Phillips Macroeconomic Model
roypd...@my-deja.com
information about economic models) about a hydraulic model created by
the man who invented the Phillips Curve relating unemployment to wage
increases. The article in the New Scientist, 9 December 2000, pages
46-47 tells the story of how Bill Phillips, a New Zealander who had
survived three and a half years in a Japanese prisoner of war camp, came
to Britain in 1946 to study sociology at the London School of Economics.
However he barely scraped a pass degree at the LSE, partly because he
kept dashing out of the exam room to smoke cigarettes, and partly
because a growing interest in economics had caused him to neglect
sociology.
Phillips sent a note to one of his economics teachers saying that as he
had difficulty in understanding his lectures he had constructed a
hydraulic model of the economy. The lecturer suggested he show it to
James Meade who encouraged him to develop it and in November 1949
Phillips unveiled it to an audience at the LSE. Coloured water
representing cash was poured in at the top and then split into different
loops to demonstrate its course through the economy. Inflation proved a
problem as if the rate rose too high water squirted out through a hole
and left a puddle on the floor!
Phillips built 14 improved models some of which ended up in British
universities, others at Harvard, the Ford Motor Company and the Bank of
Guatemala. The New Scientist article says that one of the machines is
now in the Science Museum in London. It doesn't mention whether any of
the others survive.
Digital computers have made analogue machines like Phillips' obsolete
which in some ways is a pity. If we can't understand something we often
say we cannot "see" it. The hydraulic model helped Phillips himself to
visualise the workings of the economy in a way that a mathematical model
would not have. (It also helped him to get a job at the LSE). However I
suppose that with computer graphics it would be possible to display
stocks and flows in a similar way.
Outside the mainstream of economics John Turmel, an engineer who
frequently posts messages to various newsgroups on what he regards as
the iniquities of charging interest on loans, has outlined a hydraulic
type analogy to support his arguments but I don't know if he has ever
tried to build a physical model like Phillips did.
Roy Davies
History of Money from Ancient Times to the Present Day
http://www.ex.ac.uk/~RDavies/arian/llyfr.html
Sent via Deja.com
http://www.deja.com/
David Lloyd-Jones
<roypd...@my-deja.com> wrote
<Much interesting stuff snipped.>
> Digital computers have made analogue machines like Phillips' obsolete
> which in some ways is a pity.
I don't think it's clear that analogue computers have been made obsolete by
the post-Turing (or the Post-Turing) digital revolution. Analogue computers
of various kinds are sometimes more powerful, or at least more able. The
Travelling Salesman Problem, e.g., is difficult to solve digitally, and very
rapidly blows up. The soap-bubble computer, a simple analogue device, solves
it instantly.
Similarly, the problem of arranging numbers, or anything else, in order is
O(something horrible), i.e. it blows up. The spaghetti computer, admittedly
a hypothetical device, solves it instantly: you just snip a lot of pieces of
spaghetti to lengths proportionate to the numbers to be ordered, and then
straighten up the whole bundle on a table. The highest number will instantly
stick out farthest, the second one next...
These two examples are almost jokes, but serious problems like protein
folding are also on the table. Here in Toronto IBM are spending some
astronomical amount of money over the next few years (I think it may be $400
million in five, anyway that order of magnitude) to build a machine which
might, just might, get a handle on protein folding. Meanwhile the folks
downtown at Sick Kids, one of the continent's best research centres mainly
because it's full of extraordinary Chinese, Cambodian, and other Asian
scientists, are playing with struts and elastic bands. Analogue computers
again.
IBM or Sick Kids? Six will get you five either way. But I think it is a
pretty fair bet that Sick Kids will be bringing down a Nobel in Medicine and
Physiology every year or so through the twenty-teens.
-dlj.
Paul Walker
[cut]
>Phillips built 14 improved models some of which ended up in British
>universities, others at Harvard, the Ford Motor Company and the Bank of
>Guatemala. The New Scientist article says that one of the machines is
>now in the Science Museum in London. It doesn't mention whether any of
>the others survive.
Yes there is one at the New Zealand Institute of Economic Research in
Wellington New Zealand.
--
____________________________________________________
Paul Walker p.wa...@econ.canterbury.ac.nz
roypd...@my-deja.com
It is interesting to know that analogue computers still have some
advantages in research. I was half expecting that someone would try and
defend digital computers by pointing out that unlike Phillips' machine
they would not make the floor wet when inflation got out of control!
However I am imagine that people who saw Phillips' machine spilling
water would remember it more vividly than those who had simply run some
software package!
Roy
http://www.ex.ac.uk/~RDavies/arian/money.html
Xavier Rousseau
>
>
> Phillips sent a note to one of his economics teachers saying that as he
> had difficulty in understanding his lectures he had constructed a
> hydraulic model of the economy. The lecturer suggested he show it to
> James Meade who encouraged him to develop it and in November 1949
> Phillips unveiled it to an audience at the LSE. Coloured water
> representing cash was poured in at the top and then split into different
> loops to demonstrate its course through the economy. Inflation proved a
> problem as if the rate rose too high water squirted out through a hole
> and left a puddle on the floor!
It's not the first time I hear that hydraulic models can prove theirselves
useful in economics thinking. Actually I find all this very interesting.
Years ago I have been working with Hurst exponent, long memory and arfima
models. That's a technique which was originally developped for modelling
risings of the Nile. Technique which is now presently used for dealing with
non linear time series that exhibits stationnary but very low declining acf
coefficients.
Thank you for your contribution. You tell me something I've never heard
before.
Can you give us more information about Philips model and tell us where the
paper can be found.
Ron Peterson
> I don't think it's clear that analogue computers have been made obsolete by
> the post-Turing (or the Post-Turing) digital revolution. Analogue computers
> of various kinds are sometimes more powerful, or at least more able. The
> Travelling Salesman Problem, e.g., is difficult to solve digitally, and very
> rapidly blows up. The soap-bubble computer, a simple analogue device, solves
> it instantly.
Digital electronics can emulate analog electronic devices and have faster
performance and greater accuracy.
I am not familiar with the soap-bubble computer, but although the
traveling salesman doesn't have an efficient exact solution there are
approximations that are efficient.
> Similarly, the problem of arranging numbers, or anything else, in order is
> O(something horrible), i.e. it blows up. The spaghetti computer, admittedly
> a hypothetical device, solves it instantly: you just snip a lot of pieces of
> spaghetti to lengths proportionate to the numbers to be ordered, and then
> straighten up the whole bundle on a table. The highest number will instantly
> stick out farthest, the second one next...
Sorting using a comparison sort takes O(log n) where n is the number of
items so 4 billion items would only take 32 passes.
Ron
GaryForbis
news:3a5e257a$0$63730$272e...@news.execpc.com...
>
> Digital electronics can emulate analog electronic devices and have faster
> performance and greater accuracy.
At least in the eighties Carver Meade thought differently. He beileved we
could have several orders of magnitude speed ups by working in the linear
area of the transister's behavior.
David Lloyd-Jones
"Ron Peterson" <ro...@earth.execpc.com> wrote
>
> Digital electronics can emulate analog electronic devices and have faster
> performance and greater accuracy.
>
Digital electronics can emulate some analog devices approximately, and have
faster performance and greater accuracy in problems of integer math.
-dlj.