A CONSOLIDATED ENTERPRISE REFERENCE MODEL Integrating McCarthy s and Hruby s Resource-Event-Agent Reference Models

Transcription

1 A CONSOLIDATED ENTERPRISE REFERENCE MODEL Integrating McCarthy s and Hruby s Resource-Event-Agent Reference Models Wim Laurier, Maxime Bernaert, Geert Poels Department of Management Information Systems and Operations Management,Faculty of Economics and Business Administration, Ghent University, Tweekerkenstraat 2, 9000 Ghent, Belgium wim.laurier@ugent.be, maxime.bernaert@ugent.be, geert.poels@ugent.be Keywords: Abstract: Reference Model, Resource-Event-Agent, Ontology This paper introduces a new REA reference model that integrates the transaction and conversion reference models provided by McCarthy, which aimed at designing databases for accounting information systems, and Hruby, which aimed at software development for enterprise information systems, into a single conceptual model that accounts for both inter-enterprise and intra-enterprise processes. This consolidated reference model was developed to support data integration between multiple enterprises and different kinds of enterprise information system (e.g. ERP, accounting and management information systems). First, the state of the art in REA reference models is addressed, presenting McCarthy s and Hruby s reference models and assessing their ability to represent exchanges (e.g. product for money), transfers (e.g. shipment) and transformations (e.g. production process). Second, the new, consolidated REA enterprise reference model is introduced. Third, object model templates are presented, demonstrating that the consolidated REA reference model is able to represent exchanges, transfers and transformations, where McCarthy s and Hruby s reference models can each only represent two of these features. 1 INTRODUCTION Domain conceptualisations can be represented as meta-models for domain-specific modelling languages and modelling patterns and reference models in a generic modelling language. Domainspecific meta-models provide the modeller with specific modelling languages that incorporate domain-specific constructs and modelling constraints, which supports the modeller in his modelling tasks by providing the right set of constructs (cf. vocabulary) and predefined construct interconnections (cf. grammar). Representing these domain-specific meta-models using ontology languages (e.g. OWL (W3C, 2004)) allows for formal model reasoning (i.e. inference) and correctness checking. Further, domain conceptualisations provide a basis for domainspecific modelling patterns and reference models. Modelling patterns document modelling practices and provide frequently observed solutions to domain-specific modelling problems. Reference models equip modellers with standardised domain models. In this sense, reference models come close to Borst s (1997) definition of ontologies An ontology is a formal specification of a shared conceptualisation which extends Gruber s (1993) definition. In this paper, a new enterprise reference model that is based on the Resource-Event-Agent (REA) enterprise ontology (Geerts and McCarthy, 2002, Geerts and McCarthy, 2004) is proposed. The current financial crisis has proven that accurate financial information with a foundation in real operations is essential for a healthy economy. As the Resource-Event-Agent (REA) ontology (Geerts and McCarthy, 2002, Geerts and McCarthy, 2004) was originally designed for an environment where accountants and non-accountants share information about the same economic phenomena (McCarthy, 1982), it offers a particularly interesting basis for an enterprise reference model in the light of the current crisis. Although the term REA reference model has never been coined before, to the best of our knowledge, we believe REA reference models can be found in McCarthy s (1982) foundational REA article and Hruby s (2006) book on business patterns design. Although these reference models are founded in the same ontology, they focus on different application domains, which lead to particular modelling compromises and models with

2 different representational capabilities. McCarthy s reference model focuses on transactions between individual enterprises, by representing the exchange of resources (e.g. money for product) and the resulting resource transfers between trading partners. Additionally, McCarthy s reference model allows for view integration (i.e. integrating the mirrored views of trading partners with opposing economic interests (e.g. seller and buyer)). Hruby s reference model, on the other hand, focuses on the processes inside an enterprise by representing exchanges and conversions (i.e. process input becomes process output), but not the resource transfers between trading partners. The reference model that is presented in this paper integrates both enterprise perspectives (i.e. transactions and processes) in a single reference model that is able to represent both and provides enough semantic power to discriminate transfers, exchanges and transformations. Section 2 discusses McCarthy s and Hruby s REA reference models and their specific characteristics. Section 3 presents the new, consolidated REA reference model and a number of object models for transfers, exchanges and transformations that instantiate this consolidated reference model. Finally, the consequences of the REA reference model integration are discussed in section 4 while section 5 presents conclusions and directions for future research. 2 STATE OF THE ART The REA ontology knows three main primitives, namely economic resource, economic event and economic agent, which are abbreviated to resource, event and agent in the remainder of this paper. Resources in the REA ontology are defined as goods, services or rights that have utility and are scarce and under the control of a legal or natural person (Ijiri, 1975, McCarthy, 1982, ISO/IEC, 2007). Events are occurrences in time that relate subsequent process states to each other and involve gaining (i.e. increment) or losing (i.e. decrement) control over economic resources (Yu, 1976, McCarthy, 1982, ISO/IEC, 2007). Agents are natural or legal persons (e.g. employee, customer) that are accountable for, participate in or initiate economic events. 2.1 McCarthy s Reference Model Fig. 1 shows a modernised version of McCarthy s foundational reference model (McCarthy, 1982) as it appears in ISO s business transaction scenario standard (ISO/IEC, 2007). This model differs from the initial model by decomposing the n-ary control relation in two binary participation (i.e. agent-event) relations, conforming to Weber s critique (Weber, 1986), and the merger of the economic agent and its subclass economic unit in the AGENT class. The STOCK-FLOW association between the RESOURCE and EVENT classes in fig. 1 shows which resources are involved in and affected by which events. The INCREMENT-DECREMENT association in fig. 1 represents the duality relationship between two events (i.e. one increment and one decrement). This duality balances the changes in resources due to events (Ijiri, 1975) representing of principle of economic reciprocity, which requires adequate compensation (i.e. increment event) (e.g. payment received) for lost resources (i.e. decrement event) (e.g. resource shipments). Revealing REA s accounting background, the value of the increases should equal the value of the decreases that are connected to each other through duality links. For instance, duality in market transactions dictates that when a company sells products to a customer (i.e. an event that decreases the value of the company s inventory) a requiting event like a payment or delivery of equally valued goods (e.g. as in barter trade) by the customer must follow, meaning that there is a dual event that balances the decrease in value caused by the sale. The EXCHANGE- INSIDE_PARTY association reveals the agent whose view determines which events are increment and which are decrements. The EXCHANGE- OUTSIDE_PARTY association relates the inside party s counterparty to the economic event. Figure 1: McCarthy s REA Reference Model.

3 Figure 2: McCarthy s Integrated View Exchange. With McCarthy s reference model, exchange and transfer object models can be constructed. Fig. 2 shows a template that integrates the views of both trading partners. We recognise two resource transfers with opposite directions (i.e. one represents a cash inflow for the vendor and a cash outflow for the buyer; one represents inventory goods inflow for the buyer and a goods outflow for the seller). We also discriminate two mirrored exchange templates (i.e. cash for goods in the buyer s purchase and cash disbursement duality, goods for cash in the vendor s sale and cash receipt duality). Taking the view of the buyer, an exchange is the duality between a purchase and a cash disbursement. The purchase represents an inflow of resources and an inventory augmentation, whereas the cash disbursement represents a cash outflow. Since this object model shows the buyer side of the cash for inventory exchange, the buyer is the inside party for both events and the vendor the outside party. Such exchange templates can be used to model all kinds of exchanges (e.g. money for product, product for money, product for product, money for money). For each exchange object model, a mirroring exchange object model can be created for the counterparty. In such a model the inside and outside party roles are switched and the increment/decrement and inflow/outflow role pairs are mirrored. Modelling a transfer of resources between the inventories of two trading partners, a transfer model relates the corresponding mirrored events in an exchange with each other. This mirroring relation is mediated by a resource object for which one event represents an inflow (e.g. PURCHASE) and the mirroring event represents and outflow (e.g. SALE). Although the agents connected to the mirrored events are the same, the inside and outside party roles are switched, which represents the opposing economic interests of the trading partners (i.e. SELLER and BUYER). Note that the mirroring relation between the buyer s view on an event and the vendor s view on the same event is implicit in fig. 2. The downside of this representation is that real events are artificially decomposed (e.g. purchase and sale are actually two views on the same event). 2.2 Hruby s Reference Model Figure 3: Hruby s REA Reference Model. Fig. 3 shows Hruby s (2006) reference model as it appeared in his business patterns book. By explicitly discriminating increment and decrement events, it

4 Figure 4: Hruby s REA Value Chain differs from McCarthy s reference model, which chooses to represent increment and decrement as roles of events and not as kinds of events. Hruby s reference model also incorporates provide (i.e. PROVIDER-PROVIDE) and receive (i.e. RECIPIENT-RECEIVE) relationships that connect agents with increment and decrement events. The provide relationship relates the event to the agent that experiences a resource decrement due to the event and the receive relationship relates the event to the agent that experiences a resource increment due to the event. Consequently, a provide relationship relates a decrement event to the agent that experiences the decrement (i.e. inside party in McCarthy s model) and a receive relationship relates an increment event with the agent that experiences the increment (i.e. inside party in McCarthy s model). Table 1 summarises these view relations between agents and events, and the conceptual differences for the participation relation (i.e. agentevent) in McCarthy s and Hruby s reference models. Agent Event Increment Decrement Provider Outside Party Inside Party Recipient Inside Party Outside Party Table 1: Provider and Recipient as Inside and Outside Agent. The duality (i.e. DUAL_TO-DUAL_TO) relation between the INCREMENT_EVENT and DECREMENT_EVENT classes is identical to the duality in McCarthy s model except for the fact that it relates distinct classes and not one event class that plays distinct roles. Finally, Hruby decomposes McCarthy s stock-flow relation into consume (i.e. OUTFLOW-CONSUME), use (i.e. OUTFLOW-USE) and produce (i.e. INFLOW-PRODUCE) relations. The difference between the use and consume outflow relations is that use relations temporary occupy a resource without affecting its ability to participate in subsequent decrement events, while consume relations indicate that a resource is no longer available for further decrements. Hence, Hruby incorporates part of McCarthy s object model logic in his reference model (i.e. events in an outflow role are decrement economic events and events in an inflow role are increment economic events). The basic value creating processes in an enterprise can be categorised as acquisition, conversion and revenue generating processes, which construct an enterprise s value chain (McCarthy, 2003). Hruby s reference model provides sufficient semantic power for representing these economic phenomena inside an enterprise. Acquisition and revenue generating processes can be represented using the exchange template and conversion processes are represented using the conversion template. Fig. 4 integrates an acquisition process with a conversion process. The acquisition process shows how ingredients are purchased in return for cash. The conversion process represents the conversion of ingredients into pizza. The acquisition part of fig. 4 shows the exchange of CASH and INGREDIENTS between the PIZZA_RESTAURANT and its SUPPLIER. The conversion part of fig. 4 shows the transformation of INGREDIENTS into PIZZA that is performed by the COOK for the PIZZA_RESTAURANT.

5 Hruby s exchange template is fairly identical to McCarthy s articulation (i.e. part of fig. 2). The main difference is that the increment and decrement roles in McCarthy s model are replaced by specific classes. The inside and outside party roles on the participation relations in McCarthy s model have also been replaced by provider and recipient roles. The conversion model that represents (part of) a production process, is not incorporated in McCarthy s version of REA, as it is based in event accounting and such a conversion model is more related to activity-based costing (Cooper and Kaplan, 1988). The conversion is represented as an implicit exchange between process inputs (i.e. INGREDIENTS) and process outputs (i.e. PIZZA), where the employer (i.e. PIZZA_RESTAURANT) provides the inputs and receives the outputs from the employee (i.e. COOK) and the employee receives the inputs and provides the outputs to the employer. The conversion process itself is modelled as a duality between one or more decrement events that use or consume input resources and one or more increment events that produce output resources. The downside of this kind of representation is that conversion processes are represented as exchanges, while employees never own the resources they work with, and that conversion processes are artificially decomposed in a collection of increment and decrement events. 3 CONSOLIDATED REFERENCE MODEL A remarkable feature of both McCarthy s and Hruby s reference models is that they both use the agent concept to model trading partners and event performers. The downside of this choice is that trading partners cannot be explicitly discriminated from pure event performers (e.g. the trucker cannot be discriminated from the enterprise that ships the goods). To allow for this discrimination, we reintroduce the concept of economic unit that was removed from McCarthy s original reference model as a superfluous synonym for the agent concept. Therefore, the agent and economic unit concepts need to be redefined. We redefine an Agent as the natural person that executes the event and an Economic Unit as the legal or natural person that loses or gains control over resources in decrement and increment events respectively. Consequently, economic units, which were originally defined as a subclass of Agents (McCarthy, 1982), determine the scope of the context in which economic activities take place. This new definition also fits McCarthy s original model, where economic units are inextricably bound with the inside party role that determines which events are classified as increments (i.e. gaining control over resources) and decrements (i.e. losing control over resources). Figure 5: Our REA Reference Model. Fig. 5 reflects the key role of the economic unit concept in our reference model. In this reference model, the economic unit defines the duality that connects the explicitly modelled increment and decrement roles for the event that forms the centre of our reference model. The agents in this reference model participate in events, meaning that they execute events without experiencing their economic consequences (i.e. gaining or losing control over resources). These consequences are experienced by the economic units and explicitly represented by the INCREMENT and DECREMENT classes in the reference model that mediate between the ECONOMIC_UNIT and the RESOURCE over which they gain or lose control. The reference model (fig. 5) also reveals that a single EVENT can be viewed as an increment and decrement at the same time, by the same or different economic units. Fig. 6 shows the integration of a transfer and an exchange into a consolidated view exchange model. It shows two mirrored dualities (i.e. PURCHASE and SALE) and two transfer events that connect them to each other (i.e. ACQUIRE/SELL and PAYMENT). The PURCHASE duality shows how the BUYER exchanges MONEY for PRODUCT, where the SALE duality represents how the SELLER exchanges PRODUCT for

6 Figure 6: REA Consolidated View Exchange Model. MONEY. The ACQUIRE/SELL event shows that ACQUIRE and SELL are in fact two roles of the same event and the PAYMENT event reveals that PAY and GET_PAID are also two roles of the same event. The ACQUIRE role is assigned to the ACQUIRE/SELL event by the BUYER, while the SELL role is assigned by the SELLER. The GET_PAID role, on the other hand, is assigned to the PAYMENT event by the SELLER and the PAY role by the BUYER. The exchange duality models the exchange (e.g. PURCHASE), which involves a resource inflow (i.e. ACQUIRE) and a resource outflow (i.e. PAY), from the viewpoint of a single economic unit. Taking the buyer perspective, the economic unit (i.e. BUYER) defines the DUALITY and the INCREMENT and DECREMENT roles for the resource transfer events involved. Such a transfer event (e.g. a MONEY transfer (i.e. the PAYMENT event)) is perceived by a BUYER and a SELLER. The PAY decrement role represents the view of the BUYER on the PAYMENT event and the GET_PAID increment role shows the view of the SELLER on the PAYMENT event. The CLERK object represents the person that executes the payment regardless of whether he/she experiences the economic consequences of this action. Figure 7: REA Transformation Event Model. Finally, fig. 7 shows a transformation template in which transformation inputs are transformed into outputs. The template shows that the TRANSFORMATION event takes place inside a single economic unit (i.e. EMPLOYER) and that both the inputs and outputs are owned by the EMPLOYER, without a transfer of ownership during the conversion process. This representation is closer to reality that Hruby s implicit ownership transfer during the conversion process. The template also shows that an EMPLOYEE performs the TRANSFORMATION event. For representing an entire production process, subsequent events can be modelled using the transformation template, creating multiple events that are all related to the same economic unit (i.e. the enterprise in which they take place) and share one or more resources (i.e. the output of one conversion process is the input for a subsequent one). At the start of such a process model we find one or more exchange templates (fig. 6) that represent the acquisition of the process inputs and at the end we find one or more exchange templates that represent how revenue is generated from process outputs. 4 DISCUSSION Our REA reference model integrates the features of McCarthy s and Hruby s reference models. This allows the modeling of transformations, transfers and exchanges via a single reference model, which eliminates the need for model integration efforts that replicate the same REA constructs (e.g. agent, event, resource) in a single model. Dunn et al. (2005), for example, replicate the same REA constructs in their models, because they duplicate the REA reference model for each purpose (i.e. transformation, transfer or exchange). As they discriminate three kinds of

7 exchange processes (i.e. payroll, acquisition/payment and revenue or sales/collection) and two kinds of conversion processes (i.e. manufacturing, which converts raw materials into final products and financing, which transforms earnings to cash for spending) their models tend to be large. With the consolidated REA reference model, the need for this REA construct duplication disappears. The consolidated REA reference model also resolves the ambiguous use of the agent concept, making the REA ontology easier to understand. The redefinition of the economic unit concept also allows models to represent previously implicit semantics related to the control over resources and the view on events that is implied by this control relationship. Where previously the view of every enterprise was represented in a separate model, the scope of different enterprises can now be represented in a single model via the economic unit concept and its relations with resources, events and agents. This explicit representation of enterprise boundaries allows for a central administration of transactions between and transformations within enterprises. Apart from the central administration of data generated by multiple enterprises, data interoperability is also facilitated when the consolidated enterprises use the same reference model and agreement is reached about a minimal set of attributes (e.g. identifiers). The explicit enterprise scope also eases strategic analysis in the light of vertical integration (i.e. what would the new enterprise look like when two supply chain partners (e.g. customer and supplier) merge). The redefinition of the economic unit and agent concepts facilitates the integration of data across enterprise boundaries whereas the intuitive event concept eases process modelling. Together, they can help improve product traceability by identifying the event chains (i.e. transfer and transformation events) that lead to the products, irrespective of the number of enterprises (and enterprise information systems) in which products and their constituents have their origin. Such product tracing infrastructure might support product authentication in the battle on counterfeit and other supply chain intrusions (e.g. food safety scandals) (Bechini et al., 2008). It may also help to trace the origin of money (e.g. drugs money) in the battle against money laundering. Additionally, the consolidated reference model reveals that transfer and transformation dualities (Geerts and McCarthy, 2004) actually represent transfer and transformation events, reinstalling the exchange duality as the only true duality (i.e. the balance between resource inflows and outflows). The explicit recognition of transfer and transformation events also eases the use of REA for process modelling and representation (Laurier and Poels, 2009a), as the representation of events is explicitly discriminated from representing the balance between resources flows (e.g. money for products). This purified duality concept then describes the force that balances the real economy (i.e. product flows) with its monetary equivalent (i.e. money flows) (Foster, 1922). The event interpretation in the consolidated REA reference model is made more intuitive by eliminating the transfer event decomposition (e.g. cash disbursement and cash receipt) in McCarthy s reference model and the transformation event decomposition (e.g. consume inputs and produce outputs) in Hruby s reference model. The more intuitive event concept facilitates the use of REA as a shared conceptual basis for accounting and management information systems (e.g. ERP, CRM, VMI). This shared conceptual basis will also allow analyzing process information (e.g. event logs) not only in terms of time and quantity, as is currently done by process mining tools (Rozinat et al., 2009), but also in monetary terms. 5 CONCLUSIONS The reference model integration presented in this paper supports model integration between interenterprise (e.g. for transaction recording systems) and intra-enterprise (e.g. for production process monitoring systems) models, as both kinds of systems can now rely on the same reference model. Additionally, data sharing between accountants and non-accountants, which was the original purpose of REA, is expected to be facilitated. Key to the integration of the existing REA reference models was the partial redefinition of the economic unit and agent concepts and the purification of the duality concept, which makes the REA event concept more intuitive. We are currently working on accounting data integration for subsidiaries of a single parent company (i.e. creating consolidated accounts) with the consolidated REA reference model. The proof of concept is developed as a database that implements the structure of the consolidated REA reference model and is attributed with queries for booking invoices (i.e. sales and purchases), amortisations, etc. Based on the input these queries generate, a general ledger, a journal and (consolidated) annual accounts can be constructed. Extending this proof of

8 concept with queries for cost accounting is part of future work. In the future we will also further explore the abilities of REA as a language for business modelling and use the consolidated REA reference model proposed in this paper to create strategy tools (e.g. for determining the opportunity of vertical integration, simulating the return of business process redesigns). The basic simulation models (and part of their functionality) for the creation of such strategy tools were presented in (Laurier and Poels, 2009a) and (Laurier and Poels, 2009b). Together with the data structure in the proof of concept for accounting purposes the creation of a proof of concept for such strategy tools is part of our future research. REFERENCES BECHINI, A., CIMINO, M. G. C. A., MARCELLONI, F. & TOMASI, A. (2008) Patterns and technologies for enabling supply chain traceability through collaborative e-business. 50, BORST, W. N. (1997) Construction of engineering ontologies for knowledge sharing an reuse. Centre for Telematics and Information Technology. Enschede, Universiteit Twente. COOPER, R. & KAPLAN, R. S. (1988) Measure Costs Right: Make the Right Decision. Harvard Business Review, 66, DUNN, C. L., CHERRINGTON, J. O. & HOLLANDER, A. S. (2005) Enterprise information systems : a pattern-based approach, Boston, McGraw- Hill/Irwin. FOSTER, W. T. (1922) THE CIRCUIT FLOW OF MONEY. American Economic Review, 12, 460. GEERTS, G. L. & MCCARTHY, W. E. (2002) An ontological analysis of the economic primitives of the extended-rea enterprise information architecture. International Journal of Accounting Information Systems, 3, GEERTS, G. L. & MCCARTHY, W. E. (2004) The Ontological Foundation of REA Enterprise Information Systems. Michigan State University. GRUBER, T. R. (1993) A translation approach to portable ontology specifications. Knowledge Acquisition, 5, HRUBY, P. (2006) Model-driven design using business patterns, Berlin, Springer. IJIRI, Y. (1975) Theory of accounting measurement, Sarasota, Fla., American Accounting Association. ISO/IEC (2007) Information technology - Business Operational View Part 4: Business transaction scenario - Accounting and economic ontology. ISO/IEC FDIS : 2007(E). LAURIER, W. & POELS, G. (2009a) Simulating Liquidity in Value and Supply Chains. IN AL., A. A. E. (Ed.) CIAO! 2009 and EOMAS Amsterdam, Springer-Verlag LAURIER, W. & POELS, G. (2009b) A Simulation Model Articulation of the REA Ontology. On the Move to Meaningful Internet Systems: OTM 2009 Workshops. MCCARTHY, W. E. (1982) The REA Accounting Model: A Generalized Framework for Accounting Systems in a Shared Data Environment. Accounting Review, 57, MCCARTHY, W. E. (2003) The REA Modeling Approach to Teaching Accounting Information Systems. Issues in Accounting Education, 18, ROZINAT, A., WYNN, M. T., VAN DER AALST, W. M. P., TER HOFSTEDE, A. H. M. & FIDGE, C. J. (2009) Workflow simulation for operational decision support. Sixth International Conference on Business Process Management (BPM 2008) - Five selected and extended papers. W3C (2004) OWL Web Ontology Language Overview. IN MCGUINNESS, D. L. & VAN HARMELEN, F. (Eds.). WEBER, R. (1986) Data Models Research in Accounting: An Evaluation of Wholesale Distribution Software. Accounting Review, 61, 498. YU, S. C. (1976) The structure of accounting theory, Gainesville, University Presses of Florida.