Amidst all the public debate about how states are being bled dry by militant public unions, you wouldn’t know that we just had a major housing bubble across the country followed by a financial system near-collapse and the most prolonged downturn since the Great Depression. Chris Hayes addressed this opportunism, the ignoring of the housing crisis to push long-standing right-wing priorities, in the opening segment of the Rachel Maddow show last night, and I think it’s worth throwing a graph together.
John Side posts some graphs of state budget shortfalls against public union density on his site The Monkey Cage:
And:
A commenter summarized the general finding:
I just coded the data “TheRef” posted to distinguish between states with no collective bargaining law and states with some sort of law (0=no law, 1=anything else) and used this to predict the 2011 shortfall as a percentage of budget. I have no idea if this coding is appropriate, but it should provide a rough estimate.
While the relationship was positive (like the r coefficient in the post) it explained less than 2% of the variance (R^2 = .017). This is actually less explained variance than that explained in the above post (.19*.19 = .04), though this could be due entirely to the linear compared to categorical nature of the predictors. Just to note, the unstandardized beta was 3.08.
Interesting, if that not significant. You know what is interesting, significant and recent? A multi-trillion dollar housing bubble.
I’m going to do the same graph with Total Shortfall as Percent of FY11 Budget from the CBPP (table four) as well as negative and near negative equity as percentage of mortgages, Q3 2010 from CoreLogic. Negative equity is correlated with all kinds of other bad things like unemployment, but from my point of view it’s a good first approximation for how devastating the housing bubble was to a community. The more the bubble popped, the more people that were hit by falling house prices, the more negative equity grows as a percent of mortgages. (Especially since Case-Shiller sold out and took their data to subscription only, and even then, I’m not sure there’s a particularly better state-level approximation – thoughts?)
Graph:
Significant (t-stat of 3.53), and it is significant with or without that outlier in the upper-right corner (Nevada). The mechanisms for how this contributes is important – is it the unemployment? Is it that state governments with a larger housing bubble got more confident and spent as if all those property taxes were on their way? Are there other important, casual mechanisms? These are all good and crucial questions for us to answer, ones we should take up when we finish scapegoating teachers.
Rortybomb 
Well-put in the last line.
Pingback: Housing collapse led to state budget gaps, not unions « Jeff Pelline's Sierra Foothills Report
Other important causal mechanisms? Commute distance and availability of public transit. When gasoline hit $4 dollars, people had to fill the tank before paying the mortgage because getting to work comes first. To rub salt in the wound, KochBros are doing their best to cut transit as gas threatens to go to $5.
Pingback: Predicting state budget deficits in two graphs | Charleenhandal's Blog
This is very useful. Keep up the good work.
You could try splitting the chart into two — one for recourse states, and one for non-recourse states. It would be interesting to see if the significance improves under those changes.
I wish you had reported the actual significance without Nevada — or at least posted a link to your data so others could do so. Just looking at the picture, Nevada HAS to exert a very large influence on this regression model. Even if it’s not “unusual” in the normal outlier sense, it has to be pulling the line up.
Given that I had some free time, I tried to recreate your data and run the test myself. I am not sure I was able to replicate your data perfectly from the links you listed, but when I ran the full model with the 39 states that seemed to have both variables you were using, I got really close to your picture (t-stat of 3.29 and p-value of .002). When I dropped Nevada, however, the significance disappeared as well (t-stat of 1.38 and .18). Would you mind linking to your data so I can see where the discrepency comes from?
Sorry, that last test gave a t-stat of 1.38 and a p-value of .18
Your correlation coefficient has to be close to .5 on that data, maybe .6 with Nevada included.
You can tell just by looking at the data that a linear model is going to be meaningless.
If you post the actual data, I can’t get the core logic data, I can add some more. I give you credit for at least graphing the data before reporting your stats, but you have to go to the next step and check to see if the graph and the stat values make sense.
That data has no meaningful correlation, your graph is a scatter plot with no obvious relation. Quick check, send that graph around to 10 stats folks without the line and ask them if their is a relationship…
Pingback: Links 2/26/11 « naked capitalism
Which is normal: state revenue levels during the unsustainable housing bubble or the “new normal” where, surprise, surprise, revenue levels are back where they were before the bubble. Who could have predicted that house prices and stock prices would revert to their historical mean? Who could have ever predicted that revenue levels would fall back to those dark, dark ages of say 2003?
This is frankly just pathetic analysis.
I remember when I was a kid, the argument was just starting about whether or not cigarettes caused cancer. There was a brief item in, I think, Time Magazine. This would probably have been 1947 or 1948. They showed the correlation between the importation of bananas and cancer rates. Nearly perfect correlation! I was very impressed, and a few years later, when I was 13, I started smoking.
The final graph shows very clearly that there are two clusters of states. What is common to each cluster?
Yup, a regression line is shambolic.
Pingback: Democrat-Controlled Legislatures Also Voting to Strip Union Rights | DC Skeptics