Anda di halaman 1dari 2

CONTINUITY EQUATIONS

Short Introduction to the Restricted Vector Autoregressive Model (RVAR) of Continuity Equations for Continuous Assurance
Erik van Kempen, Student Member, IEEE, <erikvankempen@ieee.org> Fontys Hogeschool Financieel Management, Fontys University of Applied Sciences

AbstractContinuous assurance can use continuity equations by modeling predictors for reported quantities in business process steps. One of the best models for these continuity equations is the Restricted Vector Autoregressive Model (RVAR). In this model lags are estimated during the modeling process as opposed to predening xed lags in the basic VAR model. Furthermore, non-correlated variables are excluded from the nal model to minimize the overall posterior R2 of the model. Index TermsContinuous assurance, continuity equations, RVAR, Restricted Vector Autoregressive model, data analysis.

I. I NTRODUCTION ONTINUITY EQUATIONS are a fundamental part of classical physics, but the application of these equations in the eld of nance has not been explored in detail. In their paper on continuity equations Kogan et al. have provided the theory and a number of solutions for applying these equations as a tool in the continuous assurance domain. [1] One of these solutions is the Subset Vector Autoregressive Model (Subset VAR) or Restricted Vector Autoregressive Model (RVAR). This model was implemented in the SAS language. In this paper the model is introduced.

In most businesses the ow of goods is the most important basis for revenue recognition. As such, it can be used to provide evidence for the completeness, timeliness and accuracy of the reported revenue. If the continuity equations hold for a certain business process, one can assert that there are no leakages from the transaction ow, i.e. the integrity of the ow of goods can be asserted. Therefore, continuity equations provide a method to evidence the integrity of the basis for revenue recognition, which makes them a valuable tool in continuous assurance. Continuity equations are based on historical data of quantities in the separate steps of business processes. For example, the sales cycle can be modeled as three separate steps: receiving the order from the customer, shipping goods to the customer and invoicing for the ordered and shipped goods. Of course the quantity of ordered goods today will show up in the invoicing step a certain amount of days later. The daily ow of goods between these steps can be dened as a certain quantity Q and lag between the steps . A. Base Model In this paper we will focus on the sales cycle, with the three previously dened process steps. The continuity equations for the sales cycle can be represented as Equation 1. In this model ordert is the quantity ordered at time t, the terms are N 1 transition vectors for a multivariate linear model and the M terms are N 1 vectors containing daily aggregates of quantities for the given dimension. ordert = oo M (order) + so M (shipped) + io M (invoiced) shippedt = os M (order) + ss M (shipped) + is M (invoiced) invoicedt = oi M (order) + si M (shipped) + ii M (invoiced) (1) Each of these sub-equations models a predictor for the reported quantities in a specic step in the business process. As previously dened, the quantities are related to quantities in the other process steps by a time delay (lag). For example, if orders are shipped in exactly one day, without exception, and invoicing is performed simultaneously with shipping, the resulting predictors can be dened as Equation 2 ordert = 1 shippedt+1 + shippedt = invoicedt = 1 ordert1 + 1 ordert1 + 2 invoicedt+1 2 invoicedt 2 shippedt (2)

July 14, 2013 II. C ONTINUITY E QUATIONS Continuity equations are a fundamental part of classical physics. These equations describe the transport of a conserved quantity, while simultaneously ensuring conversation of this quantity (like mass and/or energy). Accordingly similar relations can be dened for the transport of quantities within a system in the nancial domain. For example the movement of reported quantities between steps in the key business processes can be described with continuity equations. The term continuity equations was coined by Vasarhelyi and Halper in 1991, when they modeled the ow of billing data at AT&T. [2] Although Vasarhelyi and Halper proposed continuity equations more than 20 years ago, little research has been performed on the application in practice and implementation of a decent continuity equations model. Especially research focusing on the VAR model has been rarely performed in the last two decades. Only Dzeng [3] and Kogan et al. [1] have also considered the VAR model in their papers.

CONTINUITY EQUATIONS

B. Basic Vector Autoregressive Model In the basic Vector Autoregressive Model (VAR) the model is estimated optimizing for the overall R2 by trying different lags for the process steps. Only the maximally expected lag is provided to the algorithm and it then tries to nd the best tting model by iterating trough all lag possibilities. The exact lags do not have to be known prior to modeling as the best tting lags are determined while modeling. One can easily understand that it is not always trivial to determine lags prior to the modeling process, e.g. lags in the purchasing cycle are highly dependent on the policies and processes at third parties. Therefore, the VAR model can be a powerful tool for modeling continuity equations if lags can not be predened trivially. C. Restricted Vector Autoregressive Model Kogan et al. have shown in their studies that the VAR model did show great results. More importantly, they showed that the Subset VAR model resulted in even better results. The Subset VAR model translates roughly to optimizing for R2 of the predictor by excluding uncorrelated variables from the VAR model. For example, if the mean lag between order and shipping is less than a month the variable shippedt+365 and any other variable shippedt+ with > 31 are removed because these are obviously not correlated variables. The Restricted Vector Autoregressive Model (RVAR) not only optimizes for the overall R2 by trying different lags, it also optimizes the goodness of the t by excluding the uncorrelated variables and reiterating the VAR process. III. A DVANTAGES The RVAR model still has the advantage of the VAR model over structural models and simultaneous equations becasue the lags do not have to be xed prior to the modeling process. This is a great advantage for the analyst who performs the modeling. The step of estimating lags is very sensitive to errors in interpretation and has enormous impact on the resulting model. By removing this step from the modeling process less errors are introduced, resulting in a more accurate predictor. Furthermore, the resulting RVAR model is most likely to have a higher R2 compared to the basic VAR model, which means that the model is a better predictor in terms of accuracy for the reported quantities. A higher accuracy can lead to a lower deviation exception threshold and thus a smaller detection risk in the continuous assurance domain. Finally, the resulting model incorporates less variables and can be understood more easily. Not every individual concerned with continuous assurance is an expert in the eld of data analysis or statistics. A smaller and more easily understandable model can help potential users of the model to adapt more quickly to using it in the assurance domain.

R EFERENCES
[1] A. Kogan, M. G. Alles, M. A. Vasarhelyi, and J. Wu, Analytical procedures for continuous data level auditing: Continuity equations, 2010. [2] M. A. Vasarhelyi and F. B. Halper, The continuous audit of online systems, Auditing: A Journal of Practice & Theory, vol. 10, no. 1, pp. 110125, 1991. [3] S. Dzeng, A comparison of analytical procedures expectation models using both aggregate and disaggregate data, Auditing: A Journal of Practice & Theory, vol. 13, no. Fall, pp. 124, 1994.

Erik van Kempen Erik (born April 25, 1987) is a lecturer in Statistics, ERP and Business Intelligence at Fontys Univerity of Applied Sciences in Eindhoven, The Netherlands. Prior to this lectureship, he obtained a Bachelor of Business Administration in Accountancy. His research interests are in the areas of continuous assurance, smart auditing, business intelligence and process mining.

Anda mungkin juga menyukai