Wednesday, October 12, 2011

For Folks Smarter than Me: Can a Regression Analysis Predict the Success of Herman Cain's "9-9-9" Plan?

A question for folks smarter than I am on these matters...

We have a diverse readership here on WARN. I have chatted with a few of you via email and in other contexts and know that there are some quantitatively oriented and trained social scientists among you.

In the interest of transparency, I can read the math and the stats. I can provide some takeaway on what they convey or not. I can also do a pretty decent sniff test on a funky looking model--but that doesn't mean I can tell you exactly what is wrong and how to do the math by hand to smoke it out. In those instances I say "you need to go talk to that person down the hall."

Today's New York Times piece on Herman Cain's "9-9-9" plan--or as I like to call it the "no, no, no" plan--had a very revealing passage. We now know that the author of said policy briefing is not an economist, he is actually a hedge fund financial manager type, but the following passage really struck me:

In an interview, Mr. Lowrie said he had a bachelor of science degree in accountancy from Case Western Reserve University. On his Facebook page, he describes his political views as “free markets.” Mr. Lowrie said he had been inspired by two well-known proponents of supply-side thinking: Arthur Laffer, often considered the father of the concept that lower tax rates help pay for themselves by generating additional economic growth, and Jude Wanniski, who promoted the idea among politicians. Mr. Lowrie became involved with the Ohio chapter of Americans for Prosperity, the conservative organization supported by the billionaire Koch brothers.
The plan could have major economic and political challenges: It might result in a substantial revenue loss for the government and shift the tax burden toward lower- and middle-income people.In an interview, Mr. Cain, a math major in college, said he had asked Mr. Lowrie to do a “regression analysis” that would allow the government to eliminate all existing taxes, including those on capital gains and estates, and collect the same revenue from just three streams. “The number came up to be 9 percent,” Mr. Cain said. “And that’s how we came up with 9-9-9.”
Mr. Lowrie, who met Mr. Cain at a conference sponsored by the conservative Club for Growth, dismissed the notion that his own understanding of economics was limited by lack of a Ph.D. “I don’t list myself as an economist,” he said. “I have an accounting degree, and I’m an investment adviser. I’ve never hung out in a faculty lounge.”
A former staff member for Mr. Cain in Iowa described his and Mr. Lowrie’s relationship as “buddy-buddy,” adding, “They were just like two executives palling around together.”Their plan has drawn fire from both right and left. Conservatives are wary of a national sales tax, concerned that it would create another, easily increased method of taxation. Among the critics are The Wall Street Journal editorial page and Bruce Bartlett, an official in the Reagan and first Bush administrations, who contributes to the Economix blog for The New York Times.
First, anything that has to do with the much misunderstood Laffer Curve gives me pause. Second, anything having to do with the Koch Brothers should raise a huge red flag amongst folks concerned with the Common Good. But, the statistical lingo which is designed to impress lay people begs an especially good number of questions.
1. Given all of the variables in a nation's economy can you actually use a basic (or even simultaneous) regression model to "predict" economic outputs like GDP? In addition, during last night's debate Herman Cain bragged about "dynamic modeling," i.e. including assumptions about U.S. economic growth as a constant in the model. How reliable is dynamic modeling? How does one factor in the ups and downs of the economy to date, and are said predictions at all reliable out past a given time horizon? Is this some type of weighted variable?
2. Statistical modeling is useful to the degree that it gives you a better answer than a guess would have produced otherwise. If a basic regression analysis has revealed the genius of Cain's 9-9-9 plan, why are there ever business cycles, or ups and downs, or system shocks like the Great Recession? How and why do these models fail? And when have they ever been successful at making accurate predictions?
3. Does the ecological fallacy apply to econometrics? I primarily work with cultural texts and theory. But I have always been suspicious of grand models that purport to make predictions about the economy. Is this suspicion misplaced?
4. Finally, is Mr. Lowrie qualified to be offering these types of analyses? When he says, "“I have an accounting degree, and I’m an investment adviser. I’ve never hung out in a faculty lounge,” I get more than a bit worried. Are you?

19 comments:

Oh Crap said...

nein
nein
nein

nicoleandmaggie said...

A simple regression analysis won't take into account that behavior changes because of taxes... dynamic programming solely taking growth into account won't either.

To actually estimate what's going on you would need a much more sophisticated analysis that makes assumptions about what would happen to spending when sales tax is 9% etc. Obviously we're no good at predicting and not in agreement already about what happens to spending at slightly higher and lower marginal tax rates. Corporate tax rates as well require a lot of assumptions about how companies will move etc. Any dynamic programming or regression model is still going to be making out-of-sample predictions on things that are pretty far out of sample.

As for feasibility-- no way that the US is going to go for 9% sales taxes. The idea is ridiculous. People would stop spending when faced with a 17% or 18% sales tax. That would also cause hugely negative spillovers to state and local entities that rely on sales taxes, and they would be forced to add or increase income taxes to make up for it which would be very messy.

Next semester I get to explain to my students why marginal tax rates make sense. I've never had anybody who didn't get it after the brief explanation, but I'm guessing this year there will be more argument. But it's all in the marginal utility curve... a dollar means less to me than to my students and more to me than to Bill Gates. Therefore Gates will be just as happy at a higher tax rate than me, and I'll be just as happy at a higher tax rate than my students. When you look in terms of utility, things are fair. The question is just what to set those different numbers at.

Anonymous said...

The whole "I just have an accounting degree ... I didn't hang out in faculty louges" is a vein of the anti - intellectual, anti elitist schtick popular w/ the tea party & republicans. I'm willing to bet dude has a masters (common bar for the top jobs for accounting majors) and did not attend Podunk U. It's disingenuous for him to promote the plan as if he is "one of the regular people" when he is not.

We have 30 years of data showing lower taxes do not increase spending, hiring, etc. No predictive math can argue with existing data.

Thrasher said...

CD,

Like you and many others I can count my money, use a ATM and pretend I know all of the economic data. charts, summaries I get from my broker when he is chatting me up for more funds to invest..

Yet when politicians orchestrate platforms which are geared to get them elected than I always have reservations about any proposals they offer up to the electorate from economic paradigms to social programs and health plan projections..

Cain is pitching a game to get nominated like his co-horts it is nothing more than political propaganda designed to influence and create behavior in the voting booth..

999 could be fo fo fo ( circa Moses Malone)...lol,lol,lol

chaunceydevega said...

@Ohcrap. That counter to his argument is so obvious, I hope someone picks it up and runs with it in a debate.

@Nicoleandmaggie. I knew someone would have an answer. So is Cain's model misspecified? And given that the behavior of one of the independent variables (consumer spending) changes relative to another (the sales tax) does the 9-9-9 plan violate the basic rules of a simple OLS?

Not that it will matter, tax cuts and simple answers a religion for contemporary conservatives. I have one in a class I am teaching, nice guy, but totally talking point who can't accept that tax cuts have never (save for 2 times on margins) paid for themselves. He goes back to my opinion is... I say do you have an opinion on gravity? Because we are talking about facts not belief. He still doesn't get it.

@Anon. Stop with your sacrilege! See above.

@Thrasher. I hear you. One would think that folks vote their pocketsbooks--many do, and the public largely sees the Tea Party GOP as the party of the big business and the rich. But there are enough "values voters" and other authoritarian low information types to legitimize the viciously regressive nonsense offered by Herbie Cain.

John Kurman said...

For attention conservation, a summary: Herman Cain is a bullshit artist, a fucking liar, and his 9-9-9 plan is pure bullshit. He is spewing this crap otu knowing that hardly anyone knows what "regression analysis" or "dynamic modelling" means, but, within the context of his argument, it exactly the same as technojargon in a science fiction movie. In short, Herman Cain is an asshole who is trying to bullshit you. He should be called on it again and again.

1) Modelling the US economy to determine a dependent variable like GDP (and think of GDP as temperature in a climate model, temperature is the average of a million trillion little molecules bumping into each other, which properly would be modeled as such), requires a simulation on the order of complexity as those used in global climate models. Hundreds of parameters, dozens of variables, each connected in a thousand dependencies, and all completely unpredictable after just a few iterations. (The Wharton School once developed a 30,000 equation model, which promptly took a shit and fell flat on its face). You've heard of Chaos Theory? This is it. Modelling the proton/proton collisions at the LHC is child's play compared to modelling the national economy. Anyone who thinks that a hack bean counter can do it is an simpleton and an asshole.
2) No one has ever produced a statistical model that successfully simulates the US economy. Period, The closest simulation is the market itself, and it obviously is a shitty model. The markets can and do produce false values for prices. Witness any crash. That's when market gets the prices right. That's why it is called a "correction".
3) There is no 3).
4) Yes.
5) Mr. Lowrie is a dipshit and should be forced to bet his testicles on his economic plan. Let's see what he says when his nads are on the chopping block.

chaunceydevega said...

@John. Bringing the heat. Calm down we need you long into the fight ;)

I didn't know about the Wharton deal I will look that up to learn more. The market is the simulation is a great distillation. I am going to "borrow" it. Once folks look into his nein-nein-nein plan will joe q. public see it for what it is? An easy counter would be to have a commercial with basic staples like milk, bread, cheese, baby food, and a before and after price if Cain gets his way. You can tier it so there are other commercials showing cars, computers, ipods, homes and such with a sales tax too.

John Kurman said...

http://www.billschmalfeldt.com/2011/10/time-to-call-bullshit-on-herman-cains-9-9-9-nonsense/

nicoleandmaggie said...

You don't actually need a 30K variable model to determine how X affects Y. The type of equation being run wouldn't be to predict Y, but to predict how X changes Y. For that you only need to control for all of the things that directly affect both X and Y at the same time (omitted variables).

The problem here being that there isn't a linear relationship between X and Y. You are correct that OLS isn't going to work. In theory you could do a dynamic programming model or a structural programming model that would better determine optimal tax rates, but nobody has managed to do that yet even with small changes. (And GDP isn't the only thing that's going to need to be made dynamic in a dynamic programming model-- there are a lot of feed-back effects). If it were easy, gov't might already be doing it.

And yes, things change about the economy all the time, invalidating models... we didn't know stagflation could exist until it did back in the 70s, totally changing our ideas about the tradeoff between inflation and unemployment. Who knew that curve could shift?

It drives me nuts when people can't separate fact from opinion. One thing we do in our program is pound into our students' heads the difference. Some of the professors are nicer about it than others, but a mixture of autocratic and velvet glove seems to work on most of them by the end of their second year. Which is nice, I feel that's kind of our public service.

Recall said...

Wait, we're supposed to assume that Cain's model actually exists?

CNu said...

Ed Dunn eviscerates all-a-dat anti-Hermanator gas....,

Herman Cain was the son of a working Black father who was able to send his son Herman to Morehouse. After Morehouse, Herman Cain was entrepreneurial and opened up a Godfather Pizza restaurant chain. I knew about Godfather’s Pizza back in the day and I remember reading these Black magazines that brag about Herman Cain accomplishments. Now today, we got some bitch Black people trying to call Herman Cain a “lawn jockey” and other names like an idiot.

This is the thing I want all of you guys to notice – Herman Cain already did for self and got his coming up – what the f*ck most of these other people talking sh*t done, real talk? Someone said Herman Cain didn’t do anything for the Black community and that is a damn lie – Herman Cain done a lot for the Black community and some of yall need to shut the f*ck up if you don’t know facts.

CNu said...

(continued) Herman Cain was a job creator and created jobs and hired Blacks – how many of these other cats talking ish about Herman Cain started anything and created any jobs for Blacks? You cannot even tell me President Obama created jobs for Blacks and you know you can’t. Bottom line is some of you ignorant ass Black people is hating Herman Cain when your sorry ass need to look at your own damn self and realize Mr. Cain did more than your simple Black ass ever contributed to the game – remember that the next time you want to call him a “lawn jockey” or whatever comment your do-nothing ass got to say about him.

Now this is the thing I noticed – the same cats that want to sh*t on Herman Cain are the same cats who are looking up to and ass kissing other Black opportunists out there who are 100x worst than Herman Cain. The guy who run American Express – what the f*ck he done for bringing financial products to the Black community? Yall just admire the AMEX guy because you read about him somewhere and noticed he Black. Stanley O’Neal from Merrill Lynch – he negotiated a nice parachute package before running that company into the ground and also contributed to the Wall Street collapse of 2008. And the Black CEO of Darden that have Red Lobster chains – what economic plans he created in the Black community knowing Black people patronage the hell out of cheddar biscuits? Oh and how can I forget Rob Parsons, if that is his name from Time Warner? You want to know where he working at now? If you said “Citibank” then you fully understand what I’m talking about with the hypocrisy going on around here.

Like I said before, Herman Cain is attacked for his views but you cats ain’t got nothing to say about a closet homosexual promoting destructive Black relationship imagery wrapped up in his twisted view of Christ cross-dressing as an old Black woman. Or none of yall got anything to say about some Black woman that promotes negative music and movies on urban radio stations and cable television knowing it is affecting Black kids negatively on a mass scale day in and day out. Herman Cain ain’t never did no f*cking evil like that in his life towards Black people.

And to top it all off, Russell Simmons who peddled a fee-infested prepaid card to the poor Black community is walking around Occupy Wall Street as if he ain’t one of the cats we should be protesting against. See the hypocrisy here and the hypocrisy is based in a sheep following ignorant mentality. If the Black bloggers say Herman Cain is bad, then these Black following losers want to say something bad about Herman Cain and not think for themselves.

If you noticed the talk about Herman Cain, you would notice the people attacking him are the same sorry ass ineffectual liberal persona who don’t have any economic or empowerment solutions and focus on using emotional appeal to sway Black people. Liberal writers like that Ofari Earl Hutchinson cornball ass or the “Reverend” Al Sharpton preaching ass but you damn sure don’t got real cats with real business/social science skills making real comments because the data shows that Herman Cain is not this crazed evil Black person cats want to make him out to be.

Thrasher said...

CNu.

Calm down...Make a difference your self

CNu said...

free jiggaboo..,

my "difference" will be the subject of the keynote address at Qualcomm's national conference on one-to-one computing initiatives in education this week.

just like my difference became the focal point of the googlecity brainstorming event two weeks ago.

you issue, event, and personality driven fans of the 2nd and 3rd line - have no earthly idea...,

Thrasher said...

CNu,

What a petty nobody you are begging for recognition and attention...Get Lost! you ordinary forgettable nobody...lol,lol,lol

CNu said...

lol@free.jiggaboos craving a mid-morning mouthful of schweddy balls...,

Thrasher said...

CNu aka Jim,

Gotcha Jim..Intellectual coward hiding behind an alias..Lol Lol Lol

fred c said...

Well, that was a fascinating couple of column inches.

Throckmorton Treebeeizle said...

CdV have you seen this? I think it may be what you're looking for:

http://www.jaredbernsteinblog.com/9-9-9-in-one-really-long-graph/