Relational Accounting

beyond the paradigm of capitalism

Pieter Wisse

The world is changing. Many people in what is called the western world believe that their neighbours in the east want to implement its economic system without any reservations. This should undoubtedly include our method of accounting. Shouldn't it? The wisdom of such uncritical copying is questionable, for all parties concerned. Why don't changes stimulate critical investigation of some of our own western traditions? Again, methods of bookkeeping and accounting included. This article outlines an alternative.1

The limited scope of this presentation makes more than an incomplete introduction inescapable. In view of this, I conclude with similarly introductory remarks on the problematic character of the phenomenon of information. It seems that, especially where bookkeeping and financial accounting are concerned, positivism still predominates. However, different insight and behavior may result from interpreting information as playful elements in a hermeneutic circle. I hope the necessary recursivity and irony, its close companion, are not lost on the philosophically oriented reader.2



Bookkeeping and accounting have a long history. Recording of transactions, however, has always been the fundamental attribute of accounting. Of course, the focus of attention was and is the financial aspect of such transactions.

The quantity jump in entrepreneurial services and production is well documented. It took place in Europe at the end of the Middle Ages. The business organization came into being as a seperate entity. It took part and started to dictate many transactions in society. Ever more quantites of money circulated and were accumulated in organizational domains.

A capitalistic or business organization has assets and equity. By definition these are considered equal. Thus rules the accounting equation. Moreover, both assets and equity will be subdivided. On the same side of the equation increase in one part always equals decrease in another. The method of double-entry accounting assures conservation of these basic definitions. Double means that the entry-as-record reflects both, opposite consequences of the financial transaction under consideration.

Within the capitalistic economic order, entrepreneurial services and production are implied. In western countries all kinds of closely related phenomena have now developed beyond questioning.3 Some centuries of praxis have also turned double-entry accounting into such a truth. I contest it here. Uncritical acceptance of traditional double-entry accounting obstructs many promising opportunities as I see them. An alternative rests on another paradigm. It also shows records as double entries, but with a much restricted scope. It seems paradoxical, but flexibility of accounting is radically enhanced through this very move. It's that smaller building blocks allow for greater flexibility, as every good designer knows.


the exception of government

Relational accounting will be easier to understand after I've sketched the situation in which it originated. It started with a ministry in the Netherlands government. Double-entry accounting as used in the private sector had not been introduced, for good reasons actually. As every part of central government, the Ministry had to comply with accounting directives and regulations that certainly were, and still are, exceptional when looked at from the point of view of business. The light of history illuminates that an old, really, a pre-capitalistic accounting method is still applied. It is similar to the predominantly German Kameralistik. Or to the American fund accounting. My impression of the most notable property of such methods is as follows: the fund balance is kept as a separate account; through the activity of calculating this at the moment of recording, the balance for the remainder of the budgettary period is immediately accessible.

Financial management with the Ministry was, however, not so straightforward. Expenditures did not directly spend a budget. The most important processes had up to three intermediairy stages — or phases -, including formal decision making with external parties; these extra phases stood for as many types of commitments. Not surprisingly, mixed bookings to keep current all kinds of balances on different accounts resulted in a mess. The records were beyond control. Financial accounting was critically inadequate, and with it the internal and external management as a whole at the Ministry.

After thus identifying the causes of complexity, I tried to model accounting differently. In, say, a surge of phenomenological focussing, I composed a separate ledger for each of the recognized phases in an organizational process. And in what I practiced with entries I did nt keep a balance with respect to an earlier phase. Instead I limited every double entry to the ledger or register for the phase to which I unequivocally assigned the transaction which record was made. This way, every register had its private accounting equation. Being restricted to a separate phase, it took on a very different meaning from that for a whole process.

Of course, at first glance I seemed more removed from managable and adequate accounting. For the solitary registers obstructed collection of (information on) transactions of different phases. How could they be brought together in a comprehensive report? I had scrapped all accounts that were meant, as a result of the recording activity, to keep a balance between phases.

The problem was easily fixed. A reference could be supplied between accounts in different registers. This mechanism secured necessary reports about the financial development of the organizational process. The balance available for further decision making and pertinent spending during a certain process phase is now calculated. References show the way for compiling such reports.4 I named this alternative method: relational accounting. The name derives from those references or relationships that connect accounts. These relationships are the method's fundamental characteristic.


cash-basis accounting

The explicit relationships between accounts in different registers add marked flexibility to accounting. A new register, for example, is relatively easy to introduce and an existing one may be removed. Relevant relationships may be broken or established as a matter of quite simple maintenance. Other than that, accounting, i.e. on a relational base, can continue as before. It means a greater guarantee for accounting to stay in tune with the overall situation in which it should always be a tool for financial management.

Figure 1 shows the original implementation of relational accounting at the Ministry. Sketches like this privide a good overview because a separate register for a process phase is easy to visualize. The symbol I use is a cube. It can still remind people of a ledger filled with cards for accounts. A specific letter on the face of the cube indicates the appropriate phase, or whatever we choose to call a part of a larger process. Separate accounts in every register do not show. The same goes for their relationships. In my sketches these are summarized as relationships between registers as wholes.

The Ministry was under directives for so called cash-basis accounting. Simply put, this means allocation of a cash budget (H) for a certain organizational process; for a certain period, i.e. a calender year, cash payments (expenditures — X) are approved up to the amount of the cash budget.5 In the processes at the Ministry, the phases between cash budget and actual payments/expenditures were taken up by cash estimates of three kinds of commitments. According to the official sequence they were estimates of (program) allocations (G), (project) agreements (J) and (delivery) contracts (E). This took the total of different phases and, by direct implication, registers to five. Reports about development of the processes are possible, as I said before, through the mechanism of relationships between relevant accounts in those registers. With respect to recording, every register in itself is self-contained with its own, modified accounting equation. A relationship between two registers means that transactions of a later phase spend, to coin a word, those of an earlier phase.

figure 1


changing directives and regulations

A second ministry took an interest in relational accounting. Meanwhile, a proposal for different accounting directives for government organizations had been drawn up. A draft was sent to parliament. It was also ciculated to all ministries for comments. The new directives amounted to a departure from strict cash-basis accounting. Quite rightly, the moment of actual money transfer was considered too late for active process control. Additional focus is required on the moment where the organization formally commits itself to a financial relationship with a third party. It follows, that besides a cash budget, a budget for commitments (T) is assigned. The period in which this budget may be spent by actual commitments (M) is also stipulated as a calendar year. However, a classification of commitments, such as that used by the first Ministry, was not to be implemented.

Modeling accounting after these new directives is again quite simple. This is the advantage of separate registers as supposed by relational accounting. But, before proceeding, the integration of cash-basis and commitment accounting requires an additional concept. I call it dimension. This is a characteristic management outlook. And as a general concept, in its generality to be compared with register, it offers broad flexibility. Assigning meaning yields cash-year as a specimen of dimension. Commitment-year is another dimension.

Applying this concept, every register has a primary dimension. It can be identified from the figures as a straight track from left to right. The position of a register on a certain track marks its primary dimension. Additional dimensions appear through relationships that connect a register from right to left with one or more others. So, the primary dimension of the cash budget is the cash-dimension. According to the new directives for integration, a cash budget should also contain a commitment-dimension. Otherwise, cash budgets for different periods cannot be traced as parts/slices derived from a commitment budget. The relationship between both registers involved shows this special characteristic. It does not refer to straight spending, but to a derivation without changing the overall, absolute amount. Looking along the same lines at figure 2, the commitment-dimension is also additional for cash estimates of commitments and for actual expenditures.

figure 2

A one-dimensional model strongly resembling figure 1 restricts itself to the series of registers H, M and X. With register X for expenditures in the track reserved for the commitment-dimension, that dimension would now be primary. It would also be its only dimension. Such a model shows a strict application of commitment accounting. A similar model is possible for irreducible accrual accounting etc. In all such models, the moment of actual expediture is not recorded. This follows because the cash-dimension is lacking, even for register X. In practical accounting this clearly is undesirable. For expenditures the cash-dimension should actually always be present nd primary. In figure 2, therefore, register X for expenditures has a position in the track of the cash-dimension. The commitment-dimension is additional. Relationships along this dimension secure that expenditures may be compared to commitments on the basis of a commitment-year. Along the cash-dimension expenditures may be compared to the cash estimates as slices of those commitments.



I emphasize that figure 2 also shows just an example model. Dimensions and registers may be appointed as needed. Of course, application of relational accounting remains managable with moderation in numbers. The essential point I want to make, however, is that of increased flexibility. And from that characteristic I arrived at challenging proposals for accounting directives and regulations for (Netherlands) goverment organizations.6

I postulate that dimensions can almost always be arranged in a hierarchy, based on priorities. Assignment of a budget should, in my view, include acknowledgement of a single highest dimension. As a rule, other dimensions are subordinate. Exceptional occasions excluded, the organization has discretionary power to shift derived budgets, i.e. results of slicing or estimating along lower dimensions, between periods. As long as the budget with the highest dimension can stay valid, no political, or in general, no external decision making is required.

The new directives have not yet taken into account my challenge for a pluralistic instead of a merely dualistic order. This is understandable as the wheels of government turn slowly. Therefore, I will just apply the terminology of relational accounting to illuminate shortcomings of the latter. It can simply be stated that dualism as officially proposed for the Netherlands government restricts the dimension of a single commitment-year as the highest. As a consequence, dualism does not reflect the realities of many processes. Especially in case of an investment process or major project, there should be an overall budget (B) that is more or less independent from details like subsequent years in which actual commitments and expenditures occur. Usually, there is even much uncertainty in such estimates in the early stages of such processes. And details like that do not really matter so much that management should be biased accordingly. The process as a whole is most important. Therefore, for such an organizational process, the highest dimension should be more or less a-temporal. A model of adequate accounting now speaks almost for itself.

figure 3

A logical conclusion of pluralism is the redundancy of the dimension of commitment-year. It follows from the discretionary power for shifting estimates-as-slices at lower dimensions. But as I said, the new directives for the Netherlands central government organizations do not recognize pluralism and all implications beyond dualism.7 The result of skipping the dimension of commitment-year is shown in figure 4. Originating from an a-temporal budget, and with a different primary dimension, the character of commitments has slightly been altered. This is reflected by changing their name; on the face of this particular cube in figure 4 a letter C shows. At the same time, I accept that the need for cash management will always be represented by one or more registers with the cash-dimension as the primary one.

Sometimes, it is difficult, impossible, undesirable even, to settle on an overall budget. Sometimes, people will first of all settle on a principle (P). What subsequent action will cost is accepted as inevitable.8 Even such a principle is easy to model with relational accounting. An additional dimension and corresponding register are all that is needed. See figure 4.

figure 4



It should be clear that separate registers make an endless variety of models for accounting possible. I repeat that dimensions and registers can be chosen as needed.9 This freedom extends to the relationships between (accounts in) registers. These can be established. And again broken. And so on.

However, variety is not limited to the horizontal plane of accounting. As a result of unequivocal classification, registers can be stacked, covering levels as needed. Thus vertical relationships originate. The financial structure of cooperation, delegation and reporting is translated to a network of accounting entities. Every entity, most likely corresponding to an organizational unit, consists of registers as required. Every register comprises relevant accounts that can be horizontally and vertically related to accounts in other registers. Of course, financial data from a higher to a lower level (delegation) or vice versa (reporting) can only be exchanged between (accounts in) identical registers. For the sake of brevity, I shall skip explanations of multi-levelled accounting.10 It all amounts to models with registers on different levels.


business organization

Relational accounting is a more flexible alternative for Kameralistik, fund accounting and similar methods. The same conclusion I propose for now traditional double-entry accounting. Does not relational accounting create a synthesis of Kameralistik and such double-entry accounting? The alternative method, therefore, is also valid for business organizations. A model, limited to one level, is shown again in figure 4. The only modification would be the abandonment of the (register of) principles and the corresponding dimension. Or, what? Is an additional dimension etc. of fiscal year needed?

The play of dimensions and registers gives relational accounting the flavour of a more generic method. It can be applied at whatever organization, wherever in the world. Local differences with respect to places and levels can be entertained. In an overall organization, an accounting balance between autonomy and coordination of its entities may be developed and maintained.

Application at a typical business organization could assign the following meaning to registers in a relational model of a multi-dimensional accounting cycle. For (constituent parts of) assets, for example, register B seems fitting. It can also contain records of general agreements on funding (equity). Commitments for redemption and depreciation should be recorded in register C. That would also seem the right register for obtained and issued contracts. Actual redemptions and depreciations, of course, are recorded in register X. The registers H and E along the cash-dimension provide space for records of estimated terms, instalments etc.11Seperate registers for such records ensure, this is generally speaking of course, that modifications of estimates without changing the basic amount require minimal recording effort.

A business organization must become accustomed to the restricted scope of a double entry. Relational accounting implies that every phase has its own accounting equation; the accounts in every single register balance. A major difference with traditional double-entry accounting, is therefore that reports like the balance sheet and the profit and loss statement are not compiled with readily available data from special accounts. Each report is the result of data processing of raw data from relevant (accounts in) registers.12 And this shows relational accounting to be an application of the more general principle of the strictest possible separation between, on the one hand, recording and reporting on the other. A consistent separation is the basis of increased flexibility. The same argument applies to auditability; the number of stages between original record and subsequent reports is minimal.

Organizational processes exist where identification of more than one dimension is not warranted. In such cases relational accounting will not offer major advantages. Or else, it should be the simplicity with which accounting can be kept in harmony with its overall processes. And, by the way, this proved useful at the first Ministry. As it happened, the requirements for registers changed very soon.

With two or even more dimensions the advantages of relational accounting become evident, as a double entry is insufficient when it must be possible to present a transaction according to different outlooks as covered by dimensions. Double-entry accounting with its traditional paradigm must be relinguished; to preserve it means that triple entries etc. will be inevitable in a multi-dimensional process. Relational accounting, however, keeps every entry always limited to two opposing changes in the financial position. This kind of double entry is sufficient because a transaction is strictly related to only one, single phase. Recording does not proceed beyond the corresponding register. Only when a report is required are data collected and processed. Relationships between accounts secure the relevance of financial reporting. Balance sheet and profit and loss statement are, therefore, nothing more and nothing less than two examples of reports.



Separation between recording and reporting makes for post-coordination. This technique originated in the field of information retrieval for libraries. The basic idea is division of a record in separate elements. As required, some or even all elements can be compiled into a report. Equally important, reporting only takes place at the time of an actual requirement.

Intelligent division into elements and the possibility of handling them separately, have far-reaching, positive implications. On the one hand any report is possible as long as it is shown to be contructed out of a combination of available elements. On the other hand records can be kept to a minimum; a sensible choice of elements leads to this practical advantage.

Post-coordination, it should be noted, is a phenomenon that is always only present to a certain extent. Its limits are determined by the existing elements. This goes for every method, and relational accounting is no exception.

Double-entry accounting brings a characteristic classification with it. The chart of accounts signifies this choice. In a similar way, relational accounting also uses a chart of accounts. There is no principal difference. But elements like dimension, phase or register etc. feature predominantly in transparant models. Furthermore, and most significantly, the meaning of the concept of account shifts. Double-entry accounting is set up to reflect an organizational process as a one-dimensional cycle. In practical applications, therefore, acount numbers contain implicit information about such an artificially constructed process cycle. I reiterate that relational accounting limits the record of a transaction, in the form of double entry, to one single register. Relational accounting uses accounts especially for explicit reference. I risk an unnecessary repetition when I say that, for a process cycle with multiple dimensions, account numbers cannot easily and implicitly contain information about such cycles.

As a result of improved transparancy, it is easier to reach an optimal choice of data for elements. What I mean is that such parameters, using them more consciously as both representations and constructive elements, allow playful activity in organizations. Again, accounting included. And perhaps therein lies the critical advantage of a method like relational accounting. Accounting is now easy to model; non-specialist employees who have overall responsibilities can make confident decisions about a tool that should make a vital difference to their management results. Accounting experts will have a shock. The good old method is a thing of the past. After they have recovered, they will notice that their influence has increased. On the basis of relational accounting, they are now better understood and therefore more highly valued.


incompleteness and claiming

Choosing of elements bearing a certain meaning is called the activity of classification. An extreme separation between recording and reporting demonstrates that classification essentially precedes reporting. However, are the choices embodied in the earlier classification still relevant at the time of a particular information requirement? An outdated classification brings the problematic nature of information and reporting to light. Absolute reliability and, especially, completeness do not exist. Disputes may arise involving the concepts of claiming and accounting. The reason is that in our social order claims have no a priori limits.13 So, a random claim could, a posteriori, lead to the logical conclusion that essential accounting information is lacking.14 Any reconstruction of events-that-the-claim-is-about is impossible without other reliable sources. Is this black hole of history real? Yes, of course, according to the claimant.15

At least the dilemma is clear.16 Whatever accounting method or, for that matter, information policy in general fails to overcome it in principle. Therefore, I intend relational accounting as a vehicle for recognizing funadamental imcompleteness etc. I know that such recognition is not an integral part of relational accounting as a method, but it can remain in focus using it properly as nothing but a flexible tool. Explicit classification helps. Together with increased flexibility of accounting tools, more guarantees exist for justified claims to lead to real rehabilitation.17 And, instead of mentioning more guarantees, I should really say that I prefer the fewest obstacles. A basic condition for a humane society consists of rendering claims not invalid a priori. The opposite condition should not be what we are trying to learn from our neighbours in the east or, especially, from our own history.




1. My design of relational accounting is not related to the political developments I indicate. I pragmatically mention them as an inducement to consider relational accounting as an alternative.

2. A more comprehensive treatment of almost 600 pages is given in my book Aspecten en Fasen, to be published early 1991 in a Dutch language edition. The title translates as follows: Aspects and Phases: entries on relational accounting, organizational information policy, change etc. and vice versa

3. Another matter is whether the western social order really is, or ever was, capitalistic. Here I also evade the question whether double-entry accounting is effect or cause of capitalism. As a general rule, interaction between such phenomena is assumed. But the latter, i.e. accounting as cause, makes an equally interesting theme: Capitalism, beyond the paradigm of accounting.

4. So information processing shifts from the moment of recording to all moments where reports are compiled. Therefore, practical application of relational accounting assumes electronic information processing.

5. My examples are based on expenditure-oriented processes. The same argument holds for receipts; their target can also be a budget etc.

6. See Aspecten en Fasen, chapter 23, for a more elaborate formulation of my proposal for pluralism.

7. Fundamentally lacking is recognition of different types of commitments. And there should still be other types than those that the first Ministry acknowledged. The necessary differentiation immediately shows that every commitment type will almost invariably lead to a characteristic dimension which is its primary. See also figure 4 where the register for commitments has changed track compared to figures 2 and 3 and is thereby renamed.

8. Also, a principle is only valid as long as there is ground to carry it.

9. Of course, official directives and regulations are a determining factor for a minimal choice.

10. See Aspecten en Fasen, chapter 25.

11. In chapter 24 of Aspecten en Fasen, an example of relational accounting for business organizations is sketched somewhat less cursorily.

12. See also note 4.

13. At least, this is what the ruling ideology of the western world tries to suggest. How does it work in practice? Precisely people under the banner of a business enterprise can, for their part, claim a limited liability.

14. At first, and in its broadest sense, a claim is identical to an information requirement in general.

15. J.F. Lyotard, in his book The Different (University of Minnesota Press, 1988), starts with a story of lost claims: the murdered victim is unable to tell his own [story].

16. For the sake of simplicity, that is, simplicity of presentation, I even exclude the fundamental interdeterminacy of information. Or, in general, the fundamental indeterminacy of phenomena-as-related.

17. Such rehabilitation always presupposes, naturally, that damages can be realistically recovered. So, yet another problematic.



September 1990, web edition 2002.