Reframing imperial China’s indigenous accounting history: further discoveries in archival materials from the three centuries before 1850

Abstract

How far did the indigenous accounting of China's historically successful economy parallel Western double-entry bookkeeping (DEB)? We propose a scheme for classifying stages of bookkeeping that approach full DEB, review recently available nineteenth century Chinese accounting manuals and re-examine how far their recommendations reflect practice to be found in original account books contained in the archives of the Zigong brine wells for 1916-1917 (which have been argued to be essentially unchanged from the nineteenth century Qing era and perhaps earlier) and in the surviving accounts of the Fēngshèngtài salt traders of Henan province spanning 1854-1881. We introduce the accounting records we have now discovered from merchanting businesses in Anhui province, which span 300 years and survive from the 1590s onwards. These are all more sophisticated than the ‘merchant-banking' accounts in the vast archive of Tŏng Tài Shēng covering 1798-1850, and in the case of the Anhui merchants' accounts comprise ‘balance sheets' that include monetary values for physical as well as monetary assets, matching their owners' ‘capital'. We tentatively conclude, on the basis of the evidence now emerging, that despite its variety of forms indigenous Chinese style accounting practice may in some cases have captured the structural essentials of DEB’s content and functions and might be labelled ‘Chinese-style double-entry bookkeeping' ('CDEB'), over which Western bookkeeping had no conceptual advantages.

Keywords:

• comparative international accounting history
• Chinese accounting archives from Ming to late Qing era
• Chinese business history
• double-entry bookkeeping (DEB)
• Chinese-style double-entry bookkeeping (CDEB)

Acknowledgements

We are grateful to the administrators of the Sichuan Province archive in Zigong and in particular to Professor Chen Longgang and Mr Jiang Ming of Zigong Social Science Union for access to the accounting records held there. Also to Professor Liu Bo San, Gong Da, Dr. Guo Yanwei and Yuhua Chen-Hudson for assistance with the records from Anhui province. We also thank Professor Fu Lei and particularly Professor Song Xiaoming and his staff at the China Accounting Museum, Songjiang, Shanghai, for their advice and assistance with archival materials; and Professor David Faure for alerting us to accounting manuals. We are also grateful for the comments of participants and discussants at the various conferences and seminars in the UK, US, continental Europe and China where we have made presentations of earlier versions of the paper, including the AAA Annual Meeting, August 2019, San Francisco, CA. In particular we thank Prof. Kees Camfferman, Dr. Nadia Matringe and Dr Zhang Wen for their helpful comments as well as this journal’s editor and reviewer. Professor Yuan Weipeng acknowledges financial support from the Yongyou Foundation and the Chinese Social Sciences Foundation (21&ZD078).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Except for Canton (now Guangzhou).

2 Our emphasis added.

3 Apart from some of the materials to be discussed here, of the original archival materials listed in Appendix II only TTS and the Shanxi banks (shānxī piàohào [山西票号]) can definitely be dated before 1850. (For discussion see Yuan et al. 2017, Wu 2016.)

4 They noted that it would be valuable if future research in other archives as they are discovered could reveal contracts, business correspondence, minutes of meetings etc. that can be linked to business accounts to shed more light on issues such as profit sharing, accounting for labour, and implications of taxation (c.f. Auyeung et al. 2005, Zelin 2009).

5 Unlike most previous authors, we have standardized the presentation of Chinese technical terms and their translation, and of individual names, wherever possible, using the format: pīnyīn ([拼音] = ‘standard modern Romanized spelling with tone marks’) or equivalent, at least at the first time they appear, as an aid to reading by those not familiar with Chinese. We also transcribe the original texts into the modern, simplified characters now standard in PRC.

6 (Present authors’ note) A visit to the National Library revealed that a further 28 of the missing books have now been found in the National Library and await cataloguing and digitizing. This leaves 10 unaccounted for (together with some that may have been deposited in addition to the original gift). We have not been able to revisit the Library following the Covid-19 pandemic.

7 (Present authors’ note) The ‘three levels’ refer to day books / classified ledgers / financial statements of profit and loss and net assets (as set out in Yuan et al. 2017).

8 (Present authors’ note) In Yuan et al. 2017 ‘CDEB’ (‘Chinese double-entry bookkeeping’) was used as the label for the systems claimed by previous authors to be equivalent to Western DEB. For the reasons given below about how DEB should be defined we now distinguish ‘doubled entry’ and ‘double-entry’ bookkeeping (c.f. Sangster 2016) and will argue that some indigenous Chinese systems did probably attain the essential structural characteristic of DEB identified by Mattessich 2000 and discussed below (albeit not the Western DEB form of processing), so that we can now legitimately label them as ‘Chinese-style double-entry bookkeeping’ (CDEB).

9 (Present authors’ note) A high level of bookkeeping sophistication was reached in the maintenance of the accounts of customers of the moneylending side of the business (see Figure 4 at p.413 in Yuan et al. 2017).

10 Giersch (2014, 2020) argues that by the end of the 19th century there was increasing sophistication of accounting and demand for skilled accountants in Yunnan province. We have seen the day-book from 1859 that he refers to (2014, p. 56) but this is only a simple record of transactions.

11 Professor Guo Daoyang has argued that adoption of the ‘four column’ (or ‘four pillar’) method in government accounting during the Tang and Song dynasties (618–907AD; 960–1279AD) represented an important change in the Chinese conceptualisation of what the accounting should be ‘naming and counting’ and for what purpose (c.f. Ezzamel and Hoskin 2002). As he has indicated in conversation (November 2011), initially each period’s account had been regarded conceptually as strictly separate and so had just the ‘Three Columns’ of ‘in’, ‘out’ and closing balance: opening balances had to be confirmed as required from the preceding period’s books. However problems of loss and theft of records led government accountants to recognise the need for continuous records and to incorporate the closing balance, that strictly belonged to the previous period’s account, as the opening balance in the new period’s books, thereby creating the ‘Four Column’ system. (If there had been any commercial accounting books then one might have expected them to need ‘four column’ accounts to keep track of debts due from period to period, as illustrated in Picture 4 of Yuan et al. 2017).

12 Toms and Fleischman (2015) offer a theory of accounting change that they label the ‘Accounting Transaction-Based View’ (‘ATBV’). However in their taxonomy of developments (Table 4) they place Western DEB as the first stage, sufficient for ‘arms-length transactions between merchants’. Clearly this does not hold for Asia (and in fact not for Europe nor the US) – c.f. Macve 1996, Hoskin and Macve 2021.

13 Cash holdings may be regarded as either a physical quantity measure (specie / bullion) or a monetized measure according to the context.

14 An anonymous conference reviewer has pointed out that Robert Sterling indicated that other factors such as temperature and atmospheric pressure might affect volume and these also need to be assumed to be held constant.

15 Recognising that Chinese nouns do not distinguish ‘singular’ and ‘plural’ forms it is possible to translate e.g. zhàng [账] as either ‘account’ or ‘accounts’.

16 For an introduction to the history behind the labels that have been given to various Chinese bookkeeping systems see Section 2 above and Hoskin and Macve (2012) Appendix III.

17 We use ‘Three Feet’ (and correspondingly ‘Four Feet’ for the full CDEB system) as the translation for this bookkeeping system, to avoid confusion with the ‘Three Column’ and ‘Four Column’ methods for calculating the balances on individual accounts.

18 Lowenstein and Cao (2023) have identified doubling both of credit and of cash transactions in ‘trading’ accounts at C19th Fēngshèngtài, so these accounts reach beyond the threshold of stage 2. And although no ‘balance sheets’ have survived they argue (pp. 31-3) that circumstantial evidence from that period suggests that the accounts may have reached stage 3(ii). See Section 7.1 below.

19 The breakthrough of attaching monetary values to non-monetary assets opened up an ever expanding range of what could be ‘named and counted’. So now these include intangible assets and even ‘the present value of future profits’ (as ‘goodwill’). See e.g. Bromwich et al. (2010).

20 For discussion of the description that has been labelled lóngmén zhàng ([龙门账]  = ‘Dragon Gate’ (or ‘Joining-up the Embankment’) bookkeeping), also known as tiān dì hé （[天地合] = ‘Heaven and Earth Matching’), see Section 6 on Zigong below and Hoskin and Macve (2012) who argue that the various names were generally given later and do not refer to indigenous methods for which there is any historical evidence.

21 Historical examples of Western DEB may also utilise other accounts for gains and losses additional to those included in the Profit and Loss account and also close them to equity capital (e.g. de Roover 1956, Goldthwaite 2015). Modern IASB standards also split the total of a period’s reported income (labelled ‘comprehensive income’) into ‘Profit and Loss’ and ‘other comprehensive income’ (see e.g. Macve 2022, Appendix A). Such approaches can also be classified as 3(ii) (in China) or as 4 (in the West).

22 Lane (1977) reviews the range of scholarly views on how extensive and sophisticated the recording of doubled (or ‘dual’) entries had to be historically in Renaissance Venice, Florence and other Italian cities to constitute a double-entry system. For our purposes the distinguishing feature of double-entry (as in our stages 3(ii) and 4) is the full integration of the accounts for ‘flows’ and ‘stocks’ which constitutes Mattesich’s (2000) ‘closed system’. This is the system which it has been claimed is found in China as lóngmén zhàng ([龙门账] but c.f. Hoskin and Macve (2012). ‘Full DEB’ (our stage 4) also features the full cross-referencing of entries in paginated and indexed books as recommended by Pacoli.

23 (Present authors’ note) For modern economic analyses of ‘capital’ and ‘income’ see e.g. Bromwich et al. (2010).

24 The character 帐 for zhàng (now meaning ‘curtain/tent’) is used instead of the normal PRC simplified character 账 for zhàng (‘account’), which is used by the actual 18^th / 19th century TTS records (Yuan et al. 2017), and which incorporates the ‘phonetic derivative’ of 贝 for bèi (‘cowrie [shell]’) – i.e. the original medium of currency in antiquity – in order to provide the ‘radical’, i.e. the representation of the meaning. In discussion with Professor Guo Daoyang, he has explained that in earlier times帐 was the character used as its incorporation of the sign for ‘tent’ is believed to represent the tents of the Emperors’ travelling tax collectors.

25 In a DEB system (as also in stage 3(ii) above) some elements may be computed only at a reporting period’s end and entered as ‘closing adjustments’, e.g. depreciation, bad debts, impairments and other provisions and (depending on the system in use) closing inventory of unsold goods and production work-in-progress. (See any modern introductory financial accounting textbook.)

26 As shown in Yuan et al. (2017), TTS does not meet these criteria and we would classify it at stage 2(i) above. Given that none of the archives we examine below include recording of industrial assets (such as the brine wells at Zigong) the Chinese accounts did not include all the contents of modern DEB accounts.

27 The medieval evolution of such indexing and ordering of books, beginning in European universities with the Bible, is traced in Hoskin and Macve 1986. Sangster (2016) ascribes its adoption in commercial account books to Florentine legal requirements but without explaining why courts would have demanded such full DEB books in evidence.

28 However, Chinese account entries were dated which would reduce effort in searching for matches.

29 We are grateful to Professor David Faure for providing us with a copy of this handwritten manual. It is now published (in Chinese) in交易须知, 2013 (Volume 1, pages 17-44). We are grateful to Dr Yimin Zhao for assistance with the translation.

30 However there are internal references in the numerical examples to dates such as Xianfeng 11 ( =  1861) and ‘some year in Xianfeng’s reign’ (mŏunián [某年]) (i.e. between 1850 and 1861). Also (at page 4 line 3) to April 1, Tongzhi 1 (同治元年) = 1862). So, perhaps the 1828 edition had also been updated in some other respects before the 1891 publication, possibly reflecting new Western influences.

31 The Chinese words for buy (mǎi [买]) and sell (mài [卖]) look similar and are pronounced almost the same (in some regions identically, i.e. with the same tone) so there is risk of confusion. (More confusion can come from the fact that in Chinese ‘borrow’ and ‘lend’ are both jiè [借].)

32 Where customers prefer to record short-term transactions ‘on the slate’, i.e. just on erasable boards (shuǐpái [水牌]) an account of ‘temporary deposits’ should be kept. (See also 5.2 below: instruction #1.)

33 ‘The cooked meat is in the wok’ (ròu làn zài guō lĭ [肉烂在锅里]) (i.e. it is obvious how much one has got) and ‘What is the meaning to record books clearly if you cannot earn money?’

34 Along with the shareholder book (rùběnzhàng [入#x672C;账]) and the profit and loss account (wànjīnzhàng [万金账] – see notes to Figure 1.

35 The copy was found among a collection of business books relating to Shanxi merchants. Its original source can no longer be found (and is not in the National Library). It is held in Shanxi and has been photographed and published in 2018 and been discussed (in Chinese) and rewritten into simplified Chinese by Professor Wang Lingling and Wang Zhongliang in清代晋商典当业会计账簿组织探析ISBN: 1004-0994（2018）03-0132-6. For some reason that is not clear they date it to 1895.

36 Closing gāi ([该] = liabilities) are not specifically mentioned but the use of accrual for interest due (see below) implies that the closing position includes liabilities.

37 The numbering is ours and not in the original.

38 However some pages are missing. There must also have been some system for issuing tickets to customers for their pawned goods and recording where they were stored. Many such tickets survive from the Ming-Qing era, with templates often created by wood-block printing. However this manual does not mention them,

39 Surviving examples are common and may be written on black or white wood or on tin (similar to a Western chalkboard or ‘slate’).

40 See Yuan et al. (2017), Appendix 2.

41 As the assets dealt with in the Hebei manual are only the debts created in a pawnshop business it, like the Shanxi piàohào, is strictly only at stage 2(ii). See Appendix II.

42 https://chinadialogue.net/en/energy/9918-the-historical-roots-of-china-s-industrial-revolution-2/ (accessed 3.3. 2023).

43 For comparison with practices in British coal-mines in the 18th and 19th centuries see e.g. Fleischman and Macve 2002.

44 Or ‘joining the two ends of the embankment’ (c.f. Guo et al. 2011: see Hoskin and Macve 2012, Appendix III for further discussion of the etymology and sources of this 20th century labelling). The Zigong accounts themselves do not use this term.

45 Rongchang is a county in the Chongqing municipal area (which was formerly part of Sichuan province) and there is still an industry there with this name.

46 As demonstrated in the live reconstructions of the mines’ operations staged at the Zigong Salt Museum.

47 However a separate account for cash was not strictly needed for the accounting reconciliations if one balanced all the items of income and outgo. Surviving early Italian examples of DEB do not always include a cash book (de Roover 1956) or clearly kept it outside the general ledger (Goldthwaite 2015, p.627).

48 https://www.travelchinaguide.com/attraction/sichuan/chengdu/salt-museum.htm (accessed 10.3.2023).

49 However, Wright (1998) argues that formal dissemination of scientific knowledge (e.g. from the dockyards) through translations of Western texts remained very limited until after 1895.

50 (Present authors’note) They are referring to a 2019 working draft of the framework we have set out here.

51 The authors label this level 3(iii) in accordance with the classification previously suggested in that 2019 working paper.

52 Qianlong reigned officially from 11 October 1735 to 8 February 1796 (and actually until 1799).

53 Daoguang reigned from 1821 to 1850.

54 Guangxu reigned officially from 1875 to 1908.

55 They are published in Liu Boshan (2005) [in Chinese].

56 Xianfeng reigned from 1850 to 1861.

57 Tongzhi reigned officially from 1861 to 1875.

58 A Western contract might specify ‘DEB’ or utilise an earlier equivalent label (e.g. Oldcastle’s 1588 ‘hovv to keepe bookes of accompts after the order of debitor and creditor’. https://www.icaew.com/library/library-collection/historical-accounting-literature/highlights-of-the-collection/a-briefe-instruction (accessed 11.3.23). Lane (1977) observes the use of standardised accounting language in Renaissance Italy and de Roover (1956) argues that partnership provided the primary motive there for using DEB.

59 From a modern perspective there is an implied calculation here of the ‘return on investment’ (‘ROI’) [or ‘return on capital employed’ (‘ROCE’)] which, despite Bryer’s extensive ahistorical claims, was not explicitly deployed as a measure of ‘capitalist’ profitability in the West until the later 19th century (e.g. Toms and Fleischman 2015, fn.26).

60 One year another name appears in both the assets (cún) and the liabilities (gāi). It is not clear whether this was a way of correcting an error by a contra entry or whether there were in fact two different businesses involved.

61 So the kinds of adjustments which are beyond the scope of the supposed DEB ‘axioms’ that Sangster (2018) claims underpin Pacioli’s exposition of DEB were being made here, as they were by Italian merchants using DEB before Pacioli’s treatise (de Roover 1956, Kuter et al. 2018, Macve 2022).

62 About 240 km / 150 miles from Huizhou.

63 The books are stored in the Library of Sichuan Normal University.

64 i.e. in anticipation of their annual bonus.

65 20%pa is also used as the standard rate on capital for rewarding shareholders/partners.

66 This profit smoothing device would help fairer allocation across changes in shareholders etc..

67 Published in Zhou Shaoquan and Wang Juexin, 1991 [in Chinese].

68 The oldest example in the China Accounting Museum in Songjiang, Shanghai, is a pawnbroker’s loan account book from the reign of the Ming emperor Xuande dated Xuande 3 (1428) but no related account books survive to show what the overall system was. An older set of pawnshop records from Dunhuang, published in Tang Chang-Yu (1981−1991) (in Chinese), dates from the Tang dynasty (7th to 10th centuries AD) but contains only transaction day-books together with very detailed information about the individual customers. A series of records for a family clan in 祁门县 (qímén xiàn) county of Anhui province with two branches, including rents received, market sales of grain, and expenses of their extensive landholdings from 1575 (Wanli 3) to 1661 (Shunzhi 18), includes an inventory of the lands (and their rentals) and forests owned together with the tax payable (updated for changes) and regulations for the keeping and checking of the daily accounting records and for preparation of annual summaries of these, which would be open to all the family on the annual ‘ancestor worship’ day (now ‘tomb sweeping day’ (qīngmíngjié [清明节]) and then one copy stored in safekeeping. The forest manager’s accounts include numbers of trees planted and felled and receipts from sales of timber. This series.has been published (in Chinese) in 窦山公家议校注 by 黄山书社 in Hefei in 1993: ISBN 7-80535-258-5/Kˑ75.

69 The annual accounts were then called chá suàn ([查 算] = ‘checking and calculating’), i.e. not yet the later widely used hóng zhàng ([红账] = ‘red (or bonus) account’) (e.g. Guo et al. 2011).

70 For example: ‘To be successful in business, you had better calculate and plan clearly in advance and control the expenditure according to the income.’; ‘To choose the manager, try to select one who is both honest and talented. It will be a benefit to the company even though you have to pay him more. You should check the accounting books by yourself very carefully and frequently.’; ‘Recording every expenditure and income immediately so that you will not forget or make any errors.’; ‘Spend your money very carefully and be diligent on doing the bookkeeping. When a transaction happened, record it immediately.’; ‘Choose a partner like making a marriage and when dealing with any transaction you should keep it in your account books clearly.’ (See e.g. Cheng Chunyu (2019), ‘Important rules and tricks of a sound businessman’ (Ming Dynasty, 1626) and Wu Zhongfu (2019), ‘An Overview of Merchants rules and tricks’ (Qing Dynasty, 1792) [both in Chinese].

71 As noted, Yuan et al. (2017) argued that the abacus should be seen as an integral part of the Chinese bookkeeping ‘system’ as it was easy to calculate amounts owed at any time and continuous balancing within the written records (as in Western DEB) was not necessary. Sūzhōu măzì numerals – which were used in all the accounts referred to in this paper – accelerated the speed of such calculations. (See Yuan et al. (2017) Appendices 2 and 3; Lowenstein and Cao (2023), 17–18.)

72 The examples in Appendix II here that refer to these systems all date from the 20th century.

73 See also Ogura (1982) for traditional Japanese practice.

74 Further comparison with records from Japan and Korea (e.g. Previts et al. 2011, Zhang 2018, 2019) will also be valuable.

75 It was announced at The Silk Road International Conference on Accounting History and Culture: While East Meets West, held in Shanghai, PRC at SLUAF and the China Accounting Museum, 25–26 November 2017, that an English translation of Prof. Guo Daoyang’s latest publication was planned for publication by 2024.

76 https://www.meet-in-shanghai.net/travel-city/tourist-attraction/tourist-attraction-detail.php?id=4628 (accessed 10.3.23)

77 We are grateful to Danting Ni and Kees Camfferman for help with the transcription and translation.

78 dé jì probably reflects the business practice of continuing to use an established business name after the ownership had changed, with a note of the current owner’s name.

79 People’s / business names.

80 Records here for which sources have not been cited in this paper are accessible through the China Accounting Museum at Shanghai Lixin University of Accounting and Finance in Songjiang, Shanghai (https://www.meet-in-shanghai.net/travel-city/tourist-attraction/tourist-attraction-detail.php?id=4628 accessed 10.3.23). We are grateful to Professor Xiaoming Song, its former Director, for exploring them with us. In the case of the 1935 book titled龙门, this is held at the China Finance and Taxation Museum, Hangzhou.

81 Possibly 3(ii) but there are no surviving ‘balance sheets’. See Section 7.1.

82 When the famous Rìshēngchāng bank was founded (e.g. Morck and Yang 2010).

83 Probably 3(ii) given the date, but there are no surviving transaction records.

84 Perhaps 3(ii) but there are no surviving transaction records.

85 A Japanese investigator in 1941 refers to lóngmén jiécăi ([龙门结彩] = (literally) ‘festooning the dragon gate’) in a business in Wuxi in Jiangsu province but notes that the 总情簿 (zǒng qíng bù = summary information book) is not the same as a DEB general ledger as there is no systematic ordering (there are of course no page numbers) and various items of goods/commodities and receivables/payables of agents are recorded in different books. The process described for preparing the final accounts is essentially that of our stage 3(i). Professor Lina Xu (xŭ zǐfēn [许紫芬]) has translated this material into Chinese.

86 Professor Lina Xu (xŭ zǐfēn [许紫芬]) has examined this material.

87 Gardella 1992.

88 Fang, Zhao (2009). The Emergence of Early Modern Enterprises in China: A Case Study of Hanyeping Coal and Iron Company, Limited. Haverford College Thesis http://hdl.handle.net/10066/13642 (accessed. 21.3.2023)

89 However, as the assets dealt with in the Hebei manual are only the debts created in a pawnshop business it, like the Shanxi piàohào, is strictly only at stage 2(ii).

Additional information

Funding

This work was supported by Chinese Social Sciences Foundation: [Grant Number 21&ZD078]; Yongyou Foundation.