Heavy Lifting - thoughts and web finds by an economist
     I also contribute to Division of Labour Load HL's Front Page
Wednesday, January 10, 2007

Democrats and the Alternative Minimum Tax

There is a growing concern that as incomes continue to increase (either through productivity gains or cost of living adjustments), that many people will be pushed into the Alternative Minimum Tax. For a married couple filing jointly with tow kids, the total income required to kick into the AMT system is about $150,000 or so.

Today, the Tax Foundation released some data from IRS returns that ranked congressional districts by the percentage of returns that paid the AMT and by the average amount of (additional) AMT tax paid. In addition, the data contained the party of the current Congressman, implying, perhaps(?), there is some relationship between party affiliation and the amount of AMT or percentage of AMT returns.

I grabbed the data and threw them into STATA for a quick rundown to test this implied hypothesis. In many ways, the hypothesis is natural: generally, heavily democratic districts are also wealthy (although there are some notable exceptions) and therefore might also qualify moreso for the Alternative Minimum Tax.

Here's what I found from the Tax Foundation data. First some descriptive statistics:

Variable | Obs Mean Std. Dev. Min Max
pctamt | 436 .0227177 .0213036 .0028 .1352
amttax | 436 3718.521 938.2355 1554 10417
repub | 436 .4610092 .49905 0 1
south | 436 .3165138 .4656502 0 1
The variable pctamt is the percentage of tax returns that qualified for the AMT in a particular district. On average it was about 2.2% with a high of 13% and a low of .2%. The variable amttax is the additional AMT tax paid on average (I suppose this is for those returns that qualified for the AMT). The average was $3,718 with a low of $1,550 and a high of $10,400. I created an indicator variable for whether the House representative was a Republican and another indicator variable for whether the state was Southern (defined as a member of the CSA).

I estimated a straight linear regression model to test for differences in means:
. reg amttax repub

Source | SS df MS Number of obs = 436
-------------+------------------------------ F( 1, 434) = 3.89
Model | 3405777.05 1 3405777.05 Prob > F = 0.0491
Residual | 379518576 434 874466.764 R-squared = 0.0089
-------------+------------------------------ Adj R-squared = 0.0066
Total | 382924353 435 880285.869 Root MSE = 935.13

amttax | Coef. Std. Err. t P>|t| [95% Conf. Interval]
repub | -177.3044 89.84273 -1.97 0.049 -353.8853 -.7234009
_cons | 3800.26 61.00111 62.30 0.000 3680.365 3920.154
The results suggest that Republican districts, on average, paid $177 less in AMT taxes. How about the percentage of returns that qualified for the AMT?
. reg pctamt repub

Source | SS df MS Number of obs = 436
-------------+------------------------------ F( 1, 434) = 1.79
Model | .000810647 1 .000810647 Prob > F = 0.1817
Residual | .196611127 434 .000453021 R-squared = 0.0041
-------------+------------------------------ Adj R-squared = 0.0018
Total | .197421774 435 .000453843 Root MSE = .02128

pctamt | Coef. Std. Err. t P>|t| [95% Conf. Interval]
repub | -.0027354 .0020449 -1.34 0.182 -.0067546 .0012837
_cons | .0239787 .0013884 17.27 0.000 .0212498 .0267076
Relatively little evidence that Republican districts have fewer AMT returns, on average.

There is concern that we haven't controlled for all of the unobserved, state-level heterogeneity in the simple model. One possible extension is to include fixed effects for states:
. xtreg amttax repub , i(stateid) fe r

Fixed-effects (within) regression Number of obs = 436
Group variable (i): stateid Number of groups = 51

R-sq: within = 0.0064 Obs per group: min = 1
between = 0.0015 avg = 8.5
overall = 0.0089 max = 53

F(1,384) = 3.05
corr(u_i, Xb) = 0.0379 Prob > F = 0.0814

| Robust
amttax | Coef. Std. Err. t P>|t| [95% Conf. Interval]
repub | -143.2313 81.97951 -1.75 0.081 -304.4163 17.95356
_cons | 3784.552 64.17095 58.98 0.000 3658.381 3910.722
sigma_u | 548.43488
sigma_e | 872.46195
rho | .28322893 (fraction of variance due to u_i)

The results suggest that Republican districts tend to pay about $143 less, after controlling for state fixed effects. However, the parameter estimate is not distinguishable from zero at the five percent level (although it is at the 8 percent level - this suggests that there is a roughly 92% chance that the parameter estimate is different from zero).

If Republican districts actually do pay less than Democratic districts, does this mean that Republican districts have lower incomes, for which lower AMT taxes would be natural, or do they have better CPA's?

My broader point is that, in naive analysis, it seems that the differences between Republican and Democratic districts are not economically (nor strongly statistically) significant.


STATA data file here

I'm glad that the democratic districts pay slightly more taxes, this indicates that at least democrats have some integrity in their ideology. I don't thinkthat these so
-called "conservative" republicans are very conservative at all with our money or our lives (people dying in Iraq). I'd say they're pretty liberal in those areas. Okay, that's enough of my rant. Here might be some useful questions. Do those democratic districts have benefits because of those extra taxes? Do those republican districts have potholes on every street and sub-standard hospitals? I really miss California where I don't have to worry about driving over some humongous pothole, or breathing everyone else's cigarette smoke any time I want to go to eat in a public place. What if I wanted to make some money as a bartender while I'm getting through college, I'd have to risk my life by exposing myself to second-hand smoke. Democrats ARE good for some things.
Post a Comment


Le Chai - galerie du vin



Posts that contain Craig Depken per day for the last 90 days.


Heavy Lifting's Main Page
Email Me
Atom Feed

Heavy Lifting

Great Links

Money I Found Today

Heavy Lifting - Firehose style (56k warning)

Recent Posts

- Democrats and the Alternative Minimum Tax



Site Meter Blogroll Me!


Modified maystar design
powered by blogger