Restart : tWo

An excellent compilation about math and the credit crunch

noreply@blogger.com (---) — Fri, 16 Jan 2009 00:05:00 +0000

If you are interested in the credit crunch and the involved mathematics, mathematicians, quants etc. you should check out the excellent compilation on BBC News. You will find a lot of interesting interviews and background information nicely compiled.

And if I find some time, I will later write about my own experience with toxic waste a.k.a. the pump of death a.k.a. ratchet swaps and alike that have such highly skewed payoff profiles that coupon payments might rise up to several times of the notional. In well regulated countries products like this might not end up in the hands UHNWI with limited financial knowledge but there are less regulated countries as well... stay tuned for the full story (sometime later)?!

On: IBM Completes Purchase of ILOG (UPDATE)

noreply@blogger.com (---) — Thu, 15 Jan 2009 07:55:00 +0000

I just read the press release of IBM's acquisition of ILOG again (see also Michael Trick's blog) and I am truly disappointed that the general tendency at ILOG to focus (to put it in nice words) on BRMS (which in fact means the ILOG JRules Technology) rather than optimization seems to continue also under IBM's ownership. In fact, compared to former press releases optimization took a slightly more prominent role but effectively the magic words are still SOA and BRMS.

I personally have doubts that JRules (e.g. used to calculate car insurances and other stuff) can be an edge for companies in global competition. If you haven't been bored today you can also check out Business Rules on Wikipedia. The often cited flexibility and the speed with which changes to contracts, policies, etc can be incorporated using JRules without relying on the IT departments is a blunt joke. Yeah there are white papers and success stories but who has ever seen a fully fledged rules application knows that only rather large companies can afford to have a dedicated person managing the system which is necessary to maintain momentum. Otherwise it is effectively a rules consultant implementing these changes moving the dependency from the IT departments to some external contractor - right, if it sounds like equity tapping it is equity tapping. Anyway, that's just my personal opinion...

Given the potential of optimization technology (yeah I am preaching right here) to make a real difference in production, supply chain management, etc in particular in times of economic distress I hope that someone else fills the void left behind. I guess the problem is still that people don't understand optimization and when you talk to ibanks and private equity funds (prime customers for optimization technology) they don't really care about the operations aspect of the companies that they acquire/fund - it is all about financials. Operational aspects are too dirty, too slow, and you actually have to get into the details before you can increase efficiency. Also optimization is often put into the same category as IT solutions and thus the companies are not willing to pay the premium. And you know: IT projects always fail... ;-)

I guess that this is also one of the reasons why Gu, Bixby, and Rothberg (all involved in CPLEX development at ILOG) founded Gurobi Optimization, i.e. to fill the void.

UPDATE: I just learned from a trustworthy source (sure they all are ;-) - gossip, gossip, gossip) that the performance of the gurobi code especially for MIPs should be amazing. If true, not too surprising in terms of overall feasibility given that three gods from solver olympia worked on it, but definitely in terms of speed. The guys haven't had a lot of time... This would be really amazing!

The bird is the word

noreply@blogger.com (---) — Tue, 13 Jan 2009 10:06:00 +0000

And another one ;-)

This is my ...

noreply@blogger.com (---) — Tue, 13 Jan 2009 09:51:00 +0000

Just saw this one. Reminds me of full metal jacket ;-)

Here is the original:

GLPK 4.35 released

noreply@blogger.com (---) — Sat, 10 Jan 2009 21:58:00 +0000

The new version of glpk includes new api routines for dealing with graphs:

GLPK 4.35 -- Release Information
********************************

Release date: Jan 09, 2009

GLPK (GNU Linear Programming Kit) is intended for solving large-scale
linear programming (LP), mixed integer linear programming (MIP), and
other related problems. It is a set of routines written in ANSI C and
organized as a callable library.

In this release:

The following new API routines were included in the package:

glp_create_graph create graph
glp_set_graph_name assign (change) graph name
glp_add_vertices add new vertices to graph
glp_add_arc add new arc to graph
glp_erase_graph erase graph content
glp_delete_graph delete graph
glp_read_mincost read minimum cost flow problem data in
DIMACS format
glp_write_mincost write minimum cost flow problem data in
DIMACS format
glp_mincost_lp convert minimum cost flow problem to LP
glp_netgen Klingman's network problem generator
glp_gridgen grid-like network problem generator
glp_read_maxflow read maximum flow problem data in DIMACS
format
glp_write_maxflow write maximum flow problem data in DIMACS
format
glp_maxflow_lp convert maximum flow problem to LP
glp_rmfgen Goldfarb's maximum flow problem generator

For detailed description of these new routines and related data
structures see chapter "Graph and Network API Routines" in a new
edition of the reference manual included in the distribution.

A minor change were made in the internal routine xputc. Thanks
to Luiz Bettoni for suggestion.

A minor bug was fixed in the internal routine mpl_fn_time2str.
Thanks to Stefan Vigerske for bug report.

See GLPK web page at .

GLPK distribution can be ftp'ed from or
from some mirror ftp sites; see .

Awesome cost-cutting measure at US Airways

noreply@blogger.com (---) — Mon, 05 Jan 2009 16:54:00 +0000

US Airways apparently has to fight a huge cost pressure so that they devised a new, impressive way of saving costs:

U.S. Airways has taken the coat hangers out of its planes.

Read the full story on the "Free the Hangers" (freakonomics blog). That sounds like a lame excuse to me. I remember, once in Chicago, I asked a waitress if she could move two tables together for us as we were a slightly bigger group and she answered, she cannot do this "for security reasons"... But arguing with cost savings is even better. Probably the hangers were made of a precious metal so they sold them on the commodities market or they turned them into fuel using a new alien technology.

The swimming naked awards

noreply@blogger.com (---) — Sun, 04 Jan 2009 23:04:00 +0000

Einstein said "The secret to creativity is knowing how to hide your sources." I will refrain from following that advice although it seems to be common practice (especially in certain areas of science) lately. Therefore: Thanks Aurelie - here is her original post. And now enjoy the following gem which was published on The Economist:

The swimming naked awards

get mobile or rather get screwed?

noreply@blogger.com (---) — Sun, 04 Jan 2009 10:04:00 +0000

Spending few days over Christmas and New Year's at home is often kind of a drawback for my research. While *trying* to think about Tseitin tautologies and wondering why there are certain papers that (surprisingly?) lack proofs for their unexpected assertions (no, I won't provide the author's name) I am "forced" to listen to stupid ads on Viva, MTV, etc. Probably the most annoying one is the one from getmobile.tv: Just get two mobile contracts with two cool phones and you will also get a brand laptop and a Playstation 3 for free! Below the original ad so that you can share my frustration (German only. sorry! - but I bet that other countries are poisoned by this type of toxic waste as well):

After having given up on the tautologies due to severe brain damage inflicted by audio-visual rape of my neurons, I decided to take a slightly closer look. After all, there is no free lunch! Especially as this ad machinery and all its overhead of superfluous business engagement have to be financed somehow. So I dug through the internet. The price for the PS3, the laptop, and the two phones together roughly EUR 780,00. That is the consumer price. I bet that those guys were able to negotiate a better deal so that they can knock-off say 15% to 20% of that price. But, ok, let's not care about their business case at this point. If you look at the contracts now (you actually have to take two of those for minimum two years) you have a base fee of EUR 4,95 plus another EUR 5 for the Jamba flat subscription (GREAT!!! I can get the I-can-see-you-naked scanner, the cell phone solarium, and the cool new flash logo with a fading flower depending on my battery charge level...) per month. Also we have to use our brand new contracts for at least a minute a day. They charge at least EUR 0,19 per minute. A closer look reveals that this rate only applies on the weekend and holidays. The lowest rate on a regular day is EUR 0,49 per minute and that only applies for calls into the provider's own network. Assuming a rather balanced network distribution and that you don't choose your friends based on their network provider EUR 0,56 as a weighted average per minute (weighted over the year) is more appropriate. That makes a total cost per contract per year of EUR 323,02. As you have two of those contracts and that for at least two years that is EUR 1292,07 minimum cost. And for that you haven't really used your phone (except for the compulsory one minute a day) and the conditions of the contracts are bad compared to other offers.

Also, the laptop comes with pre-installed Linux (personally I almost appreciate this) but effectively they saved the money for the Windows license which means that you have to pay for this separately if you want to use Windows on your machine.

Now it is your turn - get mobile or get screwed.

While searching for the ad on youtube, I also found this other clip discussing the deal. Definitely, check it out (sorry, German again):

Science - it works, bitches!

noreply@blogger.com (---) — Sat, 03 Jan 2009 20:23:00 +0000

Found this one on xkcd.com:

Dinner for one - lego version

noreply@blogger.com (---) — Fri, 02 Jan 2009 11:29:00 +0000

Slightly off-topic and two days too late. Anyway, I wanted to share it with you: The "Dinner for one - lego version"

Here is the original for a (slightly unfair) comparison:

Ten Signs a Claimed Mathematical Breakthrough is Wrong

noreply@blogger.com (---) — Thu, 01 Jan 2009 11:17:00 +0000

I love this blog post from Scott Aaronson:

Ten Signs a Claimed Mathematical Breakthrough is Wrong

The all-time favorite: "The authors don't use TeX" or "The paper doesn't build on (or in some cases even refer to) any previous work"

Kind of reminds me of all the "papers" about angle trisection (with an un-marked straightedge and a compass). I remember one particular "paper" that was written in prose and the author celebrated himself in every sentence "Galois has been beaten the second time"...

Also check out the "Alternative-science respectability checklist" - priceless

Garrett Lisi, E8, and the Theory of Everything

noreply@blogger.com (---) — Tue, 30 Dec 2008 19:00:00 +0000

What happened to Garrett Lisi? In November 2007 he published the much discussed paper "An Exceptionally Simple Theory of Everything" (arXiv:0711.0770):

All fields of the standard model and gravity are unified as an E8 principal bundle connection. A non-compact real form of the E8 Lie algebra has G2 and F4 subalgebras which break down to strong su(3), electroweak su(2) x u(1), gravitational so(3,1), the frame-Higgs, and three generations of fermions related by triality. The interactions and dynamics of these 1-form and Grassmann valued parts of an E8 superconnection are described by the curvature and action over a four dimensional base manifold.

In February 2008 he gave a TED lecture about his work (see below)...

So what's the verdict? As I am not a physicist and bearing the comparative advantage in mind, I would like to pass on this question to someone of you guys!

Zeitgeist - another controversial movie

noreply@blogger.com (---) — Tue, 30 Dec 2008 18:39:00 +0000

And another controversial movie - actually two:

Zeitgeist

and

Zeitgeist: Addendum

What do you think? Opinions welcome!

The Story about Stuff

noreply@blogger.com (---) — Tue, 30 Dec 2008 18:30:00 +0000

A bit older but definitely worth a look: "The story of stuff". On the homepage it says:

From its extraction through sale, use and disposal, all the stuff in our lives affects communities at home and abroad, yet most of this is hidden from view. The Story of Stuff is a 20-minute, fast-paced, fact-filled look at the underside of our production and consumption patterns.

What do you think about it? Especially from an economic and operations research point of view. Let me know!

Samuel Israel III surrendered - let's look at his strategy!

noreply@blogger.com (---) — Thu, 03 Jul 2008 06:42:00 +0000

I am very disappointed: Samuel Israel surrendered after being fugitive for 23 days. That could have been the story for one of the greatest movies in a few years. Just imagine, the bad hedge-fund manager on the run. Of course, everything was a misunderstanding (we need that for Hollywood's paradigm). Helicopters chasing him. Burning cars, bullets, guns, bought new identifies, family or friends betraying him... All that stuff. And, based on a real story!! That would have been a great movie!

But surrendering? Definitely 0 points for style!! But let's have a quick look at his overall strategy according to WSJ.

Staging suicide in general is a good idea. Keeps other people busy while you are fleeing. +100 strategy points for that.
Hiding in the mountains at a camp site in an RV which you have been using there already weeks in advance. Extra smart! Nobody gets suspicious that you suddenly show up. Another +200 points for that.
Acting cool and not looking like someone being on the run. Another +100 points for that.

These were the positive things so far but now let us have a look at the operational implementation of that:

Calling mum. For what reason ever. Either to tell her that her "boy" is alright, or to ask for advice: Bad idea. Not only that she might try to convince you to turn yourself in, as she did your mum is also one of the first places to look for you. Definitely -500 points for that.
Asking your cute girlfriend for planning and implementation support. Well, guess: Bad idea as well. They talk! Whatever they promise you in advance, they talk. And you don't have to turn up the heat too much for that. She helped him with the planning and preparing the RV. She even told the police the color and the fact that it had a blue stripe. -500 points for that.
Having yourself picked up by a friend from the location where you pretend to commit suicide. Not a smart idea either, especially when it was captured on a video footage. They will catch your friend and bring him in for interrogation. Same as with your girl friend: They will talk!! It might take some extra heat, but they will! -500 points for that as well.

So the overall statistic looks quite bad: Great Idea, terrible implementation and definitely NO points for style!!

I guess we have to wait for the next fugitive...

The MPG illusion

noreply@blogger.com (---) — Sat, 28 Jun 2008 20:18:00 +0000

A few days ago there was an interesting article in Science ( June 20, 2008 (320) ) about the 'MPG illusion', the deception arising from the fact that the MPG is inversely proportional to gallons per mile. This effect, which is widely known to mathematicians (and other scientists), can lead to wrong conclusions when comparing gas savings in terms of MPG.

Since there are already several blog posts on this topic out there I confine myself to citing the version I found on bunniestudios.com:

Suppose you had a household with two cars, and each car needs to be driven 10,000 miles per year. One car consumes 34 MPG, and the other car consumes 18 MPG. Since gas is expensive, you want to replace one car. Because of utility constraints, you have two choices:

Replace the 34 MPG car with a 50 MPG car — a 16 MPG improvement

Replace the 18 MPG car with a 28 MPG car — a 10 MPG improvement

Which car replacement would save you the most gas?

Normally, I consider myself not bad with quantitative comparisons like this, yet initially I picked the answer of replacing the 34 MPG car with the 50 MPG car based on the superior 16 MPG improvement. Another seemingly more analytical approach also leads to the same conclusion: 50 + 18 MPG giving a 34 MPG household average seems more efficient than 34 + 28 MPG giving a 31 MPG household average.

This very interesting article in Science, “The MPG Illusion” by Richard P. Larrick and Jack B. Soll at the Fuqua School of Business in Duke University (Vol 320, June 20, 2008, p. 1593), points out the mathematically obvious truth that gas used per mile is inversely proportional to miles per gallon, which means that you have a steeper slope at lower MPG ratings, and diminishing returns at higher MPG ratings.

Read the full blog post on bunniestudios.com...

Hedging the gas price - the cost of insurance

noreply@blogger.com (---) — Sat, 21 Jun 2008 17:56:00 +0000

I was wondering lately what it would cost to hedge my fuel costs by buying options on oil futures. Chrysler offered something similar with its $2.99 per gallon deal a few weeks ago - I already estimated that cost in a previous post. So I set out this afternoon and did some research, i.e. collecting data from the Energy Information Administration, doing some (dull) regressions, and analyzing option prices. The result is a neat little table (see below) that contains rough estimates for the insurance costs depending on maturity (of the underlying future) and level of protection above which the "insurance" kicks in.

The insurance itself works as follows: You buy call options on oil futures (see below) and when the gas price rises above our protection level the value of the options we bought will have increased and will offset the additional fuel cost (by selling them). Note, that I do not consider transaction costs here which might be significant (especially if you only sell a small volume). Neither, did I consider discounting of the future contracts or convenience yield effects as their impact here is limited although they are important for longer maturities in general.

So how do we come up with some reasonable numbers for costs? The idea is pretty straightforward:

a) Relate gas prices to oil prices: Since a barrel of oil contains 42 gallons the relation should be straightforward from the theoretical point of view. Doing a regression, depending on the grade and the region in the US we obtain that approximately

gas-price-per-gallon (y) = 1.45 + 0.024 oil-price (x)

I actually adjusted the coefficients a bit to reflect the fact that "1/42 * price-per-barrel" is a lower bound for the price of a gallon.

b) Assume that an option is $1 in-the-money. According to the above formula the gas price p.g. increased by $0.024. Therefore we need 0.024 options to offset that increase by selling off the option.

Combining that with latest quotes for call options on oil futures we can calculate:

1) The expected gas price given the underlying oil price. That is the information in the top rows and is the strike price for the option and our protection level.

2) We can calculate the hedging cost for a gallon depending on the maturity of the underlying future contract. Implicitly that determines the expiration of the option as well as it usually expires a few weeks earlier.

The two tables below contain the absolute and the relative (to the protection level) hedging costs.

&amp;amp;amp;amp;amp;amp;amp;nbsp;

For example insuring with a protection level of $4.66 a gallon would cost us $0.22 or 4.62% extra (per gallon) if we want this protection for buying fuel until 08/2009. If we want to buy protection for a later point in time, say until 02/2009 we would pay $0.42 or 9.05% premium.

In order to use this strategy more efficiently for hedging exposure we could buy a mix of options to reflect the fact that if we want to fill up the car next week we do not need options with expiration in 05/2009 which are almost twice as expensive as the ones that expire in 08/2009 as we pay a higher time value.

Or we could just drive less... ;-)

The Future of Operations Research revisited

noreply@blogger.com (---) — Fri, 20 Jun 2008 09:46:00 +0000

A few days ago I read an article on Bnet about "The Future of Operations Research". In this article Lee W. Schruben from UC Berkeley argues that current models are often based on static (and partly unrealistic) assumptions and do not include forecasting and risk analysis enough.

BNET: So what is the right approach to OR modeling?
Schruben: We have to integrate forecasting and risk analysis with OR modeling. We have to integrate models with dynamic market information and forecasting. Simulation is the workhorse to do this, because it can handle that kind of dynamic complexity, whereas most OR models tend to be optimization, static kind of models.

While he makes a few valid points about the necessary integration of sensible risk parameters, robust forecasting, dynamic models, and that simulation based solutions might be able to address several of these problems I strongly disagree with his disdain of what he calls 'theoretical OR'. He argues:

Schruben: There’s a lot of theoretical OR that has given the field a bad name. This comes from the “managerial insight” section of OR research papers. Most of the insights are either obvious or wrong. And these insights are often couched in such obscure terms that they confuse and disillusion managers. So the theoretical stuff tends to give the field a bad name. But the practical application of OR is the reason we’re still in business. There’s no question that OR in practice has made a huge impact on business.

There is no doubt that most of us are unenthusiastic about 'managerial insights' but a considerable part of Operations Reseach is based on strong theoretical foundations like Integer Programming, which should just serve as one out of several examples here - I prefer to stick to my comparative advantage therefore I chose IP. While being one of the most challenging and interdisciplinary subjects in OR due to its close links to economics, computer science, and engineering, the practical impact of Integer Programming (and all the related disciplines like Linear Programming, Combinatorial Optimization etc.) is enormous. Many real world production and scheduling problems contain highly complex optimization problems (both in the computational but also in the common sense). While the modeling of the underlying assumptions under uncertainty is definitely not easy and I agree with Schruben that we have to pay more attention to these uncertainties I hardly see how we could solve the underlying combinatorial problems without the theoretical progress of the last decades. These problems are usually solved using Integer Programming solvers like CPLEX, XPRESS, glpk, coin, etc and especially modern techniques (e.g. cutting plane procedures, column generation) contributed significantly to the performance of these solvers. The dispatching of the German "gelbe Engel" from ADAC would not be, by far, as efficient as it is today without the application of modern integer programming techniques. Airline scheduling without combinatorial optimization is unthinkable. Complex stacking problems within steel factories can only be solved satisfactory if the underlying PSPACE complete problems can be solved or approximated sufficiently as well.

I understand that Schruben argues from his point of view as a simulation expert and I think he is right to a certain extent but I would appreciate a bit more tolerance towards different approaches to solving real world problems - not everything is a nail if you have a hammer.

Two more must-see TED lectures

noreply@blogger.com (---) — Thu, 19 Jun 2008 21:01:00 +0000

I got another two must-see lectures:

Again, the embedded videos below...

Article: How Not to Survive Your New Boss (BusinessWeek)

noreply@blogger.com (---) — Wed, 18 Jun 2008 07:04:00 +0000

I just read the following interesting article:

How Not to Survive Your New Boss - Hillary Clinton's actions in the days following Barack Obama's historic win would doom her in the corporate world. But then this is politics

BusinessWeek: How Not to Survive Your New Boss

Two must-see google techtalks

noreply@blogger.com (---) — Tue, 17 Jun 2008 07:52:00 +0000

A few days ago I watched two really good google techtalks which I strongly recommend to you:

The Paradox of Choice - Why More is Less

Human Computation

You can also watch them right here.

"The Paradox of Choice - Why More is Less" by Barry Schwartz:

"Human Computation" by Luis Von Ahn:

The introduction of the "Anti-Lereah Predictor (ALP)"

noreply@blogger.com (---) — Thu, 12 Jun 2008 08:15:00 +0000

It is a great country we live in. Everybody can have an opinion! One of my favorites is David Lereah former Chief Economist and Senior VP of the National Association of Realtors (NAR).

He is the authors of the book "Are you missing the Real Estate boom", later titled "Why the Real Estate book will not bust" which is only a small part of his great "predictions".

He also wrote "The rules for growing rich - making money in the new information economy" which appeared in June 2000 right when the dot-com bubble burst.

In the recent past he issued millions of bold statements highlighting the ultimately great prospect of the real estate market and others. He deceived millions of Americans like lemmings being led down a cliff. He is one of those guys whose track record exhibit a very consistent pattern. The pattern of wrong predictions. There are only few people that achieve a statistic significance in their anti-predictions, i.e the negation of their predictions. He might be one of those. Just consider this chart which I found on the David Lereah Watch:

Therefore we should profit from his consistent failure in predicting the markets by introducing the Anti-Lereah Predictor (ALP) which is the semantic negation of every prediction he will make in the future. Being extremely simple this indicator/predictor would have delivered a decent performance in the past. Of course, benchmarking against the past tells you only so much - the turkey problem. But let see how the ALP performs in the future.

It was reported that Lereah is currently advising to several Wall Street firms. I would not be too surprised if they figured it out already and use his predictions as null hypothesis.

After all, the Irvine Housing blog reported that he just published a new book:

Enjoy!

Muhammad Yunus speaks at MIT commencement

noreply@blogger.com (---) — Fri, 06 Jun 2008 15:14:00 +0000

Right now, Muhammad Yunus is speaking at the MIT commencement. He is the inventor of so called microlending and microcredits.

His story is incredible - his CV is available at the MIT commencement website (see above). It is an honor for me to listen to this very encouraging, enlightening, stimulating, and inspiring speech. I hope that it will be available later online.

MIT and Harvard celebrate commencement week in Cambridge

noreply@blogger.com (---) — Mon, 02 Jun 2008 04:49:00 +0000

This week is graduation week at MIT and Harvard. As every year prices for accommodation are sky rocketing. But this year I was smarter and I booked a few rooms well ahead of time just to re-rent them now that available rooms are pretty scarce. Applied economics and operations research blended into a nice investment. After all, who is not willing to pay to see their offspring graduating from two of the most prestigious schools in the world? I will have to check with the Coops how many of these "MIT dad/mum" or "Harvard dad/mum" shirts were sold within the graduation week. Probably next year I should create a nice package including accommodation, one "MIT/Harvard dad/mum" T-shirt, and one ticket for the Boston Duck tour.

The funniest thing I heard today was about a Sloan student who went to the Coop to get his graduation robe. Together with his wife he was examining the robes and decided that the doctoral robe was the one that he wanted to go for. After the sales woman told him that the he cannot get this robe because he was getting an MBA and not a PhD he started to argue with the girl. Even telling the story about his proud family coming all the way from India did not help him get the robe of his choice. He failed to grasp the fact that it is not a matter of money but rather a degree earned. Hapless as he was, he sighed, bought the designated robe, and left the store. Poor boy. Did Sloan forgot to teach him money cannot buy you everything?

OK let us rephrase it within the 'One-minute manager' framework by Ph.D. Kenneth Blanchard, M.D. Spencer Johnson (yes, I included the titles because it fits the story so well):

"You want this robe? Then get a PhD." - Easy as that.

But he could get the 18K large MIT ring as a substitute though.

Is Germany facing a multi drain?

noreply@blogger.com (---) — Thu, 22 May 2008 21:55:00 +0000

The recently published report on poverty in Germany (see Spiegel online [1], Die Welt [2]) fueled the perennial discussion about increasing taxes for the rich. The report states that 13% of the Germans earn less than $1,121 per month, which is also the report's definition of being poor. Additional 13% of the German population manage to avoid poverty only due to the German social welfare system. The German SPD (the left wing socialist party) immediately demanded the re-introduction of the 'wealth tax' which would furthermore increase the tax load on better earning people.

Additionally, for quite some time now the press has been full of reports about qualified Germans leaving their home country and pursuing career opportunities outside of Germany ([3], [4], [5], [6]). At the same time the German industry is struggling to fill its open positions.

So on one hand we have increased poverty in Germany which might result in increased taxes for the better earning people, and on the other hand we have an increase in the number of qualified and better earning people which leave the country. So I was wondering: What is the effect of these two (probably) interdependent developments?

So I looked up some data (freely available from German Bureau of Statistics). The data basis is a bit too small for a complete analysis but the results still might provide some (however partial) insight into the current situation. Additionally the income tax data only covers the years until 2001 and is therefore rather outdated. Nevertheless, since we only use it to understand living costs and the break-up of the taxes into income brackets (or buckets) and additionally assuming that the situation only worsened since 2001, as claimed by the report, it will still suffice for our preliminary analysis.

Let us look at the following observations separately and try to connect them afterwards.

People in Germany complain that living costs are increasing and that they have to live on less.
German industry struggles to fill open positions and at the same time well-trained people leave the country.

Observation 1. Taking the income/income tax data from 1961 to 2001 and the corresponding consumer price index (CPI) we can adjust all incomes to the 2001 level and compare them. The result is shown in the graph below. Indeed the average income per capita before and after taxes was highest in the 70' and the 80's and has been sharply declining since.

&amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;nbsp;

So it seems that there is really a downward trend and people actually are less well of than a few decades ago in terms of buying power as measured by the CPI (at least).

Observation 2. As reported in (Spiegel Online), experts fear that Germany will face a fatal brain-drain in the future. Unfortunately the available data only covers the years up to 2006 therefore the most recent developments cannot be analyzed. Resorting to data from 2006, we have 639,064 people having left Germany and 661,855 people having moved to Germany in 2006 according to the German Bureau of Statistics. So on average the number of people in the country does not change significantly due to the fact that people leave and move to Germany. But this is only one half of the story: In terms of "qualifications" and "tax money generated" these two streams across the borders of Germany might differ greatly. First, Germany is losing more highly qualified people than it is 'importing' from the outside. Second, whereas people who come to Germany in the pursuit of a better future might earn less in the beginning, people who leave Germany are often well-trained and high potentials (as claimed by the article) and might already belong to the 'better-earning' part of the society or have the better-earning potential. Therefore, in terms of tax-money, 'losing' these people might have an even worse impact on the tax money paid in the country - which hypothesis we will investigate now. The following table provides a income/tax split for the year 2001. In the last column we find the 'loss' associated with a person leaving a corresponding tax bracket (or tax bucket), expressed in terms of a number of 'average' taxpayers leaving the corresponding tax bucket. The 'average' taxpayer is defined to pay the 'average' tax, i.e. total of paid income taxes divided by the number of total taxpayers. For example, if you earn between EUR 100,000 and EUR 175,000 per year and you leave Germany, in terms of average taxpayers it is a loss of 5.53 people.

&amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;nbsp;

Using this metric rather than the actual number of people leaving the country we can obtain a more appropriate picture about the situation at hand. It allows us to assess the impact of the current developments on the overall income tax. A deficit in income taxes has to be compensated in the mid-term and might call for an increase in taxes, which is exactly what is demanded by various groups right now. This in turn may trigger second-round effects which worsen the situation. For a sound analysis we need to know what the income distribution of the streams of people leaving and moving into the country looks like. Unfortunately, I could not find any specific data on the topic. Therefore I decided to include the following small spreadsheet that allows us to play out a few scenarios and experiment with different distributions. One can use this spreadsheet to compare the effects on the overall income tax in the "change in %" field of the spreadsheet. The following assumptions were made: (a) the income bucket split of 2001 is a rather good (and conservative) estimate of today's situation; (b) The number of people leaving and moving into the country is taken from the numbers for 2006.

I have played out a few scenarios and filled different income distributions. One can create more scenarios by entering numbers in the bold+italic fields.

&amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;nbsp;

Even without knowing the exact current distributions we can observe the effect of different situations on the overall income tax and draw a few conclusions. (Any data on the topic is welcome.)

So what we observe is a certain interdependence between income tax and outflow of people across the German border: A loss of 'income tax' income for the country may lead to higher income tax rates in the following years to compensate for the loss. In turn, a higher income tax to be paid may encourage more people to leave Germany. The interplay of these two factors is nothing else but a classical vicious circle. Therefore, in Germany, we might not only be facing a loss of high potentials but also, inherently connected, a loss of tax income and vice versa. We have a 'multi drain'.

The effect we would expect is that the gap between the average yearly income per capita before taxes and after taxes is decreasing when more 'expensive' taxpayers leave Germany (assuming tax policy remains unchanged). In this case the income tax rate that is paid on average by the remaining people decreases and so does the gap. On the other hand, changes in the taxation of people would become visible as well. I would like to see if these patterns are observable in more recent data. So if someone has access (or sources) for the corresponding data please let me know.

The currently observed overall dynamic supports one point: Countries (as businesses) are providers of accommodation for a certain fee (the taxes) and the customer (the citizen of the country) has to decide if the offered deal is favorable for him with respect to his personal preference and utility function. So the same principles that apply for companies might apply for the countries as well and ultimately they may have to pay more attention to retain citizens which generate the higher tax income. Free from any political views, the underlying dynamic might demand this. In this sense it is nothing else but 'revenue management for countries'.