Talk:List of manufacturers by motor vehicle production

Source data for 2007-07-31 version
Taken from: — Preceding unsigned comment added by DeFacto (talk • contribs) 10:07, 31 July 2007 (UTC)
 * Group,Total,Cars,LCV,HCV,Bus
 * GM,8926160,5708038,3156888,43838,17396
 * Toyota,8036010,6800228,1049345,122569,63868
 * FORD,6268193,3800633,2386296,81264,0
 * VOLKSWAGEN,5684603,5429896,219537,29175,5995
 * Honda,3669514,3549787,119727,0,0
 * PSA,3356859,2961437,395422,0,0
 * Nissan,3223372,2512519,570136,134874,5843
 * Chrysler,2544590,710291,1834299,0,0
 * RENAULT,2492470,2085837,406633,0,0
 * Hyundai,2462677,2231313,966,145120,85278
 * FIAT,2317652,1753673,450544,89071,24364
 * Suzuki,2297277,2004310,292967,0,0
 * DAIMLERCHRYSLER,2044533,1275152,378278,340296,50807
 * Mazda,1396412,1169640,223995,2777,0
 * Kia,1381123,1181877,197060,0,2186
 * BMW,1366838,1366838,0,0,0
 * Mitsubishi,1313409,1008970,296431,8008,0
 * Daihatsu,1084721,905932,166667,12122,0
 * AVTOVAZ,765627,765627,0,0,0
 * Fuji,587274,507552,79722,0,0


 * But is it really correct to list GM before Toyota? Toyota is larger than GM if including Hino and Daihatsu, or should that two car companies be separated? Why are they separated from Toyota despite being subsidiaries of Toyota? 81.230.149.213 (talk) 10:37, 27 January 2017 (UTC)
 * Per WP:OR and WP:SYN we can't decide how to combine the results based on our own knowledge/understanding. We need to find external sources that make the claims.  Currently most sections in this article are based on OIAC data.  Since that is the only source the article should reflect that source including it's groupings and rankings.  If other sources can be found that group things differently then we have a WP:WEIGHT issue that could be balanced by presenting both the data as presented by OIAC as well as grouped as suggested by a RS.  Springee (talk) 14:51, 24 February 2017 (UTC)
 * What has happened to the reference ? Has it become deleted? I wonder if this reference was replacing this older, as the older included errors such as wrong numbers of produced vehicles for G.M. and Ford, as well as wrong rankings (with Volkswagen ranked as 3rd and Ford as 4th despite Ford's production which was larger than Volkswagen's), which were corrected in a new revised version. Besides this, I wonder why Toyota isn't listed as the 1st, before G.M., despite that Toyota actually had larger production than G.M., if including Daihatsu and Hino. Was it because OICA had a norm to not include subsidiaries if those were not owned to 100%? 212.100.101.104 (talk) 04:51, 4 November 2023 (UTC)
 * OICA no longer have that URL. archive.org has a single entry at https://web.archive.org/web/20070809142844/http://www.oica.net/htdocs/statistics/tableaux2006/Ranking06-2.pdf (9 August 2007) and after that it just lists the page as "ERROR 404 - NOT FOUND". archive.org lists https://web.archive.org/web/20230000000000*/https://www.oica.net/wp-content/uploads/world-ranking-2006-july-08.pdf as first seen in 2009 - whihc is after the other one was removed. I can only assume that the remaining PDF for 2006 was meant to replace the missing one.  Stepho  talk 05:07, 4 November 2023 (UTC)

OR in tables that combine manufactures
Several of the tables in this article combine counts in ways that are not supported by the referenced sources. The 2004-2008 tables note they are combining manufacturer results in a way that is not presented in the cited source. This is WP:SYN because combines data in a way that changes the reported results. It's also WP:OR because no RS is offered to justify the combinations. Springee (talk) 14:14, 24 February 2017 (UTC)


 * Ford owned Mazda 2004–2008. Toyota owned Daihatsu and Hino. Hyundai and Kia are merged into one group. But even though they are merged into one motor group, OICA lists them separately, I don't know why. That's why they are listed together here on Wikipedia. What's so strange with that? BjörnBergman 14:33, 24 February 2017 (UTC)
 * Since they are listed separately in the RS we have to list them separately in the tables. For example, the 2007 PDF shows GM as the #1 manufacture.  If the table is based on OICA's report then why doesn't it agree with the report? []  This is the same issue with the GM article and the reason why I reverted your good faith edits.  If you want the OR tag removed please find a source that supports the ranking shown in the table.  The current table shows Toyota as #1 and GM as #2.  The source shows GM as #1.  We have OR SYN issues.  I would suggest pinging some additional editors to get their views.  In the mean time I'm restoring the tag because we have a clear discrepancy between the source and the article tables.  Springee (talk) 14:41, 24 February 2017 (UTC)
 * Also prefer listing per source and no further improvements. There may be some explanation note, but no changes in cited data.--Jklamo (talk) 16:09, 24 February 2017 (UTC)

But the ranking in the source oica.net is usually not fully correct. Toyota owned Daihatsu and Hino already back in 2001 but even though, OICA lists them separately rather than together which usually is incorrect. OICA listed them together for first time in 2008. If they had done so already in 2006, it would be possible to see Toyota as the largest manufacturer already then. Why not just let the manufacturers be listed as they are listed, for example with Ford and Mazda together in 2004-2008? At that time, Mazda was a member brand of Ford but even though, Ford and Mazda were listed separately rather than together which in fact is wrong. The OICA sources have not listed the manufacturers fully correct. Only because Toyota is ranked #2 as of 2006 according to OICA, that doesn't mean Toyota wasn't ranked #1. Why it is ranked #2 rather than #1 is because they have not included Daihatsu and Hino but listed them all three separately even though they should not be listed separately. If listed together, the total production is more than that of GM and the whole Toyota group is larger than GM. BjörnBergman 17:52, 24 February 2017 (UTC)
 * Again, this isn't a case of what is right or wrong. A number of sources have raised your same concern, we can include both.  The problem here, and at the GM article, is we have RSs that say GM was #1 through 2007.  In this case the only referenced source says GM was #1 through 2007.  As editors we can't just say the source is presenting the data incorrectly and take it upon ourselves to change the data. This is an issue of WP:RS as well as WP:OR and WP:SYN.  I think your POV has merit but without reliable sources that make the claims we can't put the claims in the articles.  In the GM article we have at least three sources that say GM was #1 in 2007.  We have none that say otherwise thus we report GM was #1 in 2007.  If we have a source that says Toyota was #1 when Daihatsu and Hino are added then we should add that as a note.  In this article we should report the data as the source reports it then include a statement that some combine the three Japanese results which would have put Toyota ahead (this would require a reliable source).  Springee (talk) 18:13, 24 February 2017 (UTC)


 * But oica.net IS a reliable source. That is the source which publishes annual production of vehicles. BjörnBergman 18:16, 24 February 2017 (UTC)
 * I'm not disputing that they are a reliable source. I'm saying we can't interpret what they say because we think their groupings are incorrect.  OICA's table says GM was #1 in 2007.  That is what we have to report.  If another RS says OICA should have grouped the data differently and thus the results should have been X, we can report that as well.  Currently the article takes the reported values, regroups them then reports as if that was what the report said all along.  Per WP:OR and WP:SYN We can't do that.   Springee (talk) 18:27, 24 February 2017 (UTC)


 * But all sourced saying the largest automaker as of 2007 is G.M. are wrong. What a news article or similar source says is NOT ALWAYS correct, at least not in this case. For example, most news articles say Toyota surpassed Ford as the second-largest automaker in 2003 but according to OICA which publishes the annual vehicle production every year, that is incorrect. Ford was still second-largest in 2003, and even in 2004 if including Mazda which then was a Ford member brand. I think what OICA says is much more reliable than what other sources say. BjörnBergman 18:41, 24 February 2017 (UTC)
 * Again, that is WP:OR and WP:SYN based on the OICA results []. To put Ford over Toyota in 2004 we need to find a RS that states as much and even then, per WP:WEIGHT we need to include results that put Toyota ahead.  We simply are not allowed to change what a source says because we would group things differently.  Also, what I'm saying is we SHOULD report what OICA actually says.  They actually show GM as #1 in 2007.  They show Toyota as #2 in 2004.  That is what the table should show.
 * Per WP:SYN  ...do not combine different parts of one source to reach or imply a conclusion not explicitly stated by the source. That means don't add Ford and Mazda data to put Ford above Toyota unless that is what the source does.
 * Per WP:OR  The phrase "original research" (OR) is used on Wikipedia to refer to material—such as facts, allegations, and ideas—for which no reliable, published sources exist.[1] This includes any analysis or synthesis of published material that serves to reach or imply a conclusion not stated by the sources. To demonstrate that you are not adding OR, you must be able to cite reliable, published sources that are directly related to the topic of the article, and directly support the material being presented. Springee (talk) 18:55, 24 February 2017 (UTC)


 * Started the process. It's tedious so I'm going slowly. Springee (talk) 02:03, 25 March 2017 (UTC)


 * When reading this, you can see Ford owned Mazda by 2003/2004 but Mazda's sales were, though that, not included in Ford's total sales. If included, Ford would be larger than Toyota (including Daihatsu and Hino) in 2003 by sales. But according to OICA, Ford was No.2 and Toyota No.3 in 2003. Why rank Ford as No.3 in 2004 rather than No.2 though Ford including Mazda produced about 100,000 vehicles more than Toyota including Daihatsu and Hino?

Citation: ''Of note, Toyota's numbers include Daihatsu and Hino, two subsidiaries more than half owned by Toyota. Ford's numbers do not include Mazda, which it effectively controls, though it owns about a third of the company. If Mazda were added in, Ford would still be on top.''

This citation means Mazda and Ford should be listed together rather than separately. BjörnBergman 19:43, 25 March 2017 (UTC)


 * BjörnBergman, you were part of the discussions here [] where and I explained why the table with the combined data is WP:OR.   concurred on this page.  We do not have the liberty to combine someone else's data, especially when the source is considered highly reliable for the information in question.  Please read WP:OR before trying to combine the data in ways that are not in the source. Springee (talk) 22:37, 25 March 2017 (UTC)


 * Agree with Springee, OICA is a well respected data source that represents the automobile industry without bias and over a large number of years. It is not a regional concern and does not favour any single manufacturers or group of manufactures. It is accepted by the industry itself. Whereas newspapers will often use different methods of counting in every report of the facts (do you include subsidiaries, knock down kits, cars/trucks/buses, what percentage counts as an owned subsidiary, does a subsidiary count 100% of its sales or does a 30% ownership only count 30% of its sales, etc). Note that in 2003/2004 Ford only owned 33.3% of Mazda. That probably makes Ford the majority shareholder but not necessarily that Ford outright owns Mazda, hence explaining why OICA lists Ford and Mazda separately.  Stepho  talk 23:04, 25 March 2017 (UTC)

Notice to both and, according to the spirit of WP:BRD, it is best to not change the rankings around in the article until we after have a consensus on the discussion page.  Stepho  talk 03:02, 26 March 2017 (UTC)
 * , Stepho-wrs made a perfectly reasonable suggestion and asked that we not edit and revert the article until this matter was settled. You reverted my changes as well as changed Rayukk's numbers.  I've opened a NOR discussion to address this.  If it get's no traction in the next week then I would say we have local consensus to use the strict OICA numbers.  Certainly sources that suggests other grouping could/should be added to the article but it local consensus is the tables should match the source.  Springee (talk) 01:12, 28 March 2017 (UTC)


 * I've posted the question at NORN ] Springee (talk) 13:16, 26 March 2017 (UTC)

I've corrected the tables so their numbers are the same as the source data. I did this because we have a local consensus for this change. The NOR request for help yielded no feedback so I'm going with local consensus developed here and the related discussion on the GM talk page [] (consensus was also for setting the tables to match the source tables. Springee (talk) 01:45, 31 March 2017 (UTC)

, we have local consensus against altering the OICA data. , have both voiced objections to your changes. This morning you have reverted Jklamo and myself more than once. Again, your concerns are valid but there is simple way to handle this. In the cases where a reliable external source disagrees with the OICA groupings we can add that as a note to the article. The current groupings misrepresent the source data and thus are not OK per WP:NOR. Springee (talk) 15:21, 31 March 2017 (UTC)

, per the above discussion we can't regroup the OICA's data. We can use reliable sources to note that other sources have grouped the data differently but it's a violation of NOR for us to regroup the data. Also when you regroup without attribution in the table it misrepresents the source data. Springee (talk) 12:25, 2 April 2017 (UTC)
 * Thanks for your note, I wasn't aware of the prior discussions. I agree that we must respect the source. Actually I was about to update the data for some years where OICA issued final numbers and our table is still based on preliminary numbers they issued earlier. About combining data, we are in fact allowed to make simple calculations based on sources, per WP:CALC. Synthesis would be taking one source and another and making inferences which neither source is making. A simple addition is neither SYN nor OR. Case in point: I noticed an inconsistency in the numbers reported about Hyundai and Kia, so I went back to the sources and noticed that Hyundai and Kia figures were combined in 2004 and 2005, separated from 2006 to 2008, and combined but only labeled "Hyundai" from 2009 to 2015. If we split Hyundai and Kia over 3 years in the middle of the time series, our ranking tables would be inconsistent and the progress graph would make no sense. Then I checked whether Hyundai and Kia had merged recently, which would justify listing them separately before the merger, but it turns out they have operated as a group since 1998. I applied a similar reasoning to Chrysler, which the OICA summary lists combined with Daimler in 2004 and 2005, then separately from 2006 to 2012, and finally combined with Fiat (but only labeled "Fiat") since 2013. Our tables also had some inconsistency in the labeling between "Daimler AG" and "DaimlerChrysler", so I checked the history and found that Chrysler was spun off from DaimlerChrysler in 2007, so that's the year where we should report them separately, then it was fully absorbed by Fiat in January 2014, so that numbers should be grouped as FCA starting in 2014 (but OICA already groups them in 2013, so we stick with that). — JFG talk 13:44, 2 April 2017 (UTC)
 * , Prior to your edits we did have notes explaining the regroupings but that was still a problem with OR since we didn't have external reliable sources supporting the groupings. We have to assume that the OICA had some reason for their groupings so changing their groupings is changing what they have reported.  I would suggest we keep the rank order the same as the OICA's per each year and then either inside of the table or as a note at the end of the table explain why a source would change the groupings.  We need an external source that would support say grouping Mazda for certain years if OICA doesn't group that way.  I would suggest that, with,  and  we come to an agreement that addresses the concerns over various ways things are grouped and the concerns regarding WP:OR.  Springee (talk) 00:24, 3 April 2017 (UTC)
 * I see that added a bunch of notes, that's a good start. It shouldn't be hard to find authoritative sources confirming the groupings of automotive brands within a group. In fact, we should perhaps have a dedicated section of the article to explain this, along with various fusions and divestitures that happened during the time period we cover. I believe this would be helpful to readers in order to understand what is behind the names and figures listed.
 * Regarding the "loose" groups such as Nissan / Renault (just recently joined by Mitsubishi), they are indeed still separate companies but they also operate very much like other groups, except the capitalistic structure is not traditional. In fact, the structure of Renault Nissan cross-shareholdings + common industrial and R&D ventures became a model for some later alliances, e.g. Fiat Chrysler, who must optimize their operations while preserving the local work cultures and brand images (and investor sensitivities + local regulations). So the OICA only publishes sales figures, but a lot of the press also reports combined sales when looking at the "top 10" world automakers.
 * We could perhaps use the same process whereby the European Union is reported in various world rankings on Wikipedia: keep the individual countries listed and ranked, and add an unranked EU entry representing the de facto association of member countries. Example: List of countries by GDP (nominal). What do you think? — JFG talk 09:59, 3 April 2017 (UTC)

Gentlemen, I see that some of you are actively editing the article to show you own viewpoint. This has had the very natural response of other editors changing it back to their viewpoint. Why are you wasting your own time and our time?  Stepho  talk 23:15, 3 April 2017 (UTC)


 * For my own part, I've been doing my best to improve the clarity of the article, not putting forward a personal viewpoint, and taking into account other editors' opinions. When groupings are unclear or inconsistent, it's perfectly appropriate for editors to make them clarify or explain the inconsistencies. The footnotes that were added by and myself bring clarity. Regarding the waste of time, I think this question should be addressed to people like  who just bluntly reverted to an old version with an ugly graph and didn't participate on the talk page.  to your point about WP:V, everything is derived from the OICA sources and easily verifiable by any reader, especially as we clarify the situation when OICA summary tables are inconsistent from one year to the next. I would respectfully urge you to look at the latest state of the article and make specific comments on the talk page with a goal to improve it further. — JFG talk 00:30, 4 April 2017 (UTC)


 * The OICA summary table excludes SAIC-GM, it's as simple as that. Listing any other figure is a violation of WP:V and blatant WP:OR. Let's not make up figures out of the air. Pizzamall (talk) 15:44, 4 April 2017 (UTC)


 * We can debate the appropriate presentation of SAIC-GM figures without reverting to an ancient version which has many other issues. Several editors are working collaboratively to improve the article. If your only concern is SAIC-GM, have you read the note which explains the odd ungrouping for 2015? WP:V is satisfied per WP:CALC as explained earlier. — JFG talk 17:08, 4 April 2017 (UTC)


 * I have addressed your concern by listing both numbers and putting GM back to 4th rank as in the OICA summary source. I kept the group numbers for the graph to ensure consistency (readers might wonder how GM could suddenly lose 2 million vehicles of production capacity in 2015 barring extraordinary events). — JFG talk 17:23, 4 April 2017 (UTC)


 * , as far as I can tell, all the major players are trying to improve the article. But that doesn't mean we are doing it right. Each edit we do because we feel in our heart that it is good and right and proper but without consensus on the talk page is going to encourage edits on the other side to make the reverse edit because they also feel in their heart that it is good and right and proper.
 * The article now has a footnote that says "Including production figures from the Chinese SAIC-GM joint venture,[3] which the OICA left out from the 2015 GM total contrary to prior practice; they are combined here for consistency with previous years" and many similar ones. This is effectively saying that we got our raw productions numbers from OICA but that we can also read OICA's collective mind and determine that they made a mistake and that we are so smart that we can correct it for them. An alternative view is that OICA have some criteria for grouping companies and that in the years in question the companies did not meet that criteria. This might be the number of shares held or some other method but unfortunately they don't tell us their method. However, I do trust that they had lawyers and accountants double check the work and that each company involved had lawyers and accountants double check the work. Hands up anybody here who thinks they can outsmart a battery of lawyers and accountants on their turf! Footnotes such as this are synthesising information and is original research. The best that we can do in terms of footnotes is something along the lines of either "Including production figures from the Chinese SAIC-GM joint venture,[3] which the OICA left out from the 2015 GM total contrary to prior practice" (without actually combing the results), or "Source XXX combined these companies to make a combined ranking of #xxx [supporting reference]". Or to put it more bluntly, we are allowed to report figures from sources but we are not allowed to regroup them to suit ourselves.
 * Please, let's thrash it out here on the discussion page before we hack and slash at the article. At the moment is is useless to improve the article in other areas (eg I'd like to format the references properly and another editor would like to format the tables consistently) because it may be reverted as part of this disagreement.  Stepho  talk 22:50, 4 April 2017 (UTC)
 * I totally agree with Stepho-wrs on this. So long as we can find reliable sources that say/group things differently than the OICA data I like the idea of putting a note in the table with the grouped total but I think the table rank and stated values (ie the values not associated with notes) need to be verbatim from the OICA documents.  Springee (talk) 03:42, 5 April 2017 (UTC)
 * I believe we are largely in agreement and the current tables are respectful of the OICA summary reports. I have placed all computed values on separate lines with appropriate explanations in the footnotes. We can still tweak the footnotes if necessary, I'm happy to discuss any specific wording you'd like to suggest. I would only be uncomfortable with ungrouping Hyundai / Kia for 2006–2008 because that would be grossly inconsistent. OICA numbers for that period are correct, it's just that they chose somehow to present Hyundai and Kia separately in 2006, and then went back to showing them together in 2009. I don't suppose that "an army of lawyers" was involved, probably just the judgment of whomever compiled the statistics for publication in those years. We could even ask them for clarification.
 * On a philosophical note, Wikipedia's chief goal is to be informative to readers, so discussing what we are "allowed" to do must always take the reader's perspective into account. If sourced information is unduly confusing to readers (e.g. why did GM apparently lose 20% of its production in 2015?) it is our duty as editors to clarify the situation using an appropriate interpretation of the sources and verifiable explanations. I personally do not care who was #1 or #3 in what year, but I do care when as a reader I stumble upon something that looks weird and isn't explained properly. In fact, that's the reason I got involved in this article recently. — JFG talk 10:40, 5 April 2017 (UTC)
 * If you're gonna add SAIC-GM, you have to add every joint-venture of every automotive company in China. It is obligatory by Chinese law that any manufacturer who wants to sell vehicles in China first has to do a joint venture with a Chinese car company: source, another example: Volswagen-FAW -rayukk &#124; talk 14:39, 6 April 2017 (UTC)
 * Yes, and Volkswagen production figures for China (via FAW) have always been included in the OICA summaries. — JFG talk 12:49, 10 April 2017 (UTC)

Have you noticed that the edits for the last week have been nothing except for reverts of reverts - also known as edit warring. This is what happens when both sides think they are so absolutely right that they can ignore the talk page and force their own view of things onto the article. By all means, thrash things out on the talk page but leave the article alone until the discussion is over.  Stepho  talk 20:53, 10 April 2017 (UTC)
 * Well, we must have some combination of editing the article and discussing, otherwise it's harder to know what exactly we are discussing. With some "competition" in article versions, editors can actually choose which of the changes are worth keeping and which are not. Thanks to article history, nothing is ever lost. If you compare the article standing today to the article standing a few weeks ago, I think we can agree that there has definitely been some progress. The next step that seem to have rough consensus is to settle on a single presentation format and to limit display to the top 20 entries. — JFG talk 07:20, 11 April 2017 (UTC)

Including Mazda in Ford?
Is it correct to list Ford and Mazda together for the productions of 2004-2008 or should they be listed separately as in the OICA lists? Why are they listed separately in OICA's lists? 81.230.149.213 (talk) 19:10, 19 February 2017 (UTC)
 * I have updated your footnotes documenting the grouping of subsidiaries which are sometimes still listed separately by OICA. However, I think Mazda should be completely left out of Ford, because it was only a 33.3% affiliate at the peak of their partnership from 1996 to 2008. OICA has always listed them separately. — JFG talk 11:10, 3 April 2017 (UTC)

OICA HAS included (at least a part of Mazda) in Ford's total production, see for example this. In which way is it incorrect to include Mazda in Ford? Do you think it is incorrect to include Kia in Hyundai as Hyundai just owns about one third of Kia? BjörnBergman 21:05, 3 April 2017 (UTC)


 * You have shown that Ford and Mazda were joined in the UK. But your reference still lists them separately for every other country. You still need to prove that Ford owns Mazda world-wide. OICA obviously disagrees with you because its world-wide list show them as separate companies.  Stepho  talk 22:55, 3 April 2017 (UTC)
 * The percentage of shareholding is not the only criterion to consider if companies joined as an integrated group, and I believe that OICA are correct in their judgment to group Hyundai / Kia and to keep Ford / Mazda separate. Even in the heyday of Ford's involvement with Mazda, they labeled their relations as a "partnership", not as a "group". 33% was the largest that Ford ever had in Mazda, and today they have almost completely divested their share. The Hyundai / Kia case is different because:
 * the companies have a longstanding and still current relationship: all finance and industry sources confirm that "Hyundai and Kia have merged in 1998", there is no dispute over this fact, despite their arcane conglomerate structure;
 * the various group businesses are deeply intertwined, including steel production, parts manufacturers, distributors, shipping, finance, etc. That's the Korean way with chaebols; our own article on Kia says Kia Motors is owner in 22 different Hyundai companies. Their ownership percentages range from 4.9% up to 45.37%.
 * although Hyundai Motor Company has only a 33.88% stake in Kia Motors, the rest is largely in the hands of the public as part of the 62% stock market float; the few remaining percents are in the hands of family stockholders and 1% is self-held
 * the OICA sources have consistently reported Hyundai and Kia together except for the years 2006–2008, whereas nothing special happened to the corporate structure during that time.
 * Hope this helps. I shall restore my latest version because I put a lot of work in there and there's still much to do; we can discuss further changes up from there. — JFG talk 01:25, 4 April 2017 (UTC)


 * But the reason that OICA lists Mazda separately from Ford may be because Ford just owned one third of Mazda but that does not necessarily mean Mazda wasn't a member brand or affiliate of Ford. According to several news articles, it was. BjörnBergman 19:54, 8 April 2017 (UTC)
 * Mazda was included in Ford's total in a 1999 report. In this report, Ford was expected to increase it's production from about 7.77 million vehicles per year in 1999 to 9.15 million in 2005 (however, Ford's production now stagnated and din't increase at all).
 * OICA has reported Ford with about 6.5 million vehicles in 1999, while this report says about 7.77. It must be because the report included Mazda, which produced about 1 million and was owned by Ford, while OICA, however, excluded Mazda from Ford despite Ford's ownership in Mazda.
 * Ford owned a 33.333 % stake in Mazda, just as Hyundai did in Kia, but Hyundai-Kia is listed as one manufacturer while Mazda is listed separately from Ford. 212.100.101.104 (talk) 13:27, 4 February 2023 (UTC)
 * An important note: The German article, which has statistics for every year since 2006, has now a note in the statistics for the years 2007 and 2006 that Ford owned around 1/3 of Mazda which, however, is counted separately and not combined with Ford (despite being a brand in the Ford Motor Company just as Jaguar, Land Rover, Volvo, Lincoln). The note is that if Mazda's production is included in Ford's total, Ford's total vehicle production is more than 7 and ½ million vehicles, rather than only around 6.2 million.
 * And if combining Mazda with Ford, Toyota did not surpass Ford as the world's second largest automaker in 2003 as most people believed, as Ford had more produced vehicles than Toyota both in 2003 and 2004.
 * 2003: Ford (with Mazda) had more than 7.6 million produced vehicles - Toyota (with Daihatsu and Hino) only around 7.1 million
 * 2004: Ford (with Mazda) had almost 7.92 million produced vehicles - Toyota (with Daihatsu and Hino) only 7.87 million 212.100.101.104 (talk) 22:05, 7 October 2023 (UTC)


 * From the 2004 table:

With Ford+Mazda beating Toyota+D+H by. Which at these scales is miniscule. Nevertheless, this is all creative accounting and synthesis - we either trust our sources or we toss them out. For whatever (unspecified) reasons, OICA choose to include Toyota, Daihatsu and Hino together for most years but not in the 2003-2007 reports. Likewise, OICA never added Ford Mazda.  Stepho  talk 22:47, 7 October 2023 (UTC)


 * https://www.latimes.com/archives/la-xpm-1999-jun-07-fi-45050-story.html
 * According to this 1999 report, Ford was expected to produce around 7.77 million vehicles in 1999. However, OICA claims Ford only produced 6.5 million, as OICA excludes Mazda which produced around 1 million, while the 1999 report includes Mazda. According to a 1999 analysis, Ford (including Mazda, Volvo, Lincoln, Mercury, Aston Martin, Jaguar) was expected to increase its sales from estimated 7.77 million vehicles in 1999 to 9.15 million vehicles in 2005, and thereby surpass General Motors as the largest automaker, but now Ford's production stagnated so that Ford didn't even keep its title as the second largest automaker, but became surpassed as the second largest by Toyota, in 2005 if counting Mazda combined with Ford. 212.100.101.104 (talk) 18:33, 8 October 2023 (UTC)
 * Is it wrong to add a note that Ford produced more than a million more vehicles that what OICA has reported, if including Mazda? The German article has that note in the statistics for the years 2006 and 2007. If you think that note is incorrect or misleading information, then contact the German administrators and ask then to remove the note. 212.100.101.104 (talk) 21:45, 8 October 2023 (UTC)


 * We are going around in circles again. OICA lists them in that order and with that grouping of companies. To change them is WP:SYNTHESIS. At best we can leave a note that says other sources group/order them differently.
 * German Wikipedia can be used as a source of inspiration for us here but ultimately has no bearing on what we do here. Likewise, our decision here has no bearing on them. You can inform them if you wish and then it is up to them to make their own decision.  Stepho  talk 22:59, 11 October 2023 (UTC)


 * This can be discussed further on dewiki. 217.21.226.230 (talk) 09:48, 12 October 2023 (UTC)


 * No. That is a site for German speakers about the German Wikipedia. It has nothing to say about what happens on English Wikipedia (although its participants are always welcome to contribute to the English Wikipedia). The English Wikipedia is only bound by discussions on the English Wikipedia talk pages. Likewise, the German Wikipedia is only bound by discussions on the German Wikipedia talk pages.  Stepho  talk 10:35, 12 October 2023 (UTC)
 * The article on German Wikipedia, however, has a note on the statistics of 2007 and 2006 that Ford had more than a million more produced vehicles than what OICA has reported, if Mazda is combined with Ford. However, OICA has listed Mazda as an own automaker, rather than combined with Ford, despite Ford's ownership of a 33,4 % controlling-stake in Mazda 1996 to November 2008. 212.100.101.104 (talk) 15:38, 12 October 2023 (UTC)

Make top 20 the standard cut?
The article lists the top 15 manufacturers from 2004 to 2008, then some arbitrary large number from 2009 to 2011 (with a different format), then the top 20 for 2012, top 30 for 2013 and again top 20 for 2014 and 2015. For the sake of consistency, and to avoid bludgeoning the page with relatively smaller manufactuers, I suggest to make the cut at the top 20 for every year. Opinions? — JFG talk 16:09, 2 April 2017 (UTC)
 * I would definitely support moving to a single format for all the tables. I prefer the 1-15 tables use for 2004-2008.  I think the bar graph things are somewhat harder to read even if they do provide a level of magnitude.  They also are rather bulky.  Springee (talk) 00:15, 3 April 2017 (UTC)
 * I definitely agree with single format. But disagree with cutting, as manufacturer with 500 000+ motor vehicle is simply not small and maintaining list based on single source updated once a year is not so hard.Jklamo (talk) 15:16, 3 April 2017 (UTC)
 * If we cut at 500,000 vehicles, this gives roughly the top 30; if we cut at the top 20, that's roughly 1 million+. I don't mind either way, as long as we are consistent. I do have a slight personal preference for 20, so the page doesn't get too long. — JFG talk 16:12, 3 April 2017 (UTC)
 * I think we are largely aligning here. I'm ok with more than 15.  20 seems like a nice cut off but I'm also ok if we say production volume over x/year.  The same format for all is really my preference in this case.  I would like to keep the break out of cars, trucks, buses etc.  I'm still going to say I think we need the tables to rank the same as the source.  When I get a chance I'll put together a sample table as a proposal to address alternate groupings. Springee (talk) 16:18, 3 April 2017 (UTC)

Request for comment on including production numbers of subsidiaries in the parent company's numbers
–rayukk &#124; talk 18:43, 9 April 2017 (UTC)
 * Option 1: Take the production numbers directly from the (highly reputable) source OICA, which may in some cases not accurately represent the current market as the production numbers of subsidiaries are not (always) included in the parent company's numbers.
 * Option 2: Add the numbers of all the subsidiaries of a parent company together and provide the sum, thus altering the original source material.

Survey

 * Option 1, because I believe option 2 to be original research. –rayukk &#124; talk 18:43, 9 April 2017 (UTC)


 * Option 1 but... I don't think we should say "may in some cases not accurately". Without information from the OICA about their methods I'm not sure we can say their method is wrong or inconsistent.  We can present other sources that suggest alternative groupings (secondary groupings for lack of a better term) and I would suggest that such information be included inside the tables almost like a footnote.  But I think the core issue is, as Rayukk said, one of NOR.  If the article says the numbers come from OICA tables then we should faithfully reproduce the rankings and numbers from the table.  In cases where reliable sources suggest the groupings are wrong we should add that material as well (with citations).  We shouldn't add secondary groupings without a RS supporting that grouping.  So if no source groups say Ford and Mazda's production numbers together then we can't either.  If a NYT article says Toyota's passed GM in 200X when Hino etc are added in then we should include as a note in the table.  Reordering the OICA data in a way that isn't what OICA published is WP:OR.  Springee (talk) 01:04, 10 April 2017 (UTC)


 * Option 1 OICA numbers are consistent and mainly dealing well with double-counts.Jklamo (talk) 07:46, 10 April 2017 (UTC)
 * The devil is in the details. Some of the published summaries exclude double counts (see 2004–2010), some don't (2011 onwards, see SAIC figures). We must read the sources carefully and apply logic. The overarching goal must be to correctly inform readers. — JFG talk 12:59, 10 April 2017 (UTC)


 * It depends – Option 1 should be the general rule, except when OICA figures are inconsistent from one year to the next. For the benefit of readers, we must keep comparable statistics, even if this requires reading the OICA figures in detail and not stopping at their summaries. Hence, Hyundai and Kia should always be grouped because OICA only skipped the grouping during three years in the middle of the reporting period, whereas Ford and Mazda should always be separate because OICA never showed them grouped. The Toyota / Daihatsu / Hino situation is somewhere in between because OICA has sometimes showed them grouped and sometimes separated. My hunch is towards consistency. — JFG talk 12:47, 10 April 2017 (UTC)


 * Option 1 OICA is a professional body whose numbers are accepted by the industry. We cannot have footnotes that say OICA is outright wrong because a) you haven't proved cases like Ford owns Mazda (33% of shares is enough to control it but does that count as owning it - in a legal sense) and b) you don't know what criteria OICA used for grouping companies. However, we can have footnotes that say something like "organisation XXX differs from OICA by grouping Ford and Mazda together to rank them combined as #2 [reference]" (underlining only my emphasis for this discussion, not for the article).  Stepho  talk 21:01, 10 April 2017 (UTC)
 * I agree with your reasoning for Ford / Mazda, which OICA has never shown grouped. However, what do you think of the other cases, where OICA alternates between grouping and ungrouping depending on years? — JFG talk 07:14, 11 April 2017 (UTC)
 * The problem is we don't know why they are and aren't grouped at various times. We can guess but then we are both generating OR and misrepresenting the information from a reliable source.  Hence the argument for simply reproducing the tables and adding comments backed by reliable sources. Springee (talk) 10:27, 11 April 2017 (UTC)


 * Option 1 - no car expert - just summoned by bot. I disagree that adding two or more of an automaker's subsidiaries' production figures together and posting the sum violates WP:OR, as long as you have a source that says x, y and z make up ABC company.  Per WP:OR, such a statement can be attributable without being attributed.  In the same vein, I could add the country totals up and say that this is the world's countries' automobile output, for the top 20 car companies, and it would not be original research.  But, more importantly to this discussion, it seems that since the main source is a list that seems to change in methodology, it is simpler to pick one option.  Unless someone from OICA can come here to the talk page and explain what their logic is (I'll reach out and see) it would be too hard to maintain this, and the numerical variation would be even harder to track.  Perhaps a compromise would be use option 1 but put footnotes under each table explaining which companies could potentially be added up to get to true totals, which would at least give curious readers with a calculator some direction. Timtempleton (talk) 21:36, 11 April 2017 (UTC)
 * Quick follow up. I sent a note to OICA on their feedback form.  Hopefully someone will come by and help us understand their methodology. Timtempleton (talk) 18:43, 17 April 2017 (UTC)


 * Option 1 Similarly no expert, but when there are industry-standard figures we use them, giving whatever explanation is necessary. Not using the industry standard is original research.  DGG ( talk ) 09:42, 12 April 2017 (UTC)
 * Option 1.  OICA is a highly reputable source and it is a more concise method of giving the information. Dean Esmay (talk)
 * Option 1 - Who decides when to add the additional figures? It also amounts to original research once we start adding the sources together. Stick with the reliable source we have for each one. --CNMall41 (talk) 22:26, 17 April 2017 (UTC)
 * Option 1 - Seriously, people need to stop inventing numbers out of the air and look at WP:V - Only sourced numbers can be used in Wikipedia. Adding numbers from 3rd party sources is a blatant violation of WP:OR. Such 3rd party numbers can be challenged and removed at any time as per WP:V. Pizzamall (talk) 07:22, 25 April 2017 (UTC)
 * There are no third-party sources: all numbers are traced to various OICA reports. — JFG talk 08:09, 26 April 2017 (UTC)


 * Option 1 - (Invited here by a bot) Producing our own data is OR. If manipulation of data is required, we need to find a reliable source that does that manipulation. [Note: It's best RFC practice to state the question(s)in the simplest possible form without comments and explanation. RFC commentary is best confined to the Discussion section. And it's usually a good idea to have a Discussion section so we don't have discussion threads in the Survey section.] Jojalozzo (talk) 17:44, 27 April 2017 (UTC)
 * Option 1 - (Invited here by a bot) Yes, I agree too. It simply makes the most sense for an encyclopedic article. Peter K Burian (talk) 14:38, 30 April 2017 (UTC)

Trend plot in article
There has been a back and forth in regards to the sales volume plot that was in the article (removed here []). I agree with the concern regarding the way the numbers are grouped. I also liked the addition of the graph to the article.

Two questions: First, do people like the graph in general? (regardless of smoothing or the debate about how to group companies?) Second, if yes, can we discuss what we think it should look like?

As for question 1, yes I like it. As for question 2, I prefer the non-smoothed version since it's really just a year by year total plot, not continuous (month by month or finer) data. Also, I think it would be best to display the raw numbers (ie grouping based on the source data) then include notes to explain when groupings changed. Noting when a grouping changed is not OR since it only calls attention to the year by year changes in groupings. I think such a plot would be very informative. Springee (talk) 19:15, 25 April 2017 (UTC)
 * I like the graph (I was the one who added it in the first place). I think it's the most valuable piece of information in this article, because in my opinion visual information is way more potent than written information. I think we should add the numbers directly from the OICA into the graph, in accordance with option 1. IMO we should use the non-smoothed plot. When I added it, it was non-smoothed and I have even "unsmoothed" it once, but someone keeps reverting it. rayukk &#124; talk 09:33, 5 May 2017 (UTC)

Original research and original synthesis in tables
I updated the 2016 results following the OICA table. I suggest we should proceed that way from now on. Why? Various reasons, I'll try to summarise them:

a) We are putting in OICA's voice things the OICA didn't stated: We can disagree with the production numbers and the counting system of OICA. In fact, I wouldn't have a problem if someone add alternative tables using other sources. However, at this point in time we're saying OICA is our only source to establish the rankings here, and we keep on adding notes and fixing them because we don't like how they work. That's bad on at least two levels. Firstly, we're saying OICA doesn't know what it's doing, and therefore seriously questioning it as a reliable source. Secondly, we're making fixes based on information that is not present in this article and expecting the reader be aware of what we're talking about.

b) We are making original synthesis: There're a lot of tables that cherry-pick OICA reports and combine them based on the personal feeling of each editor. While some of them are understandable, they're all making assumptions on how things should be depicted and misusing OICA info.

c) We are misinforming: We shouldn't take sides for any manufacturer here, that's the reason we should be stricter on following OICA. At the moment, the 2015 production says it includes (in small letters) the production of "SAIC-GM" because, basically, OICA shouldn't have stopped counting them for GM. That's wrong at various levels. As I already said at the automotive industry and Wiki car talk pages, the joint venture that was excluded is SAIC-GM-Wuling, not SAIC-GM that it's still being counted for GM. Given that GM has only a 44% stake and SAIC is the majority owner, it seems a sensible decision to start adding SAIC-GM-Wuling to SAIC production. GM can still claim the sales (as it does) because it has a participation in that company, but we must trust OICA countings for production, especially if we want to avoid double-counts. At the moment we've inflated and misleading numbers in 2015, because we add SAIC-GM-Wuling for both GM and SAIC results. It goes without saying that we can't use those numbers to say GM is "3" as an editor insisted if we don't sustract them from SAIC, and SAIC would be gone of the ranking...

As fixing all the tables would be a very difficult task, I propose we keep them as they are (this will never be a featured list, but well...), while being more faithful to sources from now on. Any change should be based on OICA fixes, not ours. Thoughts? --Urbanoc (talk) 15:07, 17 October 2017 (UTC)


 * Totally agree. We either follow the source or we don't follow the source. Following the source but amending it is really slyly doing our own thing under the pretense of following the source (also called lying). The most you can do is to say that other (specifically listed) sources include or do not include certain groupings to get different rankings.  Stepho  talk 23:16, 18 October 2017 (UTC)


 * Thanks for your updates. I have further removed SAIC-GM numbers from the 2015 GM total in the graph, as the OICA has continued in 2016 their 2015 change of practice. — JFG talk 23:45, 3 November 2017 (UTC)

The case of SAIC-GM-Wuling
We need to understand whether production figures of SAIC-GM-Wuling (SGMW) are counted twice in OICA statistics, and whether they should be included in General Motors total production. Here is a table of relevant numbers as published by the OICA:

The production of SAIC-GM in Shanghai is always counted in GM total production. The production of SGMW is counted in GM total until 2014, then excluded in 2015 and 2016 (as noted in the article footnotes, discrepancy marked in red here). In the detailed production tables by country for GM, SGMW and SAIC-GM are conflated from 2006 to 2010, and both are counted in GM total. The production of SAIC Motor includes SGMW until 2005, then excludes it from 2006 to 2010, then includes it again from 2011. Specific production counts of SGMW are not available before 2011. Therefore we have four years 2011–2014 (marked in yellow) where SGMW production is counted both in the SAIC total and in the GM total.

What can we do with this state of affairs? GM shows an abrupt decrease in 2015 due to the removal of SGMW from group stats. We can either: None of these solutions fully respects the source, but given that the source is not providing comparable numbers year-to-year for this case, we should be able to clarify the information for our readers. — JFG talk 03:47, 4 November 2017 (UTC)
 * add SGMW vehicles again to GM total in 2015 and 2016, for consistency with prior years;
 * or subtract SGMW vehicles from GM total between 2011 and 2014, because they are already counted in SAIC Motor.
 * I can't participate in Wikipedia those days, but my opinion is to follow OICA whenever possible, even if it seems wrong at first sight. If that leads to inconsistencies, so be it. The alternative is to admit OICA is a shaky source and therefore we shouldn't be basing a whole article on it. I saw the opposition to OICA is widespread, that's the reason I think alternative rankings (if there are out there) should be introduced. I'm against a ranking or production numbers made by a consensus of editors against OICA, but I'll accept it if that's the case. Urbanoc (talk) 19:11, 4 November 2017 (UTC)

2018 data
I recommend the removal of the 2018 data. Reasons:


 * OICA has not published 2018 data. At the time of this writing, OICA has not even published production by manufacturer data for 2017


 * At the time of this writing, Japanese OEMs have not published their 2018 production data. They will be published tomorrow, Jan 30,2019. Renault so far has not published production data for 2018, only sales data. Any "production" data published at this time for Toyota, the Renault-Nissan-Mitsubishi Alliance, Honda, and Suzuki can only be false.
 * Volkswagen has not published production data yet. The 10.8 million in the 2018 list reflect "deliveries"


 * GM has not published global production data for 2018 at the time of this writing.
 * Ford AFAIK has not published global 2018 production at this time. Its published 2018 "wholesales" are 5,982,000 - the number does not agree with the list.

The source given in the article therefore cannot possibly be reliable. Its data can only be rough estimates. Also, according to the source, it does not reflect production, but sales. I recommend waiting until OICA data are available.BsBsBs (talk) 09:37, 29 January 2019 (UTC)
 * Your proposal sounds reasonable to me. I was worried about the different groupings as well as the different source, and difficulty in comparisons with previous years. -- DeFacto (talk). 10:30, 29 January 2019 (UTC)


 * Thank you. If anyone needs help, please contact me. Been counting cars all my life. Well, not quite. BsBsBs (talk) 14:30, 29 January 2019 (UTC)

Data 2020
There are different sources with different numbers. In this article Toyota has 7,2 millions and VW 9,33 according to Focus2move. In the articleToyota other source with 9,528 for Toyota and 9,305 for VW. It is small difference for VW, less than 30 thousands, but for Toyota. Which number is correct?--Divega (talk) 05:12, 26 October 2021 (UTC)


 * Many different reasons. Some sources count only passenger cars, while others also count light trucks. Some sources count Jan-Dec (calendar year), some Mar-Apr (US financial year), some Jul-Jun (other financial year), some Nov-Oct (US model year). Some count subsidiaries (eg Toyota + Daihatsu), some don't. Some count knock down kits assembled by other companies, some don't. And so on. All these numbers are correct - even when they differ. We used to use OCIA exclusively but they stopped publishing free data.  Stepho  talk 08:56, 26 October 2021 (UTC)

2020 original research list
The 2020 "by manufacturers" list was a hot mess, so I just got rid of it. At the moment, we only have what the manufacturers say they produce, so we can't be sure we aren't introducing double counts (OICA tried to reduce that when doing its ranking). Besides, only European and East Asian manufacturers consistently report production. The rest (including Americans) mostly don't publicly report production but just sales (the point that most concern shareholders and the government...), so we'll have a tremendously inaccurate data set if we try to put together a production ranking by ourselves. --Urbanoc (talk) 17:16, 26 January 2022 (UTC)

Statistics by manufacturer
Where is the statistics of vehicle production by manufacturer for the years 2018, 2019, 2020 and 2021? Has OICA stopped publishing statistics by manufacturer after 2017? 212.100.101.104 (talk) 15:26, 31 January 2023 (UTC)


 * Correct. OICA was perfect for what we wanted but they stopped publishing them publicly. We have not found any other source that compares the entire industry in a single table. We certainly can't use each manufacturer's data because they have different definitions (includes heavy trucks or not, includes pickup trucks or not, includes motorcycles or not, includes subsidiaries or not, includes localised assembly from knock-down kits or not, etc).  Stepho  talk 00:41, 1 February 2023 (UTC)

Update badly needed
The last year here is 2017, 6 years ago. Those data are obsolete, except for historical interest. Zaslav (talk) 04:08, 16 September 2023 (UTC)


 * Sure. But we need consistent references. We used to use OICA but they stopped publicly publishing them in 2017. It's not feasible to use the manufacturers own numbers because they are not consistent - ie some count passenger vehicles only while others count trucks too, some count subsidiaries as one huge entity while others keep them as smaller separate entities, some count joint ventures while others don't, etc, etc. If you know of consistent references then we'd love to hear of them.  Stepho  talk 09:05, 16 September 2023 (UTC)

2006 data (again)
212.100.101.104 and I have a disagreement. The anon-IP believes that OICA has grouped the companies mistakenly and has provided an alternative grouping which changes the total productions figures and the ordering. I believe that we follow the reference (as per WP:STICKTOSOURCES) and that regrouping violates WP:OR and WP:SYNTH.  Stepho  talk 04:47, 4 November 2023 (UTC)
 * I agree, and, as I recall saying before, if we consider the OICA source so bad we need to somehow "fix" its rankings, we shouldn't be using it in the first place, personal synthesis to that level is just source misrepresentation. --Urbanoc (talk) 10:00, 4 November 2023 (UTC)
 * I think this "disagreement" is finished now. I have found a new source which lists Ford as the 3rd and Volkswagen as the 4th largest automaker in 2006. This source also proves that Ford actually had 6.268 million produced vehicles rather than 6.5 million in 2006. The former source has both claimed Ford's production as 6.5 million and mistakenly ranked Ford after Volkswagen, despite Volkswagen's production which was 5.6 million and thus smaller than Ford's. Besides this, I think Toyota should be ranked as the largest, rather than 2nd largest, as Toyota (including Daihatsu) had more produced vehicles a year than G.M. for the first time in 2006. However, OICA listed Daihatsu separately from Toyota 2001 to 2007, as well as Mazda separately from Ford despite Ford's 33.4 controlling interest in Mazda 1996 to 2008.
 * If OICA claims that Ford never owned Mazda, then why did a 1999 report claim that a study in 1999 showed that Ford (including Mazda, Lincoln, Mercury, Aston Martin, Jaguar-Land Rover and Volvo) was expected to surpass General Motors in mid-2000s? According to this study, Ford was estimated to increase its production from around 7.75 million vehicles a year by 1999 to 9.15 million by 2005. However, OICA shows that Ford (excluding Mazda) only produced 6.5 million in 1999, and thus around 1 million less than this study which estimated 7.75 million, of which 1 million produced by Mazda. 212.100.101.104 (talk) 16:11, 7 November 2023 (UTC)