---

IN THE HIGH COURT OF SOUTH AFRICA

(Cape of Good Hope Provincial Division)

Case No. 7075/2001

In the matter between:

TRIPTOMANIA TWEE (PROPRIETARY) LIMITED First Applicant

NGT NETWORKS (PROPRIETARY) LIMITED Second Applicant

PETRUS JOHANNES BESTBIER Third Applicant

FRANCOIS ALBERTUS MARITZ N.N.O.

(In their capacities as Trustees for the time being of the

SourceCom Technology Solutions Employee Share Incentive

Trust. Trust No. IT 944/2001)

and

SANDRA CONNOLLY First Respondent

TIMOTHY MARK CONNOLLY Second Respondent

JUDGMENT

DAVIS J:

Introduction.

On 10 August 2001 applicants instituted proceedings against respondents in which they sought relief in terms of section 424 of the Companies Act 61 of 1973 (‘the Act’). Their claim arose as a result of a value-added-tax (‘VAT’) fraud perpetrated at SourceCom Technology Solutions (Pty) Ltd (‘SourceCom’).

SourceCom is an information technology company which provides large corporate customers, parastatal corporations and government departments with various products, including the sourcing of hardware, the administration of software licences and the provision of support services. From its formation until November 2001 first respondent was a director of SourceCom. She occupied the position of chief executive officer. Second respondent was appointed as SourceCom’s financial manager in June 1997 (or in 1998 ;there are two versions concerning the date of his appointment. To the extent that he commenced service in 1997 , his appointment would have been to SourceCom’s predecessor ‘s company. This factual issue is of no import to the resolution of this dispute).

He became the financial director in either July or September of 1999 and at the end of November 2000 he left the company’s employ and ceased to be a director.

Applicants contend that respondents, while involved in the management of SourceCom, were knowingly parties to the conduct of the business of SourceCom as from, at least, November 1998 until their leaving SourceCom, November 2000, for a fraudulent purpose. Alternatively, applicants contend that respondents acted recklessly, alternatively were grossly negligence in their conduct of the business.

Second respondent admitted his involvement in the fraud. His admission was contained in an affidavit; which formed part of applicants’ replying papers to this application. By contrast , first respondent has denied any knowledge of or involvement in such a fraud.

This application initially came before Blignault J, who, on 21 November 2001, made an order referring the following questions for the hearing of oral evidence:

1. Whether Ms Thorne (first respondent) was knowingly a party to the VAT fraud described in the applicants affidavit and

2. If so, the dates from which and to which she was such a party.

The second question is not an issue in this dispute. Applicants’ case is not that first respondent became aware of the fraud after its inception and thereafter facilitated or ignored it. They allege that she authorised the fraud from its inception in November or December 1998. In accordance with the order of Blignaut J, witness statements were filed on behalf of applicants by second respondent and by Ms Linda May Honey , an employee of Source Com, as well as on behalf of first respondent. Oral evidence was heard over five days . Evidence was given by second respondent and by Ms Honey on behalf of applicants. First respondent also testified.

The Section 424 Claim.

Section 424(1) (a) of the Act provides as follows:

‘When it appears, whether it be in a winding up, judicial management or otherwise, that any business of the company was or is being carried on recklessly with intent to defraud creditors of the company or creditors of any other person or for any fraudulent purpose, the Court may, on the application of … any creditor or member…of the company, declare that any person who was knowingly a party to the carrying on of the business in the manner aforesaid, shall be personally liable, without any limitation of liability, for all or any of the debts or other liabilities of the company as the Court may direct.’

The onus is upon the party alleging recklessness to so prove, and being civil proceedings, to establish the necessary facts on a balance of probabilities. Howard v Herrigel and Another NNO [1991] ZASCA 7; 1991 (2) SA 660 A at 672 E; Philotex (Pty) Ltd and Others v Snyman and Others [1997] ZASCA 92; 1998 (2) SA 138 (SCA) at 142 I – J.

The two phrases which are required to do the work if section 424 (1) is to be applied are ‘knowingly a party to the carrying on of the business in the manner aforesaid’ and ‘carried on recklessly or with intent to defraud creditors of the company or creditors of any other person or for any fraudulent purpose’

In Philotex , supra Howie JA said of the phrase: ‘knowingly a party’ :

‘Knowingly means having knowledge of the facts from which the conclusion is properly to be drawn that the business of the company was or has been carried on recklessly; it does not entail knowledge of the legal consequences of those facts….It follows that knowingly does not necessarily mean consciousness or recklessness. Being a party to the conduct of the company’s business does not have to involve the taking of positive steps in the carrying on of the business; it may be enough to support or concur in the conduct of the business (at 143 B – C).

In dealing with the concept of recklessness Howie JA cited with approval the dictum in Shawinigen v Vokins and Co. Ltd [1961] 3 ALL ER 396(QB) at 403 F in which the court defined recklessness as ‘gross carelessness – the doing of something which in fact involves a risk, whether the doer realises it or not; and the risk being such, having regard to all the circumstances, that the taking of that risk would be described as “reckless”. (cited by Howie JA at 143B).

Howie JA held that the test for recklessness is objective as far as respondents actions are measured against the standard of conduct of the notional reasonable person and subjective insofar as the court has to postulate that notional being as belonging to the same group or class as the respondent, moving in the same spheres and having the same knowledge or means to acquire such knowledge. (at 143 G-H)

Recklessness in this context is not limited to the more onerous test inherent in the concept of dolus eventualis. Gross negligence without a conscious or wilful regard of the consequences will be sufficient to bring a respondent within the scope of the section. See in general The Law of South Africa (volume 4 part 2(1^st reissue) by MS Blackman at para. 167. In determining the concept of recklessness within the context of section 424, a court must be careful not to adjudicate the disputed conduct with the benefit of the wisdom of hindsight. Ozinsky v Lloyd 1992(3) SA 396(C) at 414.

In determining whether the respondent has behaved recklessly, the scope of the operations of the company, the functions and powers of the director, the range of his or her responsibility, his or her background and knowledge and expertise, are all relevant in examining the director’s conduct .Within the context of this test,. it is well to remember having regard to the complexities of a large business organisation, that a director may properly leave certain tasks to a competent official who , in the absence of any reasonable ground for suspicion, can be trusted to perform such a duty honestly . in Re City Equitable Fire Insurance Co. [1925] CH 407 429; Norman v Theodore Goddard [1991] BCLC 1027: Paul Davies Gower’s Principles of Modern Company Law (6 ed) at 644; Fisheries Development Corporation v Jorgensen and Another; Fisheries Development Corporation of SA Ltd v AWJ Investments (Pty) Ltd and Others (1980)(4) SA 156(W) at 166 G-F; Lawsa para 167

‘Knowingly being a party to the carrying on of the business’ does not mean that knowledge of the facts is sufficient to bring a respondent within the scope of the section. The wording of section 424 indicates that the party must have taken part or concurred in such business practice. A director may well be a party to reckless or fraudulent conduct of a company’s business ,even in the absence of some positive steps taken in the in carrying on of the company’s business provided that her attitude to the impugned practice can, on the facts, be interpreted as condonation of or concurrence in the conduct which is the subject of the section 424 dispute. Lawsa at para 169.

The VAT Scam.

Before analyzing the nature of applicants’ case against first respondent, it is important to examine that which is common cause, namely the nature of the VAT scam. SourceCom imported Microsoft licences for resale in South Africa. In the ordinary course of a taxable supply as defined in the VAT Act, input VAT is payable to a supplier in respect of any taxable supply made by a registered VAT vendor. However VAT on imported goods is normally computed by the Department of Customs and Excise on goods entering South Africa and would be payable by the importer. Input VAT cannot normally be claimed without a tax invoice. It is common cause that the Department of Customs and Excise did not issue Source Com with any tax invoices in respect of imported Microsoft licences .Thus SourceCom was not entitled to claim any input VAT in respect of purchases from Microsoft Ireland Operations Limited (‘Microsoft IE’) as it did not pay VAT to Customs and Excise, the South African Revenue Services (‘SARS’) or Microsoft IE. According to SourceCom’s accounting records, invoices to Microsoft IE were first processed in September 1998 as a result of the award to SourceCom of a SARS’ tender to manage its Microsoft software licences in terms of a ‘Microsoft Enterprise Agreement’ .The tender was for specified software licences of 9000 computers over a period of three years. The total contract value over three years was R4 845 055.60 including VAT and a markup of 4% on SourceCom’s cost. In terms of this tender SourceCom purchased the requisite licences from Microsoft IE and sold these to SARS.

On the order date, employees of SourceCom performed the following processes:

1. A purchase order was generated which did not cause any accounting entry but created a system generated order number which was used to place the order with Microsoft IE.

2. All Microsoft IE orders were executed by SourceCom over the internet.

3. After the order was placed with Microsoft IE a Goods Receiving Note (GRN) was processed by the receiving clerk at SourceCom.

4. The GRN caused the following accounting entry:

Dr: Inventory Account (excluding VAT)

Cr: purchase accrual account (excluding VAT).

5. This entry raised the accrual for the purchase in the accounting records. No VAT was raised by this entry.

6. VAT appeared on both the SourceCom generated purchasing order and the GRN but this was a notional VAT number.

7. On receipt of the supply invoice the creditor clerk processed the invoice in the creditors’ ledger which caused the following accounting entry:

Dr: VAT

Dr. purchase accrual account.

Cr. Creditors’ account (including VAT).

According to the Pastel accounting system which SourceCom employed, when an invoice in the creditors’ ledger was processed , an adjustment could be made in respect of the VAT amount or the purchase amount. Therefore the VAT that appeared on the GRN or the purchasing order was not necessarily the VAT ultimately processed to the VAT account in the general ledger and hence not necessarily the amount claimed on the VAT return.

In summary , the VAT fraud involved two basic components:

1. The raising of fictitious input VAT on goods supplied by Microsoft to SourceCom in circumstances where the latter did not in fact owe or pay the VAT to Microsoft or to SARS .

2. The subsequent (or sometimes simultaneous) elimination of the fictitious liability or input VAT by journal entries affecting the income statement.

The raising of fictitious input VAT occurred in one of two ways:

1. by adding VAT when processing the purchase order and GRN or

2. by subsequent journal entries .

The first method was carried out as follows: The Microsoft invoice would be in dollars, (say $150.) The Microsoft invoice would not reflect VAT since Microsoft was an Irish incorporated company. A SourceCom employee would process an internal purchase order and the GRN. These documents would reflect the rand equivalent of the Microsoft dollar amount at the processing date (say R1000) and the Pastel system would automatically add and reflect 14% (R140) as an additional amount owing to Microsoft as input VAT.

The accounting entries would appear as follows:

Dr: Stock R1 000

Dr: VAT R 140

Cr: MS(Microsoft) R1 140

The effect of these entries would be to understate SourceCom’s liability to SARS for VAT by R140 (since the debit of R140 reduces its net liability to SARS for output VAT) and to overstate its liability of Microsoft by R140. At this stage the effect on the balance sheet is neutral, and the entries have no bearing on the income statement.

As regards the second method:

No VAT is added when the purchase order and the GRN is processed, so that at this stage (correctly) the account entries are:

Dr: Stock R1 000

Cr: MS R1 000

Subsequently a journal entry is made manually. Generally when this was done, the price of R1 000 was treated as notionally being inclusive of VAT, so that the fictitious VAT should not be R140 but R122,81 (14/114 x R1000). This fictitious input VAT (R122,81) would be debited to the VAT account and credited to the MS account or to the sundry creditors account.

If due to depreciation in the rand, the amount ultimately paid to Microsoft was not R1 000 but (say) R1 150, a further GRN would be processed for the difference (R150), and sometimes fictitious input VAT was also raised on this further amount.

The writing off of this fictitious VAT to the income statement was achieved by journal entries.

The bulk of the fictitious VAT was written off against cost of sales, though other income and expense accounts were also utilised. Illustrative journal entries would be:

Dr: Microsoft R140

Cr: Cost of sales R140

OR

Dr: Sundry creditors R122,81

Cr: Cost of sales R122,81

The effect of these entries would be to eliminate the fictitious liability to Microsoft (or to sundry creditors) for the bogus input VAT by reducing cost of sales.

There were instances where the first and second components of the fraud were combined, the fictitious input VAT being raised directly against cost of sales, for example:

Dr: VAT R122,81

Cr: Cost of sales R122,81

The effect of such entries was twofold:

Balance sheet: Instead of a fictitious liability to Microsoft (or sundry creditors) for input VAT, the accounts should have reflected a true liability in the same amount to SARS. By writing off the fictitious liability to Microsoft (or sundry creditors) without raising the true liability to SARS, Sourcecoms’ liabilities were understated and its NAV overstated.

Income statement The crediting of cost of sales (or any other expense or revenue account) would falsely reduce expenses (or increase income), thereby falsely overstating gross profit.

According to a report prepared by auditors KPMG on behalf of applicants, Source Com wrongly claimed R12 378 662 of input VAT from SARS for the period 1 October 1998 to 30 June 2001. The yearly amounts can be reflected thus:

Year End	Rand
30 June 1999	3 122 595
30 June 2000	4 482 917
30 June 2001	4 773 150
Total	12 378 662

The mechanism and the nature of this scam is not in dispute. Ms Honey and second respondent confirmed the fraud and their involvement in it. The dispute does not concern the question as to whether the VAT was wrongly raised by way of negligent oversight or by fraud. The fraud occurred. As Mr Rogers, who appeared together with Mr Goodman on behalf of applicants, correctly submitted the key question was whether first respondent was informed of the VAT scam and gave her approval. The basis of applicants’ case is that the inherent probabilities were that the fraud would not have occurred without her knowledge and approval.

In order to prove that first respondent’s conduct falls within the scope of section 424 ,applicants sought to rely upon the direct evidence of second respondent who claimed that first respondent was party to the scam and secondly on inferential reasoning based upon the financial position and records of SourceCom which, in applicants’ submission, admitted of but one conclusion on the probabilities, that she must have known about the fraud and was a party thereto.

Applicants’ case.

Applicants relied upon two witnesses, being second respondent and Ms Honey. Second respondent's evidence concerning the VAT scam emanated from an affidavit to which he deposed on 25 September 2001 and which was annexed to applicants replying papers (‘the replying affidavit’) , a statement made in terms of section 112 of the Criminal Procedure Act 51 of 1977 on 16 November 2001 for the purposes of a criminal case against him, a witness statement prepared for the purposes of the oral evidence in the present dispute and oral evidence which he gave before this court. As Mr Burger, who appeared together with Mr Fagan on behalf of first respondent, made much of the contradictions between the four sources of second respondent’s evidence, it is necessary to examines separately each of these sources of evidence.

In the replying affidavit second respondent said that during November 1998 Ms Honey had come to see him. She was concerned about the prospect that SourceCom would shortly have to pay an amount of some $2.5 m to Microsoft (IE) in respect of the first of the large Microsoft invoices which , in September 1998, had been submitted to SARS ,in its capacity as customer of SourceCom. According to second respondent Ms Honey explained that, in addition to the amount payable to Microsoft , there was a sum in excess of R2 due to SARS in respect of VAT. In the context of discussing SourceCom’s cash flow problems, Ms Honey suggested that SourceCom could claim the input VAT on the Microsoft (IE) (supplier) invoices by converting the dollar amount reflected in such invoices to rands, adding on the VAT and submitting the adjusted invoices to SARS.

Second respondent advised Ms Honey that he would have to discuss the matter with first respondent which he then did. He and first respondent devised a strategy in the event of a query from SARS. SourceCom would claim that its bookkeeper had made a mistake and thereafter would pay over the moneys due to SARS. It was decided that the VAT scheme would be limited to large annual and bi-annual invoices furnished to SARS, the Provincial Administration of the Western Cape and the Gauteng Provincial Government. Respondent confirmed that he had conveyed first respondent’s approval to Honey.

According to second respondent, prior to his leaving SourceCom in November 2000 , he had a discussion with Honey to the effect that the new shareholders were taking over control of SourceCom and that the VAT scam should come to an end.

In his section 112 statement, second respondent said that during November 1998 SourceCom paid an amount of approximately R14 m to Microsoft (IE). VAT had incorrectly been added to SourceCom’s purchase notes as the accounting system automatically calculated VAT on all suppliers invoices. Second respondent discussed the situation with first respondent and it was decided not to correct the artificial credits or the accounting system. At the time SourceCom was in a poor financial position owing to the loss of the Telkom contract during mid-1998.

In his witness statement second respondent suggested that Ms Honey did not speak to him in November but probably in the first half of December 1998. By this time VAT on the three Microsoft invoices had already been raised automatically by the accounting system when the purchase notes were generated. The incorrect VAT, totaling R2.323.953,42 had then been reversed by Mary Percent ,a SourceCom employee , on the instructions of Honey. When Honey and second respondent spoke, the main focus of their discussion concerned the large VAT payment which would be due to SARS. This large payment had arisen from the fact that the notional input VAT which was incorrectly raised by the system had been reversed on Honey’s instructions. According to first respondent, ‘I cannot now recall whether, at the time Honey spoke to me, SourceCom had already paid Microsoft although the currency had apparently been purchased on 12 November 1998. I very much doubt that SourceCom would have paid Microsoft on that date, as I would have ensured that he paid Microsoft on the last possible date (which would have been 30 November 1998). Either way, the fact that he had paid Microsoft or would shortly have to pay Microsoft would have placed strain on SourceCom’s cash position.’

Second respondent then referred to the manner by which the subsequent journal entries were inserted to ensure that the VAT credit was eliminated in the accounting system. He confirmed that procedures were not ‘generally discussed with the First Respondent in advance, though the fact that they had been made came up from time to time in discussions between the First Respondent and me .in the context of the management accounts. For example, she might enquire why the gross profit in a particular month was so high and I would then explain (for example) that some of the fictitious Microsoft VAT had been written off against cost of sales’.

When he gave oral evidence , second respondent’s version of the conversation which took place between Honey and himself was carefully probed by Mr Burger on behalf of first respondent. The following extracts reflect the version which he now offered to the Court.

‘….and as we read on you say in November 1998 Honey came to see me and we discussed the fact that SourceCom would shortly have to pay an amount of some $2 ½ million over to Microsoft. We know that is wrong. Your latest version is that what she came to discuss with you was the R2 m VAT? It would have been the same conversation.

No, it is not the same conversation because we know by the beginning of December Microsoft would have been paid? Yes

this is just a lie, Mr Connolly , it could not have been the discussion of $2 ½ m to be paid to Microsoft?….I was obviously confused.

It was a lie at this time and I will tell you why. On your version the foreign exchange was bought on 12 November 1998 to pay Microsoft, is that correct? That is right, yes. As you buy the foreign exchange it goes out of your bank, you pay for it, correct? Not necessarily

In this case it did? I see, it appears to have been done, yes.

Yes it is just nonsensical to suggest that the lady would in November….in November 1998 have come to discuss with you the payment of $2 ½ m to Microsoft..It just could not be?……

You take some time to answer that ?. There could be two different meetings which were confusing.

Well, that is the third version we have.

Let me give you the second version. Go to your latest statement that you have just read to the court in your evidence in chief, that is exhibit D (the witness statement)….read us the second sentence please? The main focus for the discussion concerned the large VAT payment which was – which would be due to SARS. We have got three versions. You have either discussed the Microsoft payment or you have discussed the SARS payment or you have had two meetings which one do we accept now? – No there have been several meetings all the time during November, every day, you know the exact date and time of meeting I would not be able to record to know exactly what was said in those meetings….

R8,2 m in the bank, I put it to you there is no cash flow problem in this company in this time frame ? It does not look like it. …

Yes, there is no cash flow problem but you discussed the cash flow problem with her? Yes I could have but that is nonsense we know there is no problem. What are you discussing with her? Problems is probably the wrong word to be used, cash flow issues maybe.

What cash flow issues? There are always cash flow issues, monthly, yearly always cash flow issues when to pay creditors, when to collect money or not to collect money….

You suggest that the lady who has just been appointed on probation in a new company suggests to you a VAT fraud against the backdrop of the cash flow problems debate we have just listened to? .? I here two different issues.

Why is it a different issue? One is the claiming of the VAT from Microsoft invoices and we should not have on the fraud side, and one is the cash flow implications of it.

No I do not think you follow my question. I am probing you with something much more basic. What would Ms Honey’s motive have been for making this suggestion to commit a fraud to you. What is in it for her? I do not know I’ve never asked her….

This lady says to me why do we not commit a massive VAT fraud. I say to you, you are an intelligent person. You say to her are you actually out of your mind. What does she want to do this for. You ask her, or you do not ask her? In the context the way it was said, it was not to commit a massive VAT fraud, let us claim the VAT off the Microsoft invoices.

You just steal R2 m from the fiscus? Ja

Why did you not terminate her services on this – this suggestion of fraud? No reason to….I do not know why.

In fact you appointed her permanently thereafter? Ja’

When he was cross-examined about first respondent’s knowledge of the scam, the following aspects of second respondent’s evidence are pertinent:

‘You see I do not understand the statement in your latest document which is exhibit D in paragraph 6 (the witness statement)….you say for example she, might enquire why the gross profit for a particular month was so high, and I would then explain for example that some of the fictitious VAT has been written off against cost of sales. Strangely she says to you, why are our profits so high, and you say to her, well you remember we have this

VAT fraud going and that is the VAT fraud kicking in… Its an unreal discussion between the two of you. Can you explain to me how this discussion would take place? Well Sandy would obviously know and look at every single sales figure weekly and she will know exactly the GP had been made in the company.

Yes? And when it comes to the end of the month GP will be slightly higher than she anticipated. Yes? And I’ll explain to her that was released with some of the visions(?)

we had and…..were phased into the profits.

But I thought she knew that you were releasing fraudulent figures into the sales figures? Ja.

Well then why ask about it if she is the…..- a confirmation – a confirmation of exactly what happened.

What does she confirm?. She says I know the profits must be higher I ask you why – why they are higher. Will you please confirm are they higher because of our fraud?…Ja, but some months we did not release anything into the income statements. Sometimes we kept it back. It was not a case of taking it the month it happened and reversing all the profits. Can you remember a particular month when this enquiry was raised because from reading your statement….I cannot recall.

Its seems to me that your memory again is leaving you in the lurch…..I could not recall the exact month and year.’

Second respondent was also asked a series of questions about any attempt which first respondent might have initiated to cover up on the VAT fraud when she left Sourcecom . The following passage reflects second respondent’s response:

But she made no effort on your evidence to cover up or to defuse the discovery of this VAT fraud….I don’t understand the question.

She made no effort as far as you know to cover her tracks on the VAT fraud when she left by the end of 2000 –no’.

Second respondent testified that he had instructed Ms Honey to stop the VAT scam upon his resignation from the company and that further he was a very concerned person as to the possibility of the scam being discovered. He confirmed that he made no effort to check that his instruction to stop the VAT fraud had been implemented.

Ms Honey also testified before the court. She confirmed the contents of an earlier affidavit to which she had deposed. She testified that Jessica Augustus who was the bookkeeper responsible for completing the VAT 201 forms had approached her during December 1998 when she was in the process of completing the VAT 201 return. Augustus told Honey that ,according to her calculations, SourceCom had to pay an amount in excess of R2 m to SARS. As this was an unusually large VAT liability Honey discussed the matter with second respondent. Honey explained to second respondent that the credit clerk, Mary Percent, had approached her regarding the problem of the Microsoft account reflecting an unexplained large credit balance. According to Honey she found that VAT had erroneously been added to the purchase notes of Microsoft. The system had automatically calculated VAT on suppliers invoices but as Microsoft was an overseas supplier VAT should not have been so calculated. Honey instructed Percent to rectify this situation.

Second respondent then suggested to Honey that he would discuss the matter with first respondent on return to her. On the same day he returned and instructed Honey that the journal entry should be reversed. Honey did not accept this instruction and told him that she disagreed. In her affidavit she said ‘I also told him that we were not entitled to claim these amounts as input because of the fact that it was purchases from an overseas supplier which had no VAT implication. He persisted with the decision the journal entry be reversed as well as that we must claim the calculated amounts as input VAT. I knew that my argument was correct but had to carry out the instruction as he was “the boss’.

Ms Honey also confirmed that ,contrary to the evidence deposed to in an affidavit by Ms Carol Hendricks, a forensic accountant instructed by first respondent to investigate the VAT transactions, the new Pastel system also needed a manual intervention to add 14% VAT to the purchase note of Microsoft IE .

Ms Honey was not able to confirm that she had any recollection of a conversation which took place between herself and second respondent to stop the fraudulent VAT scheme upon the latter’s resignation from SourceCom. She also confirmed that the VAT fraud continued after the resignation of second respondent from SourceCom, that is during the early period of Mr Maritz’s term of office. Mr Maritz replaced Mr Connolly as the financial officer from December 2000. Ms Honey confirmed that the auditors performed an analytical review of the business of SourceCom and its accounts, that, in her view, they would have examined turnover, profit history, gross net profit as well as the ratio engaged in a ratio analysis to test the veracity of the accounts.

Apart from the evidence of Honey and second respondent, Mr Rogers relied heavily on a process of inferential reasoning sourced in the financial statements of SourceCom and which justified, in his view, a conclusion that, on the basis of these accounts ,first respondent must have known and indeed did know of the fraud which had been perpetrated. It thus becomes necessary to examine the impact that the VAT fraud on the business by way of an analysis of key financial entries.

Distortion in the accounts caused by the VAT scam.

Mr Rogers developed a careful analysis of distortions in the reported profits caused by the Vat scam from September/November 1998. For the purposes of understanding the process of reasoning employed by Mr Rogers, the most significant distortions to be examined are those which began in September 1999.

In September 1999 a fictitious VAT input on Microsoft purchases was raised in the amount of R1,580,355. This amount was immediately written off against various accounts including a credit of R823,294,13 to cost of sales.

The consolidated income statement as at 30 September 1999 reflected a gross margin of R1,560,423 of which the bogus reduction in cost of sales constituted about 53%; a very significant figure in the context of the September 1999 accounts.

An examination of the accounts revealed a series of different indications as to the nature of the distortion:

1. The gross margin percentage in the consolidated income statement was 9,96% as against the budgeted figure of 6,7%.

2. The figures for the Cape Town branch contained a curious credit provision of R807,519 under the cost of sales and against the narration ‘Other’ This entry had the effect of yielding a gross profit percentage for the Cape Town operation of SourceCom of 73,9% as against the budgeted figure of 14,4%. The credit entry of R807,519 was primarily attributable to the bogus journal entry of R823,294,13.

3. The Johannesburg branch figures reflected software sales of R11,911,467 and software cost of sales of R11,528,166 which figures reflected a gross profit margin of software sales as 3,86%.

The sales of software by the Johannesburg branch derived almost entirely from the sale of Microsoft licences to SARS, at a gross profit percentage to 3,85%. According to Mr Rogers, first respondent would have clearly been aware of this large transaction and of the low margin at which it was concluded. If the Johannesburg software sales and the related cost of sales were eliminated from the figures contained in the consolidated income statement, the remaining sales of R3,674,024 were generated at a cost of sales of R2576,902 yielding a gross margin of R1,097,122, representing a gross margin percentage of approximately 30% gross margin in a period where professional services and technical support ,which may well have been marked up considerably, contributed little to the business. A gross margin of percentage of 30% on the business transacted during the period was thus inexplicable.

Mr Rogers also referred in this connection to first respondent’s testimony, the following extract of which is of particular relevance:

‘The big picture in shortly after September 1999 was that you had done a contract at just under R12 m with SARS at a known margin of 3.85%. That was a September transaction? Yes. You now get an income statement which you look at which has total sales of R15,6 m. You now know that just under R12m went to SARS at 3,85% but you see that you are achieving a margin on the whole lot at 9.96%. Does that fit the big picture? When you break it down like that, no. When you look at the figures as a whole it once again is not something that jumped out at me but I….in the previous month the margin was 11%. I know when you isolate and say I should have known that September was SARS and at this percentage and therefore I should have checked this and said, hang on this does not make sense, I do not do that Mr Rogers that is why I had a financial director.’

In her answering affidavit first respondent had contested the conclusion drawn by applicant that the sale of software licences to SARS had been accepted at a gross profit margin of approximately 4%. She said ‘the fundamental premise of the theory’ advanced by applicant was that the margin on the SARS contract was 4% whereas in fact it was 7,85% and that therefore applicants’ submissions had no merit. However under cross-examination she was forced to accept that the transaction in September 1999 had indeed been done at 3,85% and that it was only in December 1999 and January 2000 that subsequent exchange variations resulted in the ultimate margin increase to 7,8%,. changes which had not impacted on the September 1999 accounts.

In attempting to explain the inaccuracy contained in her answering affidavit, first respondent referred to the supporting affidavit of Ms Hendricks, the forensic accountant who had calculated a margin of 7,85%.

The relevant cross-examination reads as follows:

But it is something more than agreeing with her, because of course Ms Hendricks was not the lady who was there at the time. She can perhaps look at historical documents and do calculations and eventually when she looks at cost reports that go into the next year, she finds that this is that eventually after currency fluctuations were taken into account it ended up eventually at 7.85% but you were there at the time in September 1999, she was not. Correct? Yes.

And you knew that the September 1999 management accounts would have reflected the deal at the price it was done, that is 3.85%? At the time that I did this affidavit ? Yes – No at the time I did this affidavit which was a good eighteen months or two years later I was relying on the documentation I had and that she worked on and as it turns out this particular deal was done at 7.8% overall. The fact that that does not apply to September I agree with you but I was not aware of that at this time. I had not applied my mind to it’.

When first respondent was then probed as to whether these figures would not have created some need for further enquiry when they were first presented to her the relevant passage in her cross-examination reads as follows:

‘And the only explanation you can offer for not having said when you had the first opportunity to refute this case that the reason why you did not, the alarm bills did not go off because you had never looked at the Cape Town figure. The only explanation for that is you obviously did not apply your mind carefully enough in doing your own study? Yes, in terms of checking what Ms Hendricks had done, and that the information given to me is correct. Yes I had not.

The consolidated income statement reflected the Cape Town sales in the amount of R1,325,851 , effected at a cost of only R346,679. This produced an enormous gross margin which would have been completely out of kilter with the profit margins reasonably expected in the SourceCom business.

Mr Rogers referred to evidence given by first respondent in re-examination which indicated that she did in fact, at other times, look carefully at branch figures. When Mr Burger referred first respondent to management accounts as at 28 February 1999 where there were gross margins of just under 40% compared to the budget figure of 10,42% she admitted that these figures caught her attention and that ‘the explanation was that there was again substantial professional services in that month, it came to R230 odd thousand and if I excluded VAT then the margin became like 21% of the hardware sales.’ According to Mr Rogers these figures for professional services could only have been gleaned by first respondent after she had examined the branch reports.

The relevance of this evidence was that, when profit margins were significantly higher then the norm , first respondent was able to explain these figures by reference to the contribution of the technical and professional services which she could only have done by looking at the branch reports. According to Mr Rogers , her explanation that she only looked at the ‘big picture’ and not at the details contained in branch accounts was, at best ,a selective reflection of her conduct.

In April 2000 a fictitious VAT entry of R1, 649 349,91 was raised on a Microsoft order and simultaneously written off against cost of sales. Total sales for the month were R18,706, 849 the total cost of sales of R16,677,780 yielding a gross profit margin of R2,029,069 (10,85%). The bulk of these sales comprised software sales of R15,740,186 at a supposed cost of R13,923,840.

Included in the April 2000 figures was a sale of Microsoft licences to the Gauteng Provincial Government. The tender was in the amount of R15,458,970. According to Ms Van Wyngardt ,who prepared the tender, she was instructed by Mark Carrick ,the national sales manager at SourceCom , to submit the tender on the basis of a 2,5% markup. This was confirmed by Carrick in an affidavit to which he deposed. According to the KPMG report Microsoft licences were acquired at a cost of R13,430,380 on which figure a fictitious input VAT of R1,649,345 was claimed.

When the Johannesburg branch accounts were examined a software sales figure of R15,657,579 was included in the accounts at a cost of R13,794,249 which yielded a gross margin on the sales of 11,9%.

The overall figures of SourceCom for the relevant period as reflected in the consolidated income statement , ( period ending 30April 2000) reports the gross margin percentage on software sales as being 11,54%, most of which was derived from the sales in the Johannesburg branch. According to Mr Rogers these figures did not make sense to a person who knew (as ,in his view ,did first respondent) that the vast bulk of the software sales derived from a contract where the markup had only been 2,5%. If it assumed that the Gauteng government sale had been in the amount of R15,458,970 as per the tender and if this represented a price inclusive of a 2,5% markup, it would have been expected that the related cost of sales would have been R15,081,922 with a profit of R377,048.

On further examination of the statement if the Gauteng sale , if R15,458,970 is deducted from the total software sale for the period reflected of R15,740,186 and the related cost of sales of R15,081,922 is deducted from the supposed software sales of R13,923,840 the remaining sales of R281,216 were generated at a negative cost of R1,158,082, a patently impossible situation.

Mr Rogers offered an alternative explanation. He contended that the Gauteng sale could be divided into two components .On the assumption that the software component was R13 430,380 plus the markup of 2,5% (yielding R13,766,139) , the balance of the Gauteng sale (as a 3,5% markup) would amount to R1,692,831 made up of the cost of R1,635,585 plus the markup of R57,246. On this basis the total cost of sale for the Gauteng contract would be R15,065,965 which would mean that the remaining sales of R281,216 was produced at a negative cost of sales of R1.142,125 which again would be nonsensical.

Mr Rogers submitted that ,although these calculations demonstrated the distortions in precise numerical terms, it was not necessary to know the precise details nor to do the calculations in order to realise that ,on the basis of the tendered amounts, something was seriously amiss with the figures reported in the income statements. First respondent should have known that the vast bulk of the software sales disclosed in the consolidated income statement were done at the markup of between 2,5 and 3,5% ; hence the cost of sale figure reflected in the consolidated income statement would have made no sense.

Mr Rogers attacked first respondent’s explanation in her answering affidavit that ‘she did not analyze each line item on the pages behind the management account summary page but simply checked to see if the numbers were consistent with budget and that the revenue from the new deals done.’

This explanation according to Mr Rogers was seriously flawed for a number of reasons inter alia , that the budgeted gross margin on all sales was 10,5% against an actual figure of 10,85%. Whilst this discrepancy might be considered to be extremely modest , the budgeted gross margin of 10,5% assumed very significant hardware sales where higher margins were generally gained than would be the case with software. as well as very large sales of professional services and technical support which yielded substantial margins. Bearing in mind that sales in these categories were way below budget and that the primary contributor to sales was software, a gross margin of 10,85% ,essentially sourced in sales which could not have expected to earn a significant margin, would have been inexplicable to a person with the knowledge of the business as could reasonably be expected to have been the case with first respondent.

In the management accounts for June 2000 the consolidated income statement reflected total sales of R19,400,812, total cost of sales of R17,061,280 yielding a gross margin of R2339532 at a gross margin percentage of 12,06%. With the addition of ‘other income’ of R829,065 , the accounts reflected a gross profit for the month of June 2000 of R3,168,597. After deduction of the overheads, operating income was R1,627,255 for the month. Mainly , as a result of this significant amount of operating income for the month of June 2000 , the full year reflected an increase in operating income from R633,778 at the end of May 2000 to R2,448,360 at the end of June 2000. According to notes which were found in first respondent’s diary and confirmed in her evidence, she targeted operating income for the year ending 30 June 2000 as being approximately R2,5 with the targeted gross profit for the month of June 2000 being R3,25 m. Management accounts for June 2000 reflected that these targets were almost reached.

During the month of June 2000 fictitious VAT of R529,579,94 was raised on Microsoft orders on the 26 June 2000 and 30 June 2000, respectively, which amount was immediately written off against Cape Town hardware cost of sales. In addition , an amount of R470 000 in respect of fictitious VAT previously raised was released into the income statement during June 2000 by crediting R370,000 to Cape Town hardware cost of sales and R100,000 to Johannesburg hardware cost of sales. The effect of these figures is reflected in an examination of the relevant Cape Town accounts. Cape Town hardware sales of R6,019,209 were ostensibly achieved at a cost of R4,503,263 , giving a gross margin of R1 515,946 at a gross profit margin percentage of 25,18% compared to budgeted figures which reflected a gross margin of R584,000 at a gross margin percentage of 8%. Given that the standard margin of hardware sales was between 8-10% a margin of 25,18% certainly should have raised queries in the mind of a knowledgeable reader of these accounts.

Employing these discrepancies in the accounts, Mr Rogers sought to make out the following case.

1. First respondent was a highly qualified and experienced business woman. She had obtained a B.Bus Sci degree from the University of Cape Town with honours in finance. She was awarded a higher diploma in tax law from the University of the Witwatersrand. Between 1983 and 1986 she was a part-time lecturer at the University of the Witwatersrand’s Business School where she lectured in tax law. She worked as a tax consultant for accountants Arthur Young and Company during 1983 .Between 1984 and 1988 she worked for accountants Fisher Hoffman Stride where she was appointed as a tax director during which period she was involved in the development of various aggressive tax saving schemes on behalf of clients. From October 1988 to October 1992 she was an owner and managing director of a project finance company and then an owner and the managing director of an information technology company before joining SourceCom as chief executive officer and a significant shareholder to July 1996. In short , Mr Rogers suggested that she was an extremely financially literate .

2. According to second respondent she was a formidable ‘boss’, a micro-manager who carefully monitored the entire progress of SourceCom.

3. Between the period September 1998 to November 2000 more than R10 m was fraudulently gained by SourceCom as a result of the VAT scam.

4. She was a significant beneficiary of the VAT fraud, given her share holding in SourceCom. The company which she controlled , Elect Props (Pty) Ltd initially owned 49% of the old SourceCom company. Old SourceCom sold its business to the new and current SourceCom, where her shareholding was reduced to 20,9%. Elect Props also had a loan account with SourceCom in the amount of R10,676,393. Thus , the value of her investment depended to a large extent on SourceCom’s profitability.

5. When the business was sold to the new SourceCom she gave certain warranties during March 1998 before the loss of the critical Telkom contract which she conceded was a material blow to the company.

6. Therefore she had very significant motives for promoting the VAT fraud

7. Given her expertise , experience and knowledge of the company’s operations it was inherently improbable that she would not have detected the very significant distortions in the monthly accounts to which reference has already been made. Given SourceCom’s sales , she must have realised that very large VAT payments between the period October to December 1998 would have had to be made and accordingly she would not have failed to realise that such payment had not been made.

8. It was inherently improbable that a junior employee ,Honey, and second respondent on their own would have considered that they could perpetrate a fraud of this magnitude without her detecting it and hence without her co-operation.

9. The inherent probabilities were against a conclusion that, for over two years, a person with such expertise and skill would not have picked up the VAT scam.

10. In any event the contrary conclusion is supported by the evidence of second respondent and Honey.

11. First respondent comes across as a manifestly intelligent, quick witted and slick witness who sought to portray herself as a person of great integrity. However the evidence indicates to the contrary. Mr Rogers referred in particular to an application which she had made to the Department of Home Affairs for a temporary passport in which she had submitted a letter to the Department from her former husband on the letterhead of his company, Regis SA (Pty) Ltd, in which it was stated that she was an employee of that company and needed the passport to travel on urgent business for the company. This was a blatant untruth. Furthermore she did not declare a bonus of R250,000 in an income tax return which she had received in the tax year ending February 1999.

12. On the basis of this holistic approach to the oral testimony and to the clear inferences that could be drawn from the anomalies in the accounts, the only reasonable conclusion, on a balance of probabilities, that could be drawn was that first respondent , at the very least, had known and acquiesced in the VAT scam from the moment that it had been initiated. Whatever the criticism of second respondent’s performance in the witness box might be (and applicants did not suggest that he was a good witness) his version, namely that the scam had been discussed between himself and first respondent was a far more probable explanation than the alternative, namely that a junior employee and the financial director had done this all on their own for the benefit only of the financial director and Sourcecom .

Mr Burger sought to counter applicant’s case by attacking the value of the evidence given by second respondent, and secondly by criticizing the process of inferential reasoning relied upon by Mr Rogers in constructing applicants’ case. Mr Burger submitted that by the time second respondent came to give oral evidence before the court, he had contradicted himself in material aspects as a result of the evidence provided in his affidavit to which he deposed on 25 September 2001 and which was annexed to the applicants replying papers, his section 112 statement and the witness statement to which he deposed prior to giving oral evidence. These various statements threw up different versions of the following:

1. The time of the discussion which second respondent had with Ms Honey regarding the VAT scam in late 1998.

2. The question as to whether, at the time when Honey and second respondent discussed the VAT scam, SourceCom had already made a significant payment in late 1998 to Microsoft or whether it still required the cash in order to discharge its obligation.

3. Whether there was a discussion between second respondent and Ms Honey concerning a previously incorrect VAT add on to Microsoft’s (IE) purchase notes.

4. Whether on each and every occasion that a fraudulent journal entry was to be made or had been made, second respondent advised first respondent accordingly or whether such fraudulent journal entries were not generally discussed in advance with first respondent.

Second respondent had advanced a different version on each of these issues in the three written documents to which he deposed. To the extent that there were contradictions between the versions of Ms Honey and second respondent the former was to be preferred because she was by far the more consistent and believable witness.

Mr Burger focussed his analysis on a number of aspects of second respondent’s oral testimony, in particular:

1. Ms Honey confirmed she had discussed the erroneous calculation of VAT with second respondent and that he had ordered her to reverse the instruction which she had given to Mary Percent to correct the situation whereby incorrect VAT input credits were claimed. Second respondent was extremely vague , indeed incoherent with regard to the contents of the meeting which took place between himself and Honey. At one point he suggested that there had been three meetings between himself and Honey and that serious cash flow issues had been discussed between the two of them , notwithstanding that he later accepted that the foreign exchange necessary to discharge the Microsoft debt had already been paid by the time the meeting took place. He also sought to place the blame on Honey for the VAT fraud.

The following extract ( part of which has already been cited above ) from second respondent’s oral evidence reflects the manner in which he was asked about this version.

‘This lady says to me why do we not commit a massive VAT fraud, I say to you, you are an intelligent person you say to her are you actually out of your mind. What does she want to do this for. You ask her, or do you not ask her?…..In the context of the way it was said it was not to commit a massive VAT fraud, let us claim the VAT off the Microsoft invoices.

You just steal R2m from the fiscus? ….Ja, that is not serious in your eyes? Of course it is just the way it was said. It was not said in a serious fashion? No of course it was said in a serious fashion.

I am asking you….not the way you are putting it.

I am asking you, the lady suggests to you that you steal R2 m from the fiscus? No she did not say that we must steal R2 m from the fiscus.

OK….she says we must claim VAT on these invoices.

Yes, and that is a fraud, you know that? Yes I know before. You have pleaded guilty to that? Correct. Yes what is the motive you see, I want to argue at the end of the day that she had no motive, I am giving you an opportunity to suggest a motive? I do not know what her motive was. You did not ask her? No not that I can recall.

What I also did not understand is this, true? Why do you not terminate her services, she is on probation you look her over the first meaningful discussion you have with her, she suggests you steal R2 m from the fiscus?…..It was not.

What did you do – you appoint her permanently?…It was my first meaningful discussion with her.

Why did you not terminate her services on this – this suggestion of fraud by her? No reason to, no I do not know why. In fact you appointed her permanently hereafter, correct ? Ja’

2. In his evidence, second respondent classified first respondent ‘as a formidable boss and one who is obsessed with control’. Mr Burger submitted that it was critical to applicants’ case that first respondent be described as a ‘micro manager’ because that supported the argument that she must have known about the variations in the accounts caused by the VAT scam.

When he gave oral evidence, an affidavit deposed to by Ms Mandy Daddy, at one time the national sales manager of SourceCom, was put to second respondent; in particular Ms Daddy’s description of the management style of first respondent as being to afford considerable scope and autonomy to departmental heads so that she(Daddy) was merely required to report on a weekly basis as to whether there was any problem in her department.

The following extract from the evidence reflects his response : ‘Do you have any reason to challenge this as being true for Ms Daddy that this was her experience of Ms Thorne (first respondent)? Second respondent replied – no, not the way the sales people operated. ‘

Second respondent conceded that Ms Daddy occupied a post in the organisation at the same level as that occupied by second respondent.

Given the description of first respondent as a micro manager, Mr Burger submitted that he expected that first respondent would have been very careful regarding the termination of the VAT scam upon her resignation. But the following passage in second respondent’s evidence under cross-examination appears to suggest otherwise:

‘What did she (first respondent) do about this VAT fraud on which she was sitting? R8 m or R9 m having been stolen, running in false journals in the company , she now rides off into the sunset and she knows that there will be a disaster within weeks, the company is going to be liquidated, what does she do now to cover her tracks? I don’t know you would have to ask her that I don’t know.

She leaves the company and waits for the trouble to start coming – the only discussion we had was the question was asked to me by her that would anyone discover that fraud when we left. Oh and what did you answer? No I don’t think so.

You think you can both leave and nothing would be discovered? Well that was the opinion, Ja

And she accepted that and left…Ja.

Is she financially naïve – no.

She is not stupid? Absolutely not

But she made no effort on your evidence to cover up or to diffuse the discovery of this VAT fraud – I don’t understand the question.

She made no effort as far as you know to cover her tracks on the VAT fraud when she left by the end of 2000 – no.

While she believed that the company would have a liquidator by December of that year I won’t say liquidator, there will be in very dire straits.

Yes, and while the company is to be controlled by two innocent new purchasers, the manager of the company Dr Koornhof and Mr Harlem Correct? They manage the company, Yes.

Did you tell as you left, did you tell Ms Honey to stop the VAT fraud ? Yes I did’

Mr Burger emphasised the significance of the omission in any of the written versions put up by second respondent of any discussion which had taken place with first respondent regarding the termination of the fraud following upon their respective resignations.

3. Mr Burger also referred to second respondent’ s oral testimony regarding the manner in which he informed Ms Honey to terminate the VAT scheme: ‘And what did you tell her? That the VAT – the VAT scam that was happening in the company must stop immediately, because of the new shareholders involved. And what was he reaction, I can’t recall, it should be yes OK.

You can’t recall her reaction ?You either have a particularly bad memory or your answer is untruthful. You have a very bad memory? I don’t have a good memory, no.

So you have a bad memory, possibly that’s the answer but you don’t remember what her reaction was? Exactly no it would have been okay we must stop this.

Did you then ensure that it had been stopped from the time that you gave the instruction? I don’t know, I wouldn’t have – I would have assumed that she had stopped those were my instruction’

By contrast Ms Honey could not recall any conversation with regard to the cessation of the VAT scam. It appeared, on the basis of the written evidence and confirmed by Ms Honey , that the VAT scam continued after the resignation of both respondents.

4. Mr Burger also referred to the reason why second respondent might well have committed the fraud for his own benefit. In June 1998 he received a salary from SourceCom of between R15,000 – R18,000 per month. According to his oral testimony, at the beginning of 2000 he was required to pay R9,600 per month in respect of the bond on his house, R3,900 in respect of financial charges on the Combi and R3,000 in respect of finance charges on a BMW. His children were at an expensive school. In March 2000 the bank called up an overdraft of R240,000 . He was in serious financial difficulty.

Significantly, second respondent appeared to obtain very little advantage from the VAT fraud which he alleged he had initiated with the co-operation of first respondent. Whereas he had received a bonus of R126,000 for the 1998 financial year that is before the VAT fraud commenced his bonus for the 1999 financial year was substantially reduced to R30,000 and no bonus was paid to him for the 2000 year.

Mr Burger contended however that it could not be concluded that there were no potential financial benefits which might have accrued to second respondent as a result of the VAT scam. At one point he received a bonus of R126,000 and ,with an improved profit record ,SourceCom could award him again with a similar bonus. Secondly it was common cause that he sought a further reward through a shareholding in SourceCom and payment of R300,000 in consideration for concluding a covenant in restraint of trade..

Second respondent’s desperation for money went to even greater lengths. His ex wife Ms Adrienne Connolly, with whom second respondent was still living at the time, sought to extort payment of R1 m from first respondent and thereafter decided to expose the VAT scam to the authorities with a view to obtaining a 10% commission from SARS. Significantly first respondent made mention of this in her answering affidavit in which she stated ‘I understand that Adrienne Connolly alleged that she had a file in her possession evidencing a fraud on SARS and that she required payment of R1 m or she would report the matter to SARS and claim the reward. I immediately conveyed to Philip Connolly (my former husband and the brother of second respondent) that Adrienne Connolly should do as she pleased .as I was not susceptible to extortion’.

Apart from the criticism of second respondent’s evidence , Mr Burger suggested that there was no basis on which the applicant’s reliance on inferential reasoning could overturn the overwhelming probabilities in favour of first respondent based on the competing versions of the parties in their testimony to the Court.

Mr Burger submitted that a number of other people received the management accounts but did not discover the VAT fraud including the following:

(i) Fisher Hoffman Sithole who audited the books for SourceCom for the 1999 and 2000 financial year. (ii) Mr Crispin Sonn ,the chairperson of SourceCom and an executive director during part of the period in which the VAT fraud was being committed (June 1999 to 1 May 2000); (iii) Mr Peter Volkwyn, an executive director; (iv) Mr Peter Baird, a non executive director a chartered accountant , described in evidence as ‘a very competent business man’; (v) Mr Phares Mkwanazi, a non executive director and similarly decribed as a highly competent chartered accountant; (vi) Mr C Boorany an executive director; (vii) Mr C Millward a non executive director; (viii) Mr Neill Davies a chartered accountant and the chairman of Brait Turnarounds who in April 1999 required insight into the latest management accounts of SourceCom as well as certain other information regarding the company; (ix) A team who performed a due diligence in SourceCom in July 1999 in the text of a potential Notae Mediae transaction; (x) Loubser Du Plessis which performed due diligence in SourceCom at the insistence of Dr Koornhof and Mr Harlow in July 2000; (xi) Mr Maritz, who replaced Mr Connolly after he left the company in November 2000 and who signed the VAT cheques up until May 2001 without ever discovering that there were incorrect VAT returns.

5.Mr Burger also referred to the monthly management accounts which were clearly distorted as a result of the VAT scam but which contained other distortions which were not influenced by the VAT fraud. Under cross-examination from Mr Rogers, first respondent referred to management accounts for February 1999. In that month SourceCom achieved a margin of 40% against a budgeted margin of 10% .As first respondent told the court, ‘that really concerned me and I looked into that because the figures were there and there is no VAT contribution in that month’.

Similarly in the January 1999 accounts the removal of the VAT fraud from the accounts gives rise to some unexplained consequences. For the month ending 31 January 1999 the cost of sales was reduced by a bogus credit of R660,000. In that month total sales were reflected as being in the amount of R4,051,933 and total cost of sales were reflected as R3,281,030. Total sales figure for the month reflected professional services R114,177 and what was ..referred to as ‘technical’ in the amount of R217,853.( there was an amount of R 5690 which no one appeared to include in these calculations and which was derived from technical services from the Johannesburg branch but the amount does not appear to be to disturb the point of this calculation) As there was no cost of sales in respect of these amounts it would be legitimate to deduct approximately R335,000 from the figure of total sales in order to ascertain the margin in respect of goods sold. If the amount of R660,000 is added back to the total cost of sales, SourceCom would have had an effective margin of minus 6% on its sales. As first respondent testified ‘Nowhere in a business would you sell goods for less than what you buy them for’.

Mr Rogers submitted that applicants’ case depended to a large extent on showing by way of inferential reasoning that the mistakes in the accounts were so obvious that a person of Ms Thorne’s ability and knowledge must have been able to ascertain a distortion.

Some of the distortions were indeed extremely crude. In the accounts for September 1999 Cape Town sales are reflected in the amount of R1,327,851 , the total cost of sales being R346 679 , giving a gross profit percentage of 73,9% as compared to the budgeted figure of 14.4%. In the same amount an amount of R807,519 was described as ‘other’. This amount clearly was added back in order to obtain a gross profit amount of R981,172, which in turn produced the gross profit percentage of 73,9% .First respondent was asked about this by Mr Rogers under cross-examination. ‘So you do not ask his Lordship to find that you looked at this and thought it was genuine. You asked his Lordship to find that the only reason why he did not pick up the problem was because you never looked at this page.? Mr Rogers what I am saying is that again with hind sight and KPMG you have traced this amount back to a particular journal entry…

..Well this particular amount just for to remind his Lordship is in the cost of sales as a bracketed figure R807,000 other.

Now cost of sales is a true expense of –represented without bracket so that R800 – that negative R807,000 is at it were a credit cost of sales? Yes.

Reducing cost of sales ? H’m, H’m

And that is primary made up by a figure of R823,000 which I think you now know to be a bogus reduction…yes.

In cost of sales plus some other number ? Yes what I can say to you is, if I were a party to this fraud that your client wants me to believe there is no way I would have gone into a board meeting with figures like this. How would I explain to anybody, I would simply not have done that. It is so obvious now when one looks at it and points it out?’

6. Mr Burger submitted that, crude as these figures might have been , they were not discovered by first respondent or other board members or any other person involved in examining the accounts of SourceCom. Mr Burger submitted that, if it was argued that the scam was not so crude, then there would be no basis upon which the employment of inferential reasoning could justify the conclusion that first respondent must have known about the fraud.

EVALUATION.

As Mr Rogers submitted in his reply , applicants’ case is that first respondent’s silence in the fact of patent and obvious distortions in the management accounts showed that she was clearly aware of the VAT fraud from the outset. Whatever the finding of the credibility of second respondent , Mr Rogers submitted that the credibility of first respondent’s version had to be tested within the context of the wider probabilities of the case. He submitted that the court was required to consider the credibility of witnesses in conjunction with the probabilities .It was only where the probabilities failed to indicate where the truth probably lay , that the court should have recourse to an evaluation of the credibility of the applicants’ witnesses on the one hand and respondent’s witnesses on the other. National Employers General Insurance Company Ltd v Gagers 1984(4) SA 437 (A) at 441 A. According to Mr Rogers an assessment of the inherent probabilities of the case becomes critical. If the probabilities favour first respondent ,applicants’ would only be entitled to succeed if the court was satisfied that first respondent was lying while second respondent told the truth. Conversely if the probabilities favoured applicants’ they were entitled to succeed unless the court was satisfied that first respondent was telling the truth while second respondent was lying. If the court could not say with confidence which of first or second respondent was telling the truth but the inherent probabilities favoured applicants’ they would also be entitled to succeed.

In my view, second respondent was a completely unreliable witness. His various versions of the events which led to the conceptualization and implementation of the VAT fraud were all designed to reduce his own culpability. Whereas Ms Honey’s version that a genuine mistake had been made(caused by a software programme) which was then exploited by second respondent is a creditable explanation, second respondent’s testimony that a newly appointed person (who was later considered to have sufficient integrity to have been employed by the new shareholders) proposed a fraudulent scheme in the first few weeks of her employment is so far fetched as to raise serious questions about the balance of second respondent’s testimony.

Given that second respondent was adamant about the involvement of first respondent in the VAT scam , his testimony as to the conversation that took place between the two of them in November (or December) 1998 regarding the implementation of the scam is surprisingly vague. Similarly, a version of his testimony that first respondent would ‘enquire why the gross profit in a particular month was so high and I would then explain for example that some of the fictitious Microsoft VAT had been written off against cost of sales’ did not accord with his description of first respondent as the formidable boss, the ‘micro-manager’, who informed herself of each and every detail of the business. Had first respondent been aware of the VAT fraud she would, she may not have required second respondent’s explanation as to how the gross profit had been high for a particular month .The ‘micro manager’ would have in all likelihood been a participant in the manner in which the ill gotten gains form the VAT scam were inserted into the profits.

Second respondent’s version as to how first respondent sought to cover her tracks upon resignation is equally implausible testimony. He simply could not recall in any detail as to how he or she had covered their tracks and to what precise instructions had been given to Ms Honey in this regard. Whatever his version, it was contradicted by Ms Honey who seemed to have no recollection of any conversation with regard to the cessation of a fraudulent scheme.

When second respondent’s testimony is evaluated holistically, it consists of a number different versions which cannot ,on any reasonable basis , be considered to be probable. A junior employee, recently appointed to the organisation proposes a VAT scheme. A formidable chief executive officer described as a ‘micro manager’ (in contrast to the description of Mandy Daddy) agrees to the initiation of the scheme but never has any conversation with the junior employee ,Ms Honey ,as to its initiation or implementation. A chief executive officer who is presumed to be a careful analyst of the company accounts requires periodic reminders from second respondent as to effect of the VAT scam on the gross profit. A highly qualified chief executive officer who once taught tax at a university and was a director of a very reputable accountancy firm and was engaged in aggressive tax savings schemes makes no attempt to hide this nefarious activity upon her resignation.

First respondent was described by Mr Rogers as a slick witness. She tended to be argumentative and tried to preempt a line of questioning. In part there can be no quibble with this description. However in comparison with the pathetic, mendacious figure which second respondent cut in the court ,first respondent was prepared to concede her negligence on more than one occasion, albeit with the benefit of hind sight. At times her conduct could also be explained as a display of moral indignation under intense and most skilful cross-examination by Mr Rogers.

Absent the process of inferential reasoning in which Mr Rogers engaged, the remaining evidence particularly the key testimony of second respondent is not sufficient to support the conclusion that second respondent’s version regarding the involvement of first respondent stands to be believed.

Much of Mr Burger’s attack on applicant’s attempt to invoke a process of inferential reasoning was based on the failure of others to discover the distortions in the accounts. Mr Rogers countered this argument by suggesting that none of these individuals mentioned possessed first respondent’s overall knowledge of the business in all its facets and details. The auditors, for example, would not have received regular monthly accounts in that their function was to audit the books after year-end. They were not involved in the running of the business and would not have had first respondent’s knowledge of particular sales and the margins at which they were being concluded.

In essence , applicants imply that the emphasis placed upon the importance of the board of directors in general and non executive directors in particular to enforce a more rigorous approach of corporate governance is somewhat misplaced .(see the King Report on Corporate Governance(2002) and Wixley and Everingham Corporate Governance(2002)).

Non executive directors may act diligently and carefully but, to a large extent ,they are beholden to executive directors in general and the chief executive officer in particular for the information on which they can render management accountable to shareholders employees, customers, investors and other interested parties in the affairs of the company.

In the present case , the implication of applicants’ response to the argument about the performance of the board is that manipulation by the chief executive officer aided by the financial director prevented discovery of mismanagement or fraud. But that is not what the applicant sought to argue expressly. Applicants contended that the scheme was so crude and the manipulation of figures so obvious, that it was inherently improbable that first respondent could not have known of the effect of the VAT scam on such figures and hence on the overall performance of Sourcecom . Assuming that this description of the scam to be correct, the fact that an entire group of interested parties did not raise any questions with regard thereto is an important consideration to be taken into account in the evaluation of any conclusion that might be reached from the process of reasoning employed by applicants. If the scheme was not crude but had been conceived and implemented with a great measure of sophistication, this may help explain why first respondent did nothing to cover her tracks upon resignation ;she being confident that the sophistication of the scheme would endure without being discovered. On this basis, it is then equally probable that she did not know about the scam which had been hatched by second respondent for whatever motive with the connivance of Ms Honey. There is some support for this later conclusion in the evidence given by Ms Honey when she testified as to how she had prepared Excel spread sheets on the instruction of second respondent in which certain figures were hidden from view and required special knowledge in order to gain access thereto.

Applicant can therefore not have it both ways. Either the scheme was so crude that the distortions must have been apparent to first respondent in which case they should have been apparent to some or all of the other parties to the organization ; that is to those persons who were possessed of business expertise and accounting knowledge. If the scam was of a sophisticated nature , it may well have prevented a business person even as knowledgeable as first respondent in matters of tax and accounting from discovering the true position of SourceCom.

Much was made by applicants of the benefits which would accrue to first respondent by virtue of the VAT fraud as well as her lack of integrity . Mr Rogers referred to the inaccurate submission of a passport application to the Department of Home Affairs, a non disclosure of a bonus which first respondent received for tax purposes in order to justify the conclusion that first respondent was a person lacking in integrity.

Unquestionably first respondent did stand to benefit from increased profits in the organisation but the alternative submission that second respondent also stood to gain by way of an award of shares and a further payment pursuant to the conclusion of a restraint agreement and further bonuses cannot be discounted. That first respondent may not have always performed to a standard of integrity to be expected of a citizen in this country, does not mean that her evidence per se in this dispute should be discounted. Furthermore there is uncontested evidence that she refused to pay respondent’s ex wife in circumstances where such a payment may well have been to her advantage, assuming applicants’ version to be correct.

Applicant must show in a case such as the present that the inference sought to be drawn is consistent with all the proved facts. See Cooper and Another NNO v Merchant Trade Finance Limited 2000(3) SA 1009(SCA) at 1027-1028 where Zulman J A accepted that the first part of the test for inferential reasoning as laid out in R v Blom 1939 AD 188 at 202 is applicable to civil cases; that is the inference to be preferred must be the most appropriate ,and plausible one to be drawn from all the proved facts. See Ocean Accident and Guarantee Corporation Ltd v Koch 1963(4) SA 147(A) at 159 C-D. This dispute concerns the plausibility of the competing versions. Even if there are competing reasonable interpretations, the question arises as to which of the two test in a versions admits of the most plausible conclusion to be drawn from the facts presented to the court.

In this dispute, the account put up by first respondent is ,at the very least sufficiently plausible, to prevent the conclusion that , on a balance of probabilities, applicants argument justifies the finding of fraudulent conduct on her part. The oral testimony of second respondent must be discounted .Here is a case of a crook( who admitted that he had a serious drug problem at the time) upon whose competing versions and desire to shift blame on to other persons no reliance can be placed. His performance as a witness leads to but one conclusion ; that he lied as a matter of course on every conceivably relevant subject .His demeanor , evasiveness and preposterous explanations together with internal and external inconsistency in his testimony may be the stuff of a novelist but most certainly not of credible witness. First respondent proved to be a far more reliable witness.

So the applicants were driven back to the exclusive employment of inferential reasoning to bolster second respondents’ version . When these two competing versions are compared, it is difficult ,on the reliable evidence to conclude with any confidence that the applicants version is the more plausible. .

RECKLESSNESS.

The principle case which applicant sought to make against first respondent was one of fraudulent dealing. Mr Rogers submitted however that the incorrect deductions of input VAT constituted a reckless manner of conducting SourceCom’s business. In relation to those persons who actually knew that bogus VAT was being deducted and had the intention to defraud SARS there was fraud. However this did not preclude the possibility that other persons , who were not party to the scheme, may have been parties to the reckless deduction of input VAT.

Mr Rogers submitted that first respondent was such a person. She signed the first VAT cheque involved in the fraud and as chief executive officer she knew that large local sales of microsoft licences were concluded which were not matched by any input VAT payable to Microsoft. The amounts by which VAT was underpaid was very large. First respondent as chief executive officer was very much a party to the reckless underpayment of VAT. Any reasonable business person in her position would have realised that the VAT payment simply could not be correct and would have done something about it.

The alternative case of recklessness proceeds on the basis that the evidence of second respondent concerning first respondent’s intention to defraud stands to be discounted. Mr Burger submitted that ,if the evidence placed before the court by applicant to establish fraud is not accepted , then it could not be accepted for an alternative purpose, namely that of reckless conduct. If it cannot be shown that first respondent was aware of the fraud being perpetrated by second respondent, it is difficult to see how she could be regarded as having adopted so supine an attitude towards it which would amount to the kind of passive concurrence envisaged in Howard v Herrigel and Another NNO 1991(2)SA 660(A) at 674 H.

Furthermore the recklessness required by s 424 cannot amount to negligence. As chief executive officer first respondent might well have been negligent in not ascertaining the nature of the fraud. Certainly, this is a plausible conclusion upon which to arrive on the basis of the evidence presented to the court. It may have been the case that she relied upon second respondent to prepare the accounts and generally assumed them to be an accurate reflection of the company’s business performance.

For these reasons, I am of the view that applicants have not discharged the onus which rests upon them in terms of section 424 of the Act . Accordingly the application is dismissed with costs including the cost of two counsel, such to include the costs of the initial application brought by applicants and which resulted in the judgment of Blignault J on 21 November 2001.

DAVIS J

.

.

---