong link between rural living and conservative politics in the first place is something I'd certainly like to know.
posted by Marla Singer at 5:28 PM on November 8, 2010
I've long wondered if the real reason Europe is so much more liberal than the US is because it's dramatically more urban and less rural.
I'd wager it has more to do with the lack of a modern frontier, less isolation from other great powers, and *cough* something something natural resources. Some of America's most liberal states (Vermont, Delaware) are mostly rural and the historical home of progressivism is the western Great Lakes region (Minnesota, Wisconsin), and specifically (if I'm not wrong) areas other than the dominant metropolitan areas.
posted by kittyprecious at 5:55 PM on November 8, 2010
I'd wager it has more to do with the lack of a modern frontier, less isolation from other great powers, and *cough* something something natural resources. Some of America's most liberal states (Vermont, Delaware) are mostly rural and the historical home of progressivism is the western Great Lakes region (Minnesota, Wisconsin), and specifically (if I'm not wrong) areas other than the dominant metropolitan areas.
posted by kittyprecious at 5:55 PM on November 8, 2010
I love maps, especially the interactive kind. Wish there were more in my life.
On a side note, god damn there are a lot of white people in New Hampshire and Vermont. Coming from California its hard to imagine such little diversity.
posted by Defenestrator at 10:10 PM on November 8, 2010
On a side note, god damn there are a lot of white people in New Hampshire and Vermont. Coming from California its hard to imagine such little diversity.
posted by Defenestrator at 10:10 PM on November 8, 2010
It's not that there are a lot of white people. It's just that there are hardly any minorities. I mean, Vermont (620K) has fewer people than San Francisco proper and the three northern New England states, Vermont, New Hampshire (1.3 mil), and Maine (1.3 mil), combined have fewer people than LA proper.
posted by maryr at 7:21 AM on November 9, 2010
posted by maryr at 7:21 AM on November 9, 2010
Look, we have to grow the maple syrup and Cabot cheese somewhere.
posted by maryr at 2:49 PM on November 10, 2010
posted by maryr at 2:49 PM on November 10, 2010
My new favorite part of this map, by the by, is clicking on the "demographics" tab and then doing the mouseover Hawaii. Hmmm, there seems to be an ethnic group missing.
posted by kittyprecious at 9:06 AM on November 22, 2010
posted by kittyprecious at 9:06 AM on November 22, 2010
« Older They Live | The Letters of Constantin Brunner Newer »
This thread has been archived and is closed to new comments
posted by schmod at 10:48 AM on November 8, 2010 [1 favorite]