Tuesday, March 03, 2015

There's Legal Analysis, and then there's REAL Analysis

I usually enjoy Eugene Volokh's stuff. Though his move to the WP was disconcerting, I got over it.

But Volokh has a BIG swing and a miss in describing a 'growing' rationale for an argument that the Supreme Court should revist/reverse previous court opinions and enforce Interstate Sales Tax collection the way the greedy little state and local politicos/taxmen WANT it enforced.

The ' rationale' goes like this:
This argument has grown stronger, and the cause more urgent, with time. When the Court decided Quill, mail-order sales in the United States totaled $180 billion. 504 U.S., at 329 (White, J., concurring in part and dissenting in part). But in 1992, the Internet was in its infancy. By 2008, e-commerce sales alone totaled $3.16 trillion per year in the United States.
Sorry, those are RAW numbers. What does the trend look like factoring the overall economic environment.  I'm  assuming internet sales have supplanted most catalog sales:  a pretty safe assumption I believe.  And even if it's not, if someone can compare them directly across the timeline--well then! So can I.

Here's those two numbers Volokh used, lined up against GDP figures for private and government consumption and investment:
Correction: I had fat-fingered $3.16T as $3.6T , Heh. Now my point is even 'Truer'

The big number for 2008 compared to the little number in 1992 overstates the growth in interstate sales. Complaints about 'missing revenues' from politicians reeks of unwillingness to compete, and worse a certain sense of entitlement to other people's dollars--no matter what the source--to feed the allmighty 'state'.

A rising tide lifts all boats

As the 'Barf Box' at the bottom of the chart indicates, NO ONE ever seems to ever ask the question as to what entities receiving all that internet revenue spend it on -and where? I would normally presume states with the better internet-friendly tax laws would benefit most and the local governments would be happy about it. But I realize we're dealing with people whose lives are often immune to the direct effects of the real economy, and I think more than a few resent the denial of an "opportunity for graft".

And I'm not buying any argument that the Local and State governments NEED those taxes on internet revenues: All hail the rise of the "Social Spending-Entitlement-Complex...
  State and local government is THE 'Growth Sector' of government. 


When the data is on your side, argue the data....

Sorry Eugene.

6 comments:

S O said...

Your first diagram doesn't make much sense:

One needs to look at a logarithmic scale diagram to get an idea about "growth rates". Linear approximations are nonsensical as an illustration of "growth rates".
Create the diagram with the "$B" axis logarithmic if you want to have a relevant illustration!

Yet even IF you had that graphic and even IF it showed what you want it to show; it would be largely irrelevant to the topic. The states have a poor fiscal situation and require much federal aid even though they do little more than the basics such as paying for schools, state police and roads. Their budgets are legal and legitimate, created through a due political process by their legislative branches. This constitutes legally and legitimately a need for (more) revenues. Your private opinion about the spending is a minority opinion obviously (otherwise there would be different spending), and it's the majority that rules in a democracy.

SMSgt Mac said...

First, I didn't make the comparison between 1992 catalog sales and 2008 internet sales to illustrate a growth in intrastate economic activity. Volokh did. I'm pointing out the growth rate in that kind of personal economic activity was lower than growth in personal economic activity overall, so it is NOT supplanting traditional sales tax revenue sources overall, and I suggest that the revenues from internet sales actually stimulate other local activity but since it cannot be directly counted, the positive effect is ignored. I used to own a web design business; 98% of my Customers were small business owners who spent most of their income locally.
The lines I drew are called 'trendlines' for a reason. They're actually calculated, except for ease, format, and presentation purposes I chose to draw them.
My second point is that because an economic activity exists, does not mean a state or locality SHOULD or MUST tax it in lieu making themselves more efficient, and that the suppressing the sense of entitlement by governments (such as manifests itself in annexation practices - don't get me started) to this impulse should always be the first priority. My youngest brother is very high up in a major metropolitan city government and is a past state city management and ICMA poohbah. I've seen the effect of 'where you sit is where you stand' up close and personal. It affects even the most libertarian at heart. The impulse in a 'poor' city to find more sources for revenues in lieu of being fiscally responsible is too easy to follow, and the path to fiscal responsibility is harder. That the public is too stupid to elect people who will place responsible policies in place is at least partly the responsibility of the politicians and bureaucrats who have in places been all too happy to be irresponsible in the past hoping the consequences never catch up with them. The mechanism that prevents people from overspending and generates consequences for those who don't too often doesn't apply to government.
If you read my stuff regularly, you would also know that I know much of state and local government activity is done on the gov't teat. (http://elementsofpower.blogspot.com/2014/10/there-is-no-military-industrial-complex.html) As someone living in a state paying more federal taxes that it receives in federal benefits, I do not believe I should be subsidizing Utopian fantasies or inefficiencies elsewhere.

S O said...

Look, "trend lines" or not, growth of this nature is exponential. Any "trend line" for exponential growth can only be drawn in a logarithmised graph. The "trend lines" in the graphic here are about as purposeful and correct as a trend circle. The correct geometric shape for a trend in here would be a curve. Hence the use of logarithmic graphs, which make trends visible as lines.

The 2nd graph here shows this nicely:
http://mainlymacro.blogspot.de/2015/03/a-campaign-based-on-ignorance_6.html

You titled the text with a promise to provide better analysis than someone else, but you committed a mistake that would give you a failing grade at a university econ department.

SMSgt Mac said...

My two points were that
1) the data/argumentation Volokh used did not support the identification of some grand shift in economic activity that was shorting the taxman and
2) the personal spending portion and the government spending portion both show very linear trends over the 1992-2008 time frame (and as there are only two data points for trans-state economic activities you can't really correlate those activities to anything else to mean anything much less an 'accurate' trendline).

We could torture the data to death, but as we are dealing with rather gross summaries of data, it would only be torture and not extraction of any deeper meaning.
Within the years examined, entire industries arose and fell, shifts to services and some manufacturing occurred and shifts away from other industries and services happened as well. You cannot look at one economic element as a standalone item and declare it's impact on an economy (or tax base) is 'x'. That is the classic economics 'all other things held equal' (that never are) line of thinking.
your referenced chart about Britain's economics had no bearing on the topic at hand, and didn't showing anything 'exponential' either. (why am I always running into conflicts over 'exponential'?: http://defensetech.org/2011/02/20/the-air-force-wants-a-new-bunker-buster-bomb-for-the-f-35/)
Anyhooo....
For the data we are looking at, only the government spending trend-line R^2 value improves significantly using a exponential or polynomial fit. The personal spending data linear trend-line fits better than the exponential trend-line does for the government spending for all that bother.
In sum, You've raised no coherent argument indicating where I am wrong or why.

S O said...

I linked the graph because it shows an example of data presentation.
You not seeing anything exponential in that graph only confirms that you don't understand the math in this: It had a logarithmic scale and a strong linear trend. A linear trend in a logarithmic scale shows exponential growth.

Same problem with you thinking there's a linear trend in the diagrams here: You'd see there's an exponential trend if you only converted to a log scale and knew how to read it.

I get that you probably didn't have training in this specific math, but you should catch up with log scale diagrams and the (actually quite simple) math behind it before pretending to give "REAL analysis" with such diagrams included.

It doesn't matter much whether your original points were correct or not - it's like thinking 5x6 is more than 4x3 because 5x6 is 50 and 4x3 is only 40.

http://tinyurl.com/l7skoyh

SMSgt Mac said...


I let your little 'training in specific math'comment ride at the time because I thought it was funny. But since you're leaving now, I should probably close this loop. The 'tinyurl' you left doesn't work anymore but it didn't make the point you had hoped. The point it DID make was that log plots are good for viewing variations within trends to tease out any internal data. If you had been a regular reader, you would have read in 2012:

"The reader should immediately note that the chart above is plotted in a ‘Log-Log’ format: both chart axes are plotted using a logarithmic scale. This has the effect of allowing the clear display of early values, where wider variations in data are to be expected and of showing trends (and deviations from same) more accurately. As more statistically relevant data is accumulated, on through to where the system maturity point is selected for determination as to whether or not the system meets the reliability requirement, the deviation from the mean value should lessen (more about that later)."
From : http://elementsofpower.blogspot.com/2012/07/strange-silence-on-gao-f-35-june-2012.html