Oct 29, 2011

Why banks have become dysfunctional

James Saft, the Reuters' columnist, has a piece in the IHT about a talk by Andrew Haldane, the executive director for financial stability of the Bank of England. Once in a while, somebody writes a few lines we ("we") really should read, and here they are:

The purchaser of a portfolio of global banking stocks in the early 1990s is today sitting on a real loss. So who exactly is it extracting value from today's banks? The answer is twofold: shorter-term investors and bank management. Because banks have, over the past two centuries, migrated to a limited liability, shareholder-owned model, there is a natural tendency for owners to make riskier loans and trades and to increase the bank's assets.

Andrew Haldane
A bigger, riskier balance sheet with more leverage produces terribly volatile results, with many good-size profits mixed in with the occasional catastrophic loss. But with limited liability, executives and shareholders can simply walk away from the smoking wreckage, having pocketed the gains when times were good.

Bank of England
Banks then have a built-in incentive always to increase leverage, and the tyranny of quarterly earnings places huge pressure on them to enlarge their asset books, even if there is no one creditworthy left to lend to.
That was one of the main causes of the subprime episode. Faced with the prospect of not increasing earnings, banks simply began to manufacture borrowers where none really should have existed.
The situation is exacerbated by the fact that debt is tax-deductible while equity is not, giving banks even more incentive to borrow. While the typical leverage of an American or British bank in 1900 was five or six times equity, that figure peaked at about 30 times before the crisis, and is higher still now for many euro zone banks.
Bank bondholders have been unwilling to play their role as vigilantes, in part because they quite rightly expect to be bailed out by governments if banks go to the wall.
In the past 30 years, many banks have moved to measure their performance -and set their bonuses -on the basis of a measure called return on equity, which measures profit compared with equity. What return on equity does not adjust for, of course, is risk, and it looks as if return-on-equity targets in a leverage driven business have produced a lot of risk in the form of extreme bank earnings volatility, and badly compensated volatility at that.

PS: Conspiracy, conspiracy. If you search for James Saft on IHT's web site, it comes back with " 'James Saft' did not match any documents under Past 30 Days." If you search for the column's title "Why banks take such huge risks," it comes back with all sorts of articles (about Berlusconi, among others), but not with Saft's column. However, if your search for the same title on Google, it comes back with a mirror site of the column as first result. (Lol)

Oct 22, 2011

We missed rapture day

Rapture had been scheduled for May 21, and was then rescheduled for Oct 21 (an error in the calculations). Yesterday, folks. And we missed it.

 

An alternative explanations is, of course, that we all got raptured yesterday (or at least everybody we know), and are now in heaven. You decide.

Oct 19, 2011

History of the world: Apple Computers

Act II. Somewhere in 1978 or 79, the Amsterdam department store De Bijenkorff opened a new sales corner on its 4th floor, mysteriously named "huiscomputers" and it featured a new product, the Apple II home computer. At that time most people, including myself, would conceive of computers as "electronic brains" (Germans called them "Elektronengehirne," before they called them "computers," before they called them "Rechner,") all built by IBM, all infinitely expensive, large, and remote.

Act I. My first contact with computer had been in 1972, when I took an algebra class at the Free University of Berlin and we were tasked to program matrix inversions and some such in Algol68, the programming language du jour. This was done by (1) punching Hollerith cards in the right places, on special machines located in the university's computing center, then (2) placing the cards in the intray located in the hallway outside the main operating room where the computer was located (there was only one), (3) waiting for an operator to appear to empty the intray (he would open a wing door, and allow you a glimpse at the electronic brain, humming and chugging along in fluorescent light, tape decks clicking back and forth), (4) then waiting another hour or so for the operator to reappear with the "output," --- folded stacks of paper in a very large format, the name of the "job" (no pun intended) printed in very large letters on the first page. If your stack was very thin (as it usually was) this could mean only one thing: something had gone wrong. You would (5) try to find the error, or try to find some help to find the error, correct it, (6) resubmit your job, and repeat the process ad infinitum. Usually, it would take only a few days  until a program of a few lines code would finally run properly.

Act II, cont'd. So far so good. Back to the department store. What could you do with a home computer, I asked the sales person. Well, he said, you could store cooking recipes and call them up as appropriate. I didn't buy one.

Act III. We're now at Dartmouth College, NH, and the day is Jan 16, 1984. In  between, I had become interested in a computer simulations,  and was visiting there Dartmouth's Research Policy Center, run by Dennis Meadows of The Limits to Growth fame (the book) to learn more about his approach, called "System Dynamics." To repeat, the day is Jan 16, a Monday, and we all must go and have a look at the new Apple computer, the Macintosh. So we cross the icy, snowy campus, and arrive in a dedicated room of the computing center, where a passionate lady demonstrates to us what a rectangular box, white, with a small screen, and a funny device, called "mouse," linked to the computer could achieve together. There was also a small matrix printer with ugly output. But, but, you could create sketches on the Macintosh screen by moving the mouse across the table, and then print them on the printer. Also, you could use different fonts, when printing a text. This led to typographic orgies of the worst kind for months on end (don't ask).


Act IV.  A year and a half later. I'm returning to Dartmouth College on a regular basis for various projects, and spend a lot of time with Perry LaPotin, the polymath grad student, who has become an invaluable part of the Cold Region Research Lab of the Corps of Engineers, conveniently located next to the college. Perry was already writing programs for the Macintosh. There was only one small problem. You could not write Macintosh programs on the Macintosh, since its memory was too small. Apple had built another machine, the Lisa, sold only to professionals, whose memory was large enough, since it had a hard disk (HARD DISK). The hard disk was really large, 10 megabytes, but the was a little glitch in the hard disk space management. Lisa didn't always know when the hard disk capacity was exhausted, which led to hard disk malfunction, which then Perry had to repair by using a mix of erratic reset activity (the escape button, yes), brawn, and black arts. He spent roughly half of his working day resetting the hard disk. I still see him sitting there, patiently kicking our Lisa back to work. When we would finally go home, belatedly, exhausted, we turned our attention to the regrettable downward tendency of the Apple Computer stock price. Apple was already on its way out, since the Macintosh was fairly useless.

Act V. Now comes the part that is omitted in all the obituaries. A few weeks later, still 1985. The Apple laser printer appears on the market. And it prints like a professional printer, plug and play, 50 different fonts, some very convincing ones. Your manuscript looks just great, your letter looks just great, your writ, opinion, table of content, graphics (Graphics) they all look great. It looks almost as if you can stop arguing, just putting your breath-taking graphics of the front page of your important contribution (the PC-world of MS-DOS, might, just might be able to connect to some laser printer and print something in Courier font until the next software glitch puts an end to such pretentiousness) but we, with our Apple laser printer, we rool (we meant "rule," but rool is even better) we rule the world. My research grant applications are looking so much better than those of the competition, I'm collecting one grant after the other, until I get a Pioneer Grant from the Dutch goverment that allows me to start my own research institute, the Applied Logic Laboratory. I'm still convinced that my success in those years hinged on the flawless Macintosh laser print of my submission, and in particular on the flawless laserprinted  tables of content done by the best text processor of those days, Wordperfect. For example, the committee for the Pioneer grant met only once, with forty longish application to evaluate, and only one grant to award. You an bet that they started reading the stuff when stepping on the train for their meeting in The Hague (much Dutch work gets done on trains, as Paul Krugman), and they had barely time to read the tables of content. Mine was the best.

Anyhow, the laser printer constituted a quantum leap, and many people understood, got their Macintosh laser act together, bought the stuff. and saved the company.

Stay tuned.


Act VI.  It's three years later 1988, and I'm back in Amsterdam. The Macintosh II appears on the market, the first bona fide machine with a color screen. Somebody wrote a program that would generate Mandelbrot's fractals in real time and the annual Dutch software exhibition features nothing but magic lanterns that move according to the incorruptible logic of Mandelbrot's algorithm.

A typical Mandelbrot image (Helix 2)

Interlude (short). We're also getting a research contract with IBM, since IBM has now a Unix machine, a mini computer, half way between a PC and a small mainframe (for insiders only: think VAX). We would get the computer for free (listprice perhaps 100 kay, regardless of the currency), work with it, and produce a report. Nerd alert: IBM has a Unix machine. Not Unix of course (nor Linux, which didn't exist in those days), but some Unix dialect that is supposedly compatible with standards Unix (of course it isn't). A small step for mankind, but a big step for IBM.

Now, in order to use the machine, we had to connect it to our network. And it's an ethernet-(work). "Is your new machine ethernet-compatible," we ask IBM. "We are the best," the man in the blue suits sing in unison while pummeling their breasts (there was a dress code at IBM in those days, blue suits, white shirt, tie), "so our machine is ethernet compatible."

So we connect the IBM machine (something with lot's of "8" in the name) to our Ethernet. Nothing happens, of course. We call IBM. "It's your fault," the men in the blue suits sing over the telephone while pummeling their breast, have you though of switching your network on?" This goes back and forth for a few month. "Have you thought of this, have you thought of that?" Yes, we have. One fine day, A delegation from IBM descends from heaven in the spaceships that Emmerich's Independence Day made so famous. Several people. They switch on the machine, they think of this and of that, but nothing happens. This takes the whole day. Finally, finally, they have the answer. "Yes, they say, it's obvious, you are using the latest Ethernet version, and our machine is not yet compatible with your version. It's your fault."

I'm not making this up.



Act VII. My research center (initially cursed with the hopeless name CCSOM) is growing, and we need more computers

Oct 15, 2011

History of the world: Apple Computers (5)

(Go  here for earlier acts)

Act V. Now comes the part that is omitted in all the obituaries. A few weeks later, still 1985. The Apple laser printer appears on the market. And it prints like a professional printer, plug and play, 50 different fonts, some very convincing ones. Your manuscript looks just great, your letters look just great, your writs, opinions, protestations, tables of content, graphics (Graphics), indexes, they all look great. You look great. A picture values a thousand words, a laser-printed graphic is invaluable; (in the PC-world of MS-DOS of 1985, you might, just might have been able to connect to some third party laser printer and print something in Courier font until the next software glitch put an end to your pretentiousness, but graphics where an entirely different animal and would have had to be printed separately anyhow).

My research grant applications are looking so much better than those of the competition, I'm collecting one grant after the other, until I get a Pioneer Grant from the Dutch government that allows me to start my own research institute, the Applied Logic Laboratory. I'm still convinced that my success in those years hinged on the flawless Macintosh laser print of my submissions, and in particular on the flawless laser-printed  tables of content. For example, the committee for the Pioneer grant met only once, with forty longish applications to evaluate, and only one grant to award. You can bet that the committee members, all busy, distinguished scholars, didn't start reading the stuff until they stepped on the train for their meeting in The Hague (much Dutch work gets done on trains, ask Paul Krugman), and they had barely time to read the tables of content during the journey. Mine was the best.

First Apple laser printer (plug & play)

Anyhow, the laser printer constituted a quantum leap, and many people understood, got their Macintosh laser act together, bought it together with the Macintosh, and saved the company.

Go here for the next act.

Oct 13, 2011

Oct 12, 2011

History of the world: Apple Computers (3)

(Go  here for earlier acts)

Interlude. A friend sends this picture and writes...

Apple store in Palo Alto, CA
Apple Store, Palo Alto, CA.

..."did you know I hate Post-It stickers, and the people who use them, almost as much as I hate Apple?" 

Go here for the next act.

Oct 10, 2011

History of the world: Apple Computers (2)

(Go  here for earlier acts)


Act III. We're now at Dartmouth College, NH, and the day is Jan 16, 1984. I had become interested in computer simulations,  and was visiting Dartmouth's Research Policy Center, run by Dennis Meadows of The Limits to Growth fame, to learn more about his approach, called "System Dynamics." To repeat, the day is Jan 16, a Monday, and we all must go and have a look at the new Apple computer, the Macintosh. So we cross the icy, snowy campus, and arrive in a dedicated room of the computing center, where a passionate lady demonstrates to us what a rectangular box, white, with a small screen, and a funny little device on the desktop, called "mouse," could achieve together. There is also a small matrix printer with very ugly output. But, but, you could create sketches on the Macintosh screen by moving the mouse across the table, and then print them on the printer. Also, you could use different fonts for your text, and print them as they appeared on the screen (WYSIWYG). This led to typographic orgies of the worst kind for months on end, campuswide (don't ask), printed in very ugly ways by this matrix printer.

Apple Macintosh

Act IV.  A year and a half later. I'm returning to Dartmouth College on a regular basis for various projects, and spend a lot of time with Perry LaPotin, the polymath grad student, who has become an invaluable part of the Cold Regions Research Lab of the Corps of Engineers, conveniently located next to the college. Perry was already writing programs for the Macintosh. There was only one small problem. You could not write Macintosh programs on the Macintosh itself, its memory was too small. Apple had built another machine, the Lisa, available only to professionals, whose memory was large enough for Macintosh programming since it had a hard disk (HARD DISK) that could be made to work as virtual memory. The hard disk was really large, 10 megabytes, (MEGABYTES) but there were glitches. Lisa didn't always know when the hard disk's capacity was exhausted, which led to hard disk malfunction, which then Perry had to repair using a mix of erratic reset activities (eg. the escape button), brawn, and black arts. He spent roughly half of his working day resetting the hard disk. I still see him sitting there, patiently kicking the Lisa back to work. When we would finally go home, belatedly, exhausted, we would turn our attention to the regrettable downward spiral that constituted the Apple Computer stock price. Apple was on its way out, since the Macintosh was fairly useless.

Perry LaPotin


Go here for the next act.

Oct 8, 2011

History of the world: Apple Computers (I)


Act II. Somewhere in 1978 or 79, the Amsterdam department store De Bijenkorff opened a new sales corner on its 4th floor, mysteriously named "huiscomputers," which featured a new product, the Apple II home computer. At that time most people, including myself, would conceive of computers as "electronic brains" (Germans called them "Elektronengehirne" before they called them "computers" before they called them "Rechner"), all built by IBM, all infinitely expensive, large, and remote.

Standard IBM Hollerith punch card

Act I. My first contact with computers had been in 1972, when I took an algebra class at the Free University of Berlin and was tasked to program matrix inversions and some such in Algol68, the programming language du jour. This was done by (1) punching Hollerith cards in the right places, on special machines located in the university's computing center, then (2) placing the cards in the intray located in the hallway outside the main operating room where the computer was located (there was only one computer), (3) waiting for an operator to appear to empty the intray (he would open a wing door, and allow you a glimpse at the electronic brain, humming and chugging along in fluorescent light, tape decks clicking back and forth), (4) then waiting another hour or so for the operator to reappear with the "output," --- folded stacks of paper in a very large format, the name of the "job" (no pun intended) printed in very large letters on the first page. If your stack was very thin (as it usually was) this could mean only one thing: something had gone wrong. You would (5) try to find the error, or try to find some help to find the error, (6) correct it, (7) resubmit your job, and repeat the correction loop as appropriate. Usually, it would take only a few days  until a program of a few lines of code would finally run properly. 

IBM mainframe, system 360 (1964 - 78)


Act II, cont'd.  So far so good. Back to the department store. What could you do with a home computer, I asked the sales person. Well, he said, you could store cooking recipes and call them up when needed. I didn't buy the Apple II.

Go here for the next act.