There's been this meme floating around the internet for a year or two now. Anytime a conservative or libertarian mentions welfare or the 47%, the standard liberal retort is to cite that red "welfare" states receive more federal money than the IRS collects in taxes from that state. Liberal economist Paul Krugman would have you believe that Social Security and Medicare is what constitutes mooching. Their narrative is that Republicans are bigger moochers than Democrats.
The problem with both of those claims is that it doesn't get to the heart of what "mooching" is. The problem with the first claim is that federal spending is made up mostly of defense expenditures at the the state level. Defense is not welfare. Actual welfare and poverty programs only amount to about 10% of the expenditures at the federal level. Now if a state received only funds for poverty programs, then you could claim that it is a welfare state. But unfortunately for their argument, this is not this case. PBS states, "In all but a handful of states, Department of Defense dollars account for by far the majority of federal dollars." (Other big ticket items that round out the list for state money from the federal government include farm subsidies, retirement programs and infrastructure projects) It's simply a function of flowing from the states without large defense operations and retired people to the states with them. If a less populated state has a large military base with a legion of personnel conducting operations should we be surprised that there might be an imbalance of funds? No, because it is a government organization that is not producing goods, but is consuming ammo, gas, food, electricity, salaries etc.. Also, most states don't tax military paychecks, which would somewhat offset the federal expenditure, so overall there is going to be a net draw of funds. But more to the point, national defense is a common good that benefits the whole country, so it can hardly be classified as mooching.
And then there are other problems with Krugman's claim. Does he really consider Social Security and Medicare recipients, who paid payroll taxes into the system their whole lives only to get a payout during retirement, a moocher? I don't think that is what constitutes a moocher in anyone's definition, except maybe for a very special liberal like Krugman. But wait, then there's another problem; how do you control for a state that is a retirement haven like Florida or Arizona? A good portion of these people worked in other states only to migrate to the retirement haven during their golden years. So this would show up on the books as a contribution in one state and later a draw in another. These are not the welfare queens that are what people have in mind when they are talking about government dependency. But they want to argue that the red-states/Republicans are moochers, so they've got to fit the right set of facts to fit their narrative somehow.
Bloomberg jumped into the fray recently with an article attempting to stir up the debate. They titled it "Food Stamp Cut Backed by Republicans With Voters on Rolls" with the obvious implication that Republicans are the ones that benefit the most from food stamps. What did they find?
Among the 254 counties where food stamp recipients doubled between 2007 and 2011, Republican Mitt Romney won 213 of them in last year’s presidential election, according to U.S. Department of Agriculture data compiled by Bloomberg.
Aside from the obvious bias painting Republicans as heartless villains and Democrats the saviors with the "commitment to help those struggling to meet basic needs" line, the claim could sound pretty damning, right? Well, not if you are a critical thinker and ask the right questions.
The first point that came to mind for me was that Romney won a majority of the counties in the country. Ergo, there are more red counties for events to happen. They didn't use a rate and instead a raw number, so therefore even if the food stamp growth rate were uniformly distributed across the country's counties, more would register for red counties. I bet the counties that had a doubling of traffic accidents also went to Romney. The second point that came to mind is the fact that these counties are much smaller than counties that went to Obama. In fact, they are less than half the size as shown in Table I. There are more opportunities for doubling, because it is much easier for a small county with a small food stamp population to double it than it is for a large county with a large food stamp population to do the same. But does a small county with 5,000 people and 500 food stamp recipients have the same weight as a large county with 50,000 people and 5000 food stamp recipients? A related point is that, a jump from 1% to 2% is a doubling, but from 51% to 100% is not. What if the Republican counties were just that much more self-reliant to begin with? More on this in a bit, but Romney had 164 counties with food stamp use under 5%, compared with only 33 for Obama.
What they should have done was report the population-weighted food stamp rate by party according to county voting preferences. Seems to me that would get to the heart of the matter. Since they went through all the datasets to aggregate the figure they reported, I bet they calculated this figure and would have published it had it been the result they were looking to report. But since they didn't publish this figure, it tells you a lot of what they didn't find, huh? Sometimes omission contains information.
*The Guardian dataset reports about 1000 more counties/areas than officially reported by the census. If matched on the second dataset discussed below, the results become even more favorable to my argument.
To solidify their conclusion in the article they pull out a single county "which backed Romney with 81 percent of its vote, [and] has the largest proportion of food stamp recipients among those that he carried." What was the proportion? 10%? 20%? How does it compare to Obama's county with the highest proportion? No mention of that. We'll get back to that later.
But anyway why pussyfoot around with all the misleading metrics about federal spending and food stamp growth with sleight of hand? If you want to determine who are the biggest moochers, why not go straight to the welfare and food stamp data? (Especially since those who are on welfare are generally on a whole host of other dependency programs, pardon the pun). Well we know the reason, but since I am on the other side of this debate, I will go straight to the relevant and inconvenient facts. This probably won't surprise anyone unless they really believe the red state-moocher narrative. We'll start with the welfare data; blue states have an over-representation of welfare cases relative to their population sizes and as a consequence of the former, red states have under-representation based on their population sizes.
About 1.3% of the nation are on welfare proper. I say proper because there is a whole treasure trove of backdoor welfare programs that don't have the welfare label (think food stamps, tax credits and Social Security Disability), but I hope to run some analysis in the future if I can find all of these datasets. If the welfare population were uniformly distributed across the nation then each state would have 1.3% of their population on welfare. But what we find is that Democratic states have 24% more representation than the national average after adjusting for population size while Republicans have 43% less. And this is even giving Democrats Florida, which is one of the most evenly divided states in the union. Without Florida going to either party, Democratic states would be overrepresented by 33.7%. A shocking fact is that a full third of the welfare cases in the nation come just from the state of California, yet they only represent 12% of the nation's population.
If you want raw number of welfare dollars, you'll see that only two states out of the top 15 welfare spenders are red states (if I'm playing their game).
Even still, this is all a very indirect way of getting at the underlying issue of which party mooches more. Going back to the food stamps, here's a thought experiment: You could have a highly bifurcated county with lots of poor Democrats and lots of rich Republicans. Say all 49% of the Democrats are on food stamps and all 51% of the Republicans in the county are wealthy. Well the county would vote Republican, but the food stamp benefit ratio would also register highly (or vice versa, for that matter). In a more realistic case, you'd have 5% of the county on food stamps and the county could vote either way, but that wouldn't tell you if the 5% was Republican or Democrat. So it's really hard to pin down which party mooches more when looking at state or county level more unless you have individual dependency data matched with voting behavior. Alas, we do.
Before we go there, I could draw your attention to the summary page that shows the largest growth in food stamp use went mostly to blue states and the highest proportion of food stamp use goes to Washington DC followed by a mix of blue and red states. But I'm not going to play this game. Frankly, because I don't need to.
If you look at the interactive county map that shows the highest concentration of food stamp use, it does appear at first glance that the southern red states are the biggest offenders. But look a little deeper. Most would consider it to be racist to just point to a map showing racial representation, and since we have actual data on food stamp use, we can do better than just simple assumptions. But you will notice that in a majority of the cases, black recipients, a factor 93% correlated with Democrats, outnumber whites (a category that includes Hispanics) by a factor of two to one on average (even though whites outnumber blacks by a factor of 5 in the general population). In fact, running a beta regression on the percentage of blacks on food stamps against the percent of the population on food stamps (3.39, t-stat: 13.7) and the log of the county size (.12, t-stat: 10.1), and another on the percentage of whites on food stamps against the same variables (-1.71, t-stat: -7.3; 0.049, t-stat: 4.22), we find that county size is positively related to food stamp dependency in both cases (larger counties vote for Democrats), but at a faster rate for blacks, and that as the percentage of food stamp use in a county increases, it is positively associated with blacks and negatively associated with whites showing that, on average, most of the food stamp dependency comes from blacks, and therefore Democrats.
I decided to combine the two datasets on food stamps and voting to see what information would shake out. Some FIPS county numbers did not match between the two, but 2,654 counties resulted from this merged dataset. The average Romney county has 12.74% on food stamps, while the average Obama county has 14.94%, a statistically significant difference. I then ran a regression on the percentage of the population on food stamps against the population size, number of Obama votes, number of Romney votes and the percentage of the total county population that voted. As Table III shows, all variables are statistically significant. We find results that are similar to the above analysis. Food stamp use increases with the size of the population, which shows that that larger urban areas are meccas for government dependency. The kicker is that after holding the population constant, we find that food stamp use is positively related to Obama voters and strongly negatively related to Romney voters. This clearly shows that food stamp dependency is a factor mostly associated with Democrats.
What else? For the years 2004-2007 the Maxwell school at Syracuse University has conducted a nationally representative poll that surveys questions pertaining to government dependency and political party affiliation. Given all the analysis already performed it should not be surprising that Democrats make up a majority of direct usage for these programs. For instance, as shown on the chart to the left, a full 81% of public housing, and 74% of Medicaid is consumed by Democrats. The worker's compensation proportion is interesting. Either Democrats get injured more often because they work in more dangerous industries, or they are more likely to use it as another backdoor welfare program.
We also know that Republicans earn about 40% more income than Democrats, on average. Voter exit polls for the most recent election showed that 63% of the sub $30,000 per year vote went to Obama and Obama voters were less likely to have jobs (and were younger). From a recent NPR poll surveying the long term unemployed, the Democrat's proportion is twice that of Republicans.
Liberals repeatedly question why conservatives "vote against their own interests." But maybe the simple answer is that they don't. Think about it, who would? The answer to this 'paradox' is because the ones who are actually voting for Republicans aren't mooching off the government. The Occam's razor answer was staring them right in the face the whole time, but they either couldn't see the facts, or deliberately ignored them for their agenda.
So there you have it - the "Red States are Welfare Queens" meme is only a Myth - Case Closed.