This is an archive of past discussions with User:Nbarth. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.
One wants to be very cautious about attributing the origin of notions of utility functions and of marginal utility to the work of Cramer or of Bernoulli on the St Petersburg Paradox. On the one hand, some read Aristoteles as having a concept of marginal utility; on the other hand, many later thinkers arrived at such notions independently of any concern with this particular paradox. —SlamDiego←T18:01, 1 January 2009 (UTC)
BTW, quite a while back, I was in repeated conflict with an editor at that article who wanted to replace every reference to diminishing marginal utility with a reference to risk aversion. Perhaps you already see the problem there in full, but part of it is that Cramer and Bernoulli specifically conceptualized things in terms of diminishing marginal utility, and part of it is that, while diminishing marginal utility can be expected to imply risk aversion, it is really only equivalent to risk aversion on the assumption of something like the strong independence axiom (id est, that expected utility is linear in the probabilities). (Some technically define “risk aversion” to be equivalent to diminishing marginal utility, but this turns what was a description into a mere name, dangerously separated from ordinary language.) —SlamDiego←T06:27, 2 January 2009 (UTC)
This adjusts for scale, yielding a dimensionless quantity which can be compared between different sample sets
That seems horribly misleading, because it seems likely to leave the reader under the impression that comparability across data sets, resulting from dimensionlessness, is the point.
That is wrong.
In typical linear regression problems, even when all of the errors have the same variance, the variances of residuals vary greatly. That is the main point! Think of a very simple regression problem: just a scatterplot in the plane, to which you fit a line. If an x-value is extreme, then the corresponding residual is highly sensitive to the estimated slope, and thus has a large variance; by contrast, for a nearly average x-value, the residual is relatively insensitive to the estimated slope. Errors, in contrast to residuals, on the other hand, in no way depend on the estimated slope. Thus when errors have the same variance, residuals don't. Michael Hardy (talk) 00:54, 21 February 2009 (UTC)
Now I see what I had forgotten: in the article it said:
The fact that the variances of the residuals differ, even though the variances of the true errors are all equal to each other, is the principal reason for the need for studentization.
I think you significantly misrepresented the cited source on ringing and windowing. Please take another look, and if you need help interpreting it, ask. Dicklyon (talk) 02:52, 29 March 2009 (UTC)
Simple truncation causes severe ringing artifacts. …
To reduce artifacts originating from truncation the ideal reconstruction filters can be multiplied with appropriate functions which drop off more smoothly at the edges.
We will discuss…these…windows.
I don’t see how my wording:
In image processing applications, windowing often causes ringing artifacts, and different choices of windowing function can significantly affect a filter.
significantly differs from it.
Should I write instead: “Simple truncation causes ringing artifacts, which can be mitigated by different choices of windowing functions”?
I'd say the more sensible interpretation is that windowing reduces ringing artifacts. An unwindowed sinc, or a truncated sinc, has ringing artifacts; a tight window that cuts off negative lobes will have no ringing at all. The windowing doesn't cause ringing – it reduces it. Unfortunately that particular source doesn't say what happens with an unwindowed filter. And it is not clear why you put the mention of ringing into the context of images. Dicklyon (talk) 03:39, 29 March 2009 (UTC)
Good point about truncation being the main cause, and windowing reducing this – that’s a better way of putting it, and I’ve incorporated it.
I’ve rewritten (in rather more detail) the contributions to the various pages, drawing on the sources and adding another reference for Lanczos. Do they look better?
Truncation isn't the main cause, either, as it never adds negative lobes that weren't there already. Ringing is caused by filter impulse responses that ring, corresponging to sharp frequency-domain cutoffs; windowing reduces the negative lobes and smears out the f-domain transitions, reducing ringing. But I wouldn't focus too much on ringing in the context of windowing; the frequency-domain view of windowing is usually more important. I haven't looked at your new contribs yet, and probably don't have time right now, but later. You can reply here without the talkback thing and I'll see it. Dicklyon (talk) 16:03, 29 March 2009 (UTC)
Ok – thanks for clarifying.
So if I understand what you’re saying, brick-wall filters cause ringing, due to the discontinuity (sharp cutoff) in the frequency domain.
Thus sinc, the time-domain view of a brick-wall low-pass filter, causes ringing, even ideally?
…and this is reflected (in the time-domain) in the negative lobes?
Are there any references you might recommend?
The authors of the paper are either confused or unclear though, as they attribute ringing artifacts to truncation, as in the quote – they don’t say that truncation fails to fix the artifact.
From doing a little reading, it seems people attribute ringing to the Gibbs phenomenon, i.e., discontinuities in the signal.
Regardless, ringing is not the emphasis of the revised changes.
Right, that source is very unclear on that point; it should say that the rectangular window does little to reduce ringing (and has other bad properties when viewed in the frequency domain). Also, if you're going to reference that paper, you need to say where it's published; a link to a PDF that doesn't say really doesn't quality as a reliable source. As for Gibbs, yes, that's pretty much what a brick-wall filter is. Here are some books you look at. Dicklyon (talk) 22:13, 29 March 2009 (UTC)
I took out all that stuff you added to Lanczos filter. From the figures in the paper it appears that the authors just got very confused. Their plots do not show it as one of the ones with negative frequency-domain lobes. I think you should read some of those books and base your edits on better sources. Dicklyon (talk) 22:56, 29 March 2009 (UTC)
Actually, I mis-characterized the problem. Looking with more zoom, I do see a slight negative lobe on the Lanczos plot. But the conclusions that it causes "severe artifacts" and that it's "unsuitable for reconstruction" are not justified. If there's something about the 3D "reconstruction" problem that makes it a lot different from image reconstruction, they haven't explained that. And the figures at the end suggest that they didn't bother trying the Lanczos filter. If you want to cite the published opinion of these authors anyway, you need to at least show that they are credible experts first. Dicklyon (talk) 23:17, 29 March 2009 (UTC)
This is an automated message from CorenSearchBot. I have performed a web search with the contents of Ringing (signal), and it appears to be very similar to another Wikipedia page: Ringing. It is possible that you have accidentally duplicated contents, or made an error while creating the page— you might want to look at the pages and see if that is the case. If you are intentionally moving or duplicating content, please be sure you have followed the procedure at Wikipedia:Splitting by acknowledging the duplication of material in edit summary to preserve attribution history.
This message was placed automatically, and it is possible that the bot is confused and found similarity where none actually exists. If that is the case, you can remove the tag from the article and it would be appreciated if you could drop a note on the maintainer's talk page. 209.51.196.26 (talk) 23:15, 12 April 2009 (UTC)
Invitation
I would like to commend you for your substantial contributions to articles in the telecommunications and signal processing area. I would also like to invite you to Wikiproject Telecommunicaitons WP:TEL. Mange01 (talk) 20:32, 14 April 2009 (UTC)
Thanks! Glad to be helpful – I’ve added myself as part-time.
Nils, I removed again the bit about Babylonians' sums and differences being Fourier transforms – it's too much of a stretch to say such a thing without an explicit source; the wikipedia page you linked was itself unsourced, and removed it from there, too. I'm worried too about the Lagrange resolvents; I can't find anything in the books you cited that connect the resolvents with an order-3 DFT. Can you tell me what pages to check? Dicklyon (talk) 02:56, 21 April 2009 (UTC)
Hi Dick,
The Lagrange resolvents for the cubic are (obviously) the order-3 DFT, though people don’t tend to refer to them as such (separate culture of algebraists and analysts/signal processing).
I don’t have a reference to hand that explicitly links them, so you’re welcome to remove them if you feel uncomfortable, though mathematically they are identical.
OTOH, digging around, I’ve actually found several references that explain how the Babylonians used (continuous) Fourier series in astronomy – I’ll add those (referenced); also mentioned earlier 18th century Fourier theory.
We can't be using our own observation of similar function form to make such claims about the order-3 DFT; if it's not sourced, let's leave it out. I've looked at the some of those refs that mention Babylonian and Fourier and Neugebauer, but I don't find any explicit support for the statement in the article that the Babylonians used a form of Fourier analysis. Can you point to anything that actually says so? Dicklyon (talk) 03:39, 21 April 2009 (UTC)
Re: order-3 DFT – ok, feel free to take it out.
Re: statement that the Babylonians used Fourier analysis.
I’ve added a reference for resolvents being Fourier theory, and reworded it to accord with what the source says (Knapp calls it “a Fourier decomposition relative to a cyclic group”, though he doesn’t call it a DFT).
Digging more, I’ve found a source that discusses (with facsimile copy of Lagrange’s paper!) a further use of the DFT by Lagrange (“The DFT” book), which I’ve included with reference.
Well, the rest look pretty marginal, but that one looks good. Thanks for working on it. I added some missing titles. Dicklyon (talk) 04:25, 21 April 2009 (UTC)
No worries, and thanks for helping – it looks rather better than before.
AFAICT, there’s a conflict between “Common naming” versus “Neutral naming” (Wikipedia:Neutral point of view#Article naming), which I was trying to resolve: while the European-only names are common in English, the Persian-European names are used in the literature, and preferred in the historical literature, as discussed in the references. Hence mentioning both, particularly in light of the centuries of historical precedence and significant and earlier contribution, and use in the historical literature, seems a more neutral position. (For the purpose of linking it’s of course moot.)
Nils, your recent edits suggest that PTP and DCF are somehow linked, but I can find no evidence for that. Do you have some sources that indicate some connection between them? The fact that most cameras implement both might be worth mentioning, if we have a source that says so, but as far as I know it doesn't go beyond that. Dicklyon (talk) 18:23, 3 May 2009 (UTC)
Hi Dick,
I figured that if one were interested in how images are stored on a camera or transfered between them, one would likely wish to know about both PTP and DCF, so I linked them.
Sorry if my edit suggests that they are linked more strongly than that; I’ve elaborated at Picture Transfer Protocol, with ref (PTP was designed to work with DCF, Exif, etc., though it’s a different layer) – how d’you like it now?
Great addition. Just as the latest G8 was kicking off in Italy. There was a slight issues with the number of pics which I explained on the articles talk page. FeydHuxtable (talk) 12:34, 9 July 2009 (UTC)
I'm contacting you as you have contributed to Paraboloid, and I hope you don't mind me asking you a purely hypothetical question about maths and the wooden hyperbolic paraboloid roof on the Church Army Chapel, Blackheath. We know the roof is a wooden structure as it was thus described by Terry Peck of Capital Roofing, who re-surfaced it last year. It was important to know this as, if I remember rightly, they were often made of carbon fibre and had a tendency to blow off the roof if a gale occurred during transport to site or fixing. Also, carbon fibre roofs don't support spires. So the idea of a wooden version makes sense. But how was it done?
My question is - is it possible to construct a hyperbolic paraboloid using straight lines (e.g. wooden beams) only? Forgive me - I have no mathematical knowledge - but the drawings on the Paraboloid page have lines delineating curves only, so any straight-line element within the shape can't be seen in the illustrations.
I am also interested to know whether - if the roof could have been constructed entirely with straight beams - the roof would be rigid enough to support a fairly lightweight but tall aluminium spire? I suspect that most of the required strength and rigidity would be for keeping the spire steady in a gale, and not quite so much for supporting weight.
I am also wondering if such a structure would be rigid by the same principle as the Mathematical Bridge. I don't understand the rigidity-principle of the bridge myself, but I do appreciate that if there is a named principle which could also be applied to a wooden hyperbolic paraboloid roof structure, then even a dumbo like me would have simple language by which to demonstrate that the roof could theoretically be rigid enough to stabilise the spire.
Please forgive me for all these questions - I appreciate that you may not be interested in the question, and I quite understand if you don't feel like answering. My problem is that the plans are missing and I am not an architect, so I have to look for other means of understanding the building. I shall be attempting to visit it next week, so any advice on what too look for or photograph would be gratefully received.
Hi Storye book – no worries re: question (fair game)!
To your first question – yes, a hyperbolic paraboloid can be built using only straight beams, a so-called saddle roof – that’s the practical meaning of being a doubly ruled surface. I’ve added the attached image to Paraboloid: Properties so one needn’t dig through 2 or 3 levels of links to learn this.
Regarding rigidity – I’m no engineer, but I believe that saddle roofs are rigid (i.e., like triangles, which don’t flex, unlike squares, which do), and I believe that you are correct that the key issue with light-weight spires is keeping them steady in wind, not supporting them.
In the Mathematical Bridge, the straight lines (tangent to the bridge surface) lie outside the curved bridge surface.The Mathematical Bridge is built by an entirely different principle – it has a weak curved bridge surface, supported by external trussing.
The difference can be seen as follows: in a saddle roof, the straight beams are part of the curved surface (tangent lines to the surface lies inside the surface, in the 2 directions that form the rulings), while in the Mathematical Bridge, the straight lines are not part of the curved surface, but rather project out and form the trussing.
Hope these replies help!
You may find the Wikipedia:Reference desk most helpful in answering questions, since it seems to be peopled by friendly folk for just this purpose!
And regarding photographs of Church Army Chapel, perhaps a photo of the roof from some angle and height that shows its shape to good effect (shadows help, of course) to contribute to the roof section of the article, together with an overall photograph of the chapel, and of any notable details, since there are no photos of the church at the current article.
I've updated Church Army Chapel, Blackheath with your information. Of course you are welcome to edit it if you see fit. I worry that my new para could get deleted on the grounds of being opinion, however the architect's own drawing at the top does appear to show straight beams on the roof, and the beams go in the same direction as the ones in your image. Meanwhile I have just been given a new lead towards a possible copy of the plans. You never know, we may be lucky. Thanks again.--Storye book (talk) 17:07, 23 July 2009 (UTC)
Glad to help!
I’ve updated your edits; you were being timid, saying that it is “possibly” a saddle roof (hyperbolic paraboloid) – it is obviously a saddle roof – as you note, it’s built with straight beams – so I’ve stated it more directly, and also clarified some points.
Thank you very much indeed for your careful edits - the article is much improved now - and more importantly we are closer to understanding the building. Cheers.--Storye book (talk) 10:58, 24 July 2009 (UTC)
Will do – I’ve added link concordance (other knot theory articles on which I’ve worked lately were already included) and a link to Category:Knot theory for reference.
I have not come across this terminology before: is there a reference? Unless I am misunderstanding the definition, oriented cobordism is strong, not weak as claimed.
ranicki (talk) 14:20, 24 July 2009 (UTC)
Oops, sorry! This was me confusing myself: in an old edit, I had replaced “a very simple theory” with “weak theory” (as this sounded more formal), and then in a later edit thought “hey, this term should be more prominent!”
I’ve fixed this, removing all weak/strong words.
Regarding the distinction, it’s basically “cohomology theories that geometric topologists care about (unoriented, oriented cobordism)” and “cohomology theories that algebraic topologists care about (complex cobordism and related)”.
As I understand it, the distinction is that the oriented and unoriented cobordism Thom spectra are products of Eilenberg-MacLane spaces: and and this is what is meant by “reduce to ordinary (co)homology”.
If I recall correctly, oriented cobordism is exactly characteristic classes (or rather, characteristic numbers): the cobordism class is not determined by the (ordinary) cohomology ring – they are/include normal data (classifying map) – but the representing object is just ordinary cohomology.
Is this correct? (And perhaps it could be made clearer at the page.)
Thanks for fixing this. Unoriented cobordism is a product of Eilenberg-MacLane spaces (or rather spectra), but oriented cobordism is such a product only at the prime 2, i.e. if odd torsion is ignored. However, both unoriented and oriented cobordism are determined by characteristic numbers: Stiefel-Whitney in the unoriented case, Stiefel-Whitney and Pontrjagin in the oriented case. The Edinburgh Fringe has all kinds of esoteric shows, but not (as yet) anything quite as esoteric as algebraic surgery! ranicki (talk) 07:41, 26 July 2009 (UTC)
No problem – and I’ve clarified/elaborated that while oriented cobordism is determined by characteristic numbers, the ring is complicated to describe.
A vandal removed the section on Sep 13 (see history); I got my WikiGnome on and restored it, fixing these redirects. (As you perspicaciously suggest, "List of Coho" is referenced.)
I did make a mistake in my edit (the log prior is not the only Jeffreys prior on the positive reals). I've fixed the mistake and incorporated a correct discussion of log prior on the page.
Thank you for fixing this (and those above). Wikipedia's mathematics is amongst the best it has to offer, perhaps because it is, ironically enough, the field where WP:OR in general and WP:SYN in particular is most acceptable, as everything correct is just a "routine calculation" (for someone) from verifiable material. -- ToET00:28, 17 September 2009 (UTC)
Glad you like 'em!
Your philosophical points on WP mathematics are well-taken; conversely, one could argue that WP's math articles are some of the least well-referenced, since everyone just figures it out themselves, half-remembered, rather than turning to the bookshelf. This also points to how much of math knowledge is oral - "everyone knows that" (mathematical folklore).
This has been present for me lately in contributions to economics articles - economics is a more scholarly and contentious field than math, so references are more prized and more necessary. For "obvious" or "well-known" statements, Google scholar and the (quite active) economics blogosphere have been invaluable; the former often reveals that concepts are centuries older than commonly understood, while the latter helps transcribe the folklore and the divisions.
Thanks again, and feel free to advise me on other nigh-insoluble problems!
you have a formula for computing the WAL without knowing the amortization schedule. Do you have a source for that formula you can share with me? I see it has been some time since you updated that page, but I'd appreciate any help you can provide.
Thanks.
unsigned comment by Sixhertz (talk) 9:28, 2009 September 18
I don't have a reference to hand (it's just a restatement of "Total Interest = WAL * interest * principal"), but you might find Fabozzi helpful -- I've added him as a ref to WAL; hope this helps.
Hi there. If you have time please check the merge discussion you started at Talk:Long boom with a view to merging Golden age of capitalism with long boom. Im hoping we can close the discussion with a view that the Golden age article will be kept with its current title, i dont mind whats decided about the other articles. FeydHuxtable (talk) 12:42, 10 September 2009 (UTC)
Hi Feyd,
Thanks for sorting this and your extensive research -- agree with conclusion that long boom should be a disambig, as Yvwv has done.
Pursuant to this, I've started a merge discussion at Talk:Golden Age of Capitalism#Merge Post-World War II, with the main question to my eye being whether to call the merged article "Golden Age" or "Post-War" -- perhaps you could weigh in?
I’ve merged the articles, as per discussion. The merged article is at Post-World War II economic boom, which was not your preferred title; I’ve suggested resolutions there, if you feel strongly that it should be “Golden Age” instead.
Thanks for your contributions to discussion and article!
Great work on the merge, the article's looking much more well balanced and well organised now. Shame about the name, but one cant have everything. FeydHuxtable (talk) 13:13, 27 October 2009 (UTC)
You introduced the section "polar coordinates" in the article about the planimeter. I do not understand this section. Maybe you can give a more explicit explanation. Nijdam (talk) 11:51, 25 October 2009 (UTC)
Ok, will do – apologies for the unclear explanation, and thanks for the feedback.
I've re-written the section some, though I grant that it's not terribly clear; it's just making the connection of planimeter to Green's theorem in polar coordinates instead of vector calculus, which I find clearer. Hope it helps though.
I just spent some time getting rid of some of the "inline" TeX in Automorphisms of the symmetric and alternating groups. Some of this was in section headings. When you put TeX in a section heading, it is invisible in the table of contents. That means you had a section heading appearing in the T.O.C. as follows:
The exceptional outer automorphism of
That's it: it ended with "of". When it's not in a section heading, you get things like in which the "2" appears lower than the surrounding letters and the characters are between two and four times the size of the surrounding letters, whereas 23 matches the surrounding letters perfectly. You need TeX to write things like
Regarding LaTeX vs. HTML – your point regarding the poor typesetting of simple inline LaTeX is well-taken.
That said, use of LaTeX vs. HTML is a debated and contentious point – I find LaTeX markup much easier to write and edit (compare X \to Y to ''X'' → ''Y''), and more semantically correct.
The specific rendering problems you mention, while vexing, are technical issues, and are solved by better rendering.
My suggestion is to use LaTeX inline, and fix the rendering if necessary – this ensures consistent style and simplifies edition (and makes your life easier!) – and regardless, we’d like LaTeX rendition to be good.
Specifically:
I have no baseline problem with basic formulae.
The “math is too big” is due to the use of font-size: 125% in Monobook.css, which was discussed at: MediaWiki talk:Monobook.css#Math font-size – it was designed to make serif math work better with sans-serif text, but as you note, it breaks for many (myself included).
Removing the 125% fixes all problems for me (middle is HTML, two ends are LaTeX – right end removes the 125%, fixing the size); note that math is by default set in serif:
Aa Bb 23 Cc Dd
Aa Bb Sn Cc Dd
Aa <math>2^3</math> Bb 2<sup>3</sup> Cc <math style="font-size: 100%">2^3</math> Dd
Aa <math>S_n</math> Bb ''S''<sup>''n''</sup> Cc <math style="font-size: 100%">S_n</math> Dd
Thanks for bringing this up; what I’d suggest concretely is:
Setup a WP:MATH subproject to discuss LaTeX rendering issues, as this is a big deal and warrants dedicated attention; I’d be happy to do so.
Discuss what WP:MATH would like the LaTeX-to-HTML rendering and style to do.
For example, LaTeX could render exactly as you desire (sans-serif, normal size) by making 2 small changes to
Probably couldn't hurt to start a subproject. These issues have been talked about since 2003. When TeX first became available on Wikipedia, "inline" TeX was formatted so that the bottom of the line of text aligned with the bottom of the TeX display, and that looked ridiculous in most cases. It was soon changed to centering, and that also looks ridiculous in some cases. I'm not sure there's been any progress since then. Michael Hardy (talk) 14:54, 1 November 2009 (UTC)
Systematic error
Howdy, your recent edits have introduced a systematic error: PGL is not the automorphism group of the incidence structure associated to projective geometry. It is similar but not isomorphic. The automorphism group of the projective plane is called the collineation group and is larger than the projective general linear group. Similar statements hold in higher dimensions. JackSchmidt (talk) 15:51, 31 October 2009 (UTC)
Oops‼ I thought they were identical!
Thanks for the catch – I’ll fix it and ref to and from collineation.
Howdy, thanks for splitting off the covering groups material.
The section Covering groups of the alternating and symmetric groups#Construction of double covers currently trails off with "As there are two pin groups, there are two distinct 2-fold covers of the symmetric group, but only one of these". From the lead, it looks like one wants to say "For the symmetric group, there are actually two different covers (corresponding to the two different Pin groups), but only one of these is of interest in classifying projective representations." but I'm not sure the comment about "interest" is accurate. Could a citation be added for this discounting of one of the covers? JackSchmidt (talk) 18:34, 18 November 2009 (UTC)
Oops, thanks -- got cut off editing (there's another diagram coming too). The statement discounting one of the covers is from the Encylopaedia of Mathematics (SpringerLink); I've added a citation. Unfortunately, they don't say *which* cover is of interest, and I can't find other refs (well, other than Schur, but looking through a 100-page article of 100-year old German mathematics is not exactly encouraging). Hope it looks better now!
Hi Nbarth, after noticing one of your recent articles at NewPage patrol, I've set the Wikipedia:Autoreviewer flag on for your account. So in future any articles you create will bypass the new page patrol process, as they will be automatically marked as patrolled. Take care and happy editing. ϢereSpielChequers14:38, 25 November 2009 (UTC)
Thank you for the flag and the explanation! (Now I can edit without fear of cluttering the New Page Patrol!)
You're welcome:) More importantly than just new articles, I think that when and if flagged revisions come in the edits of Autoreviewers will be preflagged as not vandalism. So the more editors like you we can give the autoreviewer flag to the less likely it is that flagged revisions will break the wiki. ϢereSpielChequers21:29, 25 November 2009 (UTC)
OIC – thanks!
Over at Wiktionary, there’s a very active patrolling and autopatrolling culture and system, which y’all may find interesting – see WT:WL.
Thanks, I've only made a couple of edits there so I'm not familiar with their culture, but I love the idea of a sysop based patrol of recent changes. We patrol both newpages and recent changes, but many, perhaps most, of our patrollers are not admins. At Newpage patrol we have a whole load of taggers who are not admins, and many of the admins just focus on stuff that others have tagged for deletion. ϢereSpielChequers22:51, 3 December 2009 (UTC)
Huh?
What did you have in mind with this edit? The sentence preceding the dash is false, and what comes after the dash seems to assume that infinitely divisible distributions cannot be decomposable; whereas in fact all infinitely divisible distributions are decomposable. Nor would it have made sense if you had said "need not be indecomposable", since in fact such distributions are NEVER indecomposable, immediately from the definition. Michael Hardy (talk) 23:28, 28 November 2009 (UTC)
Oops – it was a wording fix; I wanted to emphasize that indecomposable is clearly not closed under sum and emphasize that infinitely divisible is much stronger than decomposable, but I introduced an error in the process. Thanks for catching this; I’ve given another stab at wording in this edit, which is hopefully both clear and correct.
Sorry, and thanks for catching it; I fixed a few other instances of it. (Fortunately there isn’t a Borel–Harish-Chandra theorem or the hyphenation would be a mess.)
hi, I saw your new article, and am impressed. I pointed out on the talk page that your internal link to commute contains no mathematical case of the word, but I think I'm the only one who's visited the talk page. Just thought I'd bring that to your attention. Cheers! —Preceding unsigned comment added by Throwaway85 (talk • contribs) 00:51, 3 December 2009 (UTC)
A criticism has been raised of an edit of yours on the talk page of the article J-invariant. You may like to read the criticism, and perhaps respond to it, even though it is now some years since you made the edit in question. JamesBWatson (talk) 11:44, 10 December 2009 (UTC)
Thanks for bringing this to my attention! (And for digging out that I wrote it.)
I’ve clarified the article and replied on the talk page.
(I’d assumed that explicit year/month/day was useful so that the date field didn’t need to be parsed, but as you say, it’s redundant, so I shaln’t need to include in future.)
The money creation articles are a bit of a mess – I’ve a long-standing to-do item to fix them – with information duplicated several places.
I’ve split money multiplier to its own page, as it’s currently 15k, which is rather long – hope you like it!
The point I was making, which is mentioned but often missed, is that the multiplier is a maximum, not an equality – though for long periods it functions as an equality, hence easily conflated. You’re right that this has important implications for monetary policy, though it is also important for the nature of the multiplier itself (Samuelson’s “no 5 for 1 [in a depression]” states this clearly), and I’ve reworded and retitled the section to reflect this (it was not “criticism” so much as “clarification”), and edited the article so that it reflects this.
I’ve also fixed the section in money creation so that it is written in summary style and defers to the main article, though I’ve not fixed the section at Fractional-reserve banking nor included relevant info at monetary policy, though you’re correct that something belongs there.
Hope this begins to address your concerns; I’ll be a bit busy over holidays, but I’ll try to continue integrating and cleaning up, and of course you’re quite welcome to do so yourself – we can review and reconcile when we’ve both time.
It's good that you're putting in the effort to clean up the money and banking articles. Thanks! You're right that they're in a bit of a mess. Part of the problem is the people who've read Rothbard and then decide to change the articles to reflect Rothbard's rather idiosyncratic viewpoint.
I've had a look at the money multiplier page you've put together, and it's nice. You're of course right that the classic 1/rr reflects a maximum for the money multiplier - there's a nice discussion of this in the Mankiw intermediate Macroeconomics textbook, Chapter 18 I think. The standard way to treat this is to define the money multiplier as the ratio between M2 and M0, and explain how it will be between 1 and 1/rr depending on the amount of currency that people decide to hold and the excess reserves kept by banks. I notice that your treatment is different - at some point I may wade in and change it a bit.
Unfortunately, I've been having a busy few months (putting together a new course, and just sent off a paper to a journal) and haven't really been contributing. But I do check for messages pretty regularly. Do let me know if I can help in any way.
No worries re: such an Augean task; I’ve added some references to Mankiw and Krugman, which should help, and mentioned M2 and M0, though it could certainly use revision (the worked re-lending example, in particular, is rather rough).
My main concern for the “Money multiplier” article itself is that readers are clear on the different ways this particular concept is used and defined, since the differences are subtle but have significant impact on monetary policy – hence I’ve fronted this issue, but started with the simple “reciprocal of reserve rate” definition. On other pages we needn’t (and shouldn’t) make a big point of it, as it obscures matters, but on this page we should – for example Krugman, Mankiw (Intro) and Mankiw (Macro) define it at least three different ways (reciprocal of reserve rate, empirical ratio, and component of model), sometimes in slightly different ways over the course of a few pages, and this can be very confusing.
Thanks for the warning about Austrians – agreed that while we want to include all significant views (and Austrian economics is a significant view, albeit heterodox), they shouldn’t dominate or distort main pages like Money; specialized pages like criticism of fractional-reserve banking are rather better places for such argument. I’ll try to avoid stepping on any golden toes!
You asked if there was any way you might help; if you’ve time or familiarity with the literature on this,
one question that would be of particular interest to me is research that analyses the multiplier (presumably there’s something in Friedman & Schwartz, at least data), and especially the re-lending model of money creation; the only treatments beyond the intro “here’s $100 at 10% reserve rate – ta-da, geometric series!” with which I’m familiar are criticisms, in turn by monetary circuit theorists (rather different model), by financial practitioners (PIMCO, dismiss it as “textbook, not reality”), and by Kydland & Prescott (from econometric data/Real Business Cycle theory) – are there empirical studies that support the re-lending model, or is it mostly just used in textbooks (or in modeling?)?
Deep questions. Unfortunately, I don't have the knowledge or expertise to answer them as well as I wish, and as you deserve.
First, for something a little more complicated than multiplier = 1/rr, Mankiw's intermediate macro textbook (chapter 18) has a simple model showing that the money multiplier depends on the actual reserve deposit ratio (which is equal to the mandatory reserve ratio plus the excess reserve ratio), and on the currency deposit ratio (the ratio of cash vs deposits that the public chooses to keep). Specifically, the actual multiplier will be equal to:
money multiplier = CR + 1 / CR + RR
where CR is the currency deposit ratio, and RR is the actual reserve deposit ratio. Even if the mandatory reserve ratio remains constant, the money multiplier will tend to shrink as people hold more cash, and as banks hold more excess reserves. This happens during a financial crisis, which is why financial crisis were associated in the past with a shrinking money supply and economic depression (before the days of quantitative easing). This point was also brought up in Friedman & Schwartz in their "Monetary History of the United States".
Second, unfortunately I'm not familiar with criticisms from monetary circuit theorists or financial practitioners so can't comment on them. As for RBC ala Kydland & Prescott, the simple RBC models do not include money, and make no comment on the effect of money on the economy. However, no one can deny that an increase in the growth rate of money supply reduces nominal interest rates in the short run and increases it only in the long run, and so people working in the RBC literature have attempted to create models with money and the appropriate frictions to generate this result. The RBC and the new Keynesian models seem to be converging, in that as the RBC models add more and more frictions, and the new Keynesians try for more and more micro foundations, the models end up looking very similar to each other. As far as I know, the RBC people have no problems with the concept of a money multiplier, and would not disagree with the mainstream view that new money expands slowly into the economy as bank lending rises.
Third, a major strand of the post-Keynesian literature is on money, and the effect of money on the economy. They hold to a theory of 'endogenous money', which I understand to mean that the amount of money in the economy is actually very flexible, and subject to the demand for money and for bank lending. In the mainstream view, this would mean that both the money multiplier and the velocity of money are not constant, and are elastic to the demand for money. They also do work on credit cycles and economic bubbles. I am not very familiar with the post-Keynesian literature, but Paul Krugman has spoken of them approvingly.
Lastly, about an empirical literature on the money multiplier, there is work on how the multiplier shrinks when either banks or the public want to hold more cash. You can find this in the literature on financial crisis. Also, work on hyperinflations show that monetary injections affect broad money with a lag, essentially because it takes time for banks to relend the money out, and so when the money supply is accelerating very quickly (as happens in hyperinflationary environments) the money multiplier will tend to fall. Unfortunately, I do not have any handy citations available, as I read those papers a long time ago.
Hope this helps. If I wasn't clear about anything, please drop me a line asking me to clarify. best, LK (talk) 10:14, 29 December 2009 (UTC)
Thanks for the (detailed!) explanation and background! As and when I read up on money more (and work on the WP pages), I'll try to incorporate (and cite) the above, and whatever else I may find.
