PROVINCIAL COURT OF NOVA SCOTIA
Citation: R. v. Mullins, 2021 NSPC 35
Date: 2021-07-23
Docket: # 8311820, 8311821
Registry: Sydney
Between:
Her Majesty the Queen
v.
David Joseph Mullins
|
Judge: |
A. Peter Ross |
|
Heard: |
September 21 – 25, 2020; November 9, 10, 2020; January 5,6, 2021; May 11, 2021 in Sydney, Nova Scotia |
|
Decision |
July 23, 2021 |
|
Charge: |
s.334 and 380, Criminal Code of Canada |
|
Counsel: |
for the Crown, Alex Keaveny for the Accused, Christopher Conohan |
REASONS FOR DECISION
Introduction
[1] The accused, David Joseph Mullins, is charged as follows:
1. Between 14 July 2014 and 14 September 2016 at or near Sydney, NS, did steal money, the property of the Sydney Garrison Messes of a value exceeding five thousand dollars contrary to s.334 of the Criminal Code;
2. And did on the same date and place by deceit, falsehood or other fraudulent means defraud the Sydney Garrison Messes of money contrary to s.380 of the Criminal Code.
[2] The proceedings were by Indictment. The accused elected trial in provincial court on 23 December 2019. His trial commenced on 21 September 2020 and continued on various dates until closing submissions on 11 May 2021. One witness appeared by video from Germany pursuant to s.714.2. Over 9000 digital files, many of them accounting records, were tendered into evidence on DVDs and USBs. Expert opinion evidence was given by a forensic accountant.
[2] Crown argues that there is compelling circumstantial evidence of theft from observed inventory discrepancies, and more direct evidence of theft from unofficial receipts and cash found in possession of the accused.
[3] Defence argues that there is no evidence of loss, much less theft. It suggests that the control measures and accounting systems in place during the time when the theft is alleged to have occurred were incapable of tracking such losses reliably and that it is now impossible to draw safe conclusions from an analysis of available accounting information, related data, and the other evidence in the case. It argues that “criminal law is made of sturdier stuff” than the Crown has mustered here.
[4] Defence argues further that even if the evidence does prove loss of funds, it does not show clearly that the accused is the person responsible for this. It acknowledges that the accused may have violated some policies and procedures but points out that this is not proof of a crime, that the accused was no less lax than many others who were in some sense responsible for the Sydney Messes.
[5] Defence has attacked the Crown’s case at many flanks. It suggests that there are many sources of doubt about the guilt of the accused, saying “there are too many holes in the bucket”. I will attempt to address these various concerns under separate headings. Some overlap is unavoidable; the redundancy is regrettable.
[6] I am mindful that doubt can accumulate from various sources. At the end of the day every element of the offences must be proven beyond a reasonable doubt, otherwise a finding of not guilty must be entered. The onus of proof rests with the Crown.
[7] In these reasons I do not intend to tell the entire story of the trial. I will instead discuss the most contentious aspects. Having concluded that the Crown has proven the guilt of the accused, I will attempt to show how the evidence leads to that conclusion without doing an exhaustive review of the entire record. This will, I fear, be tiring enough.
[8] Whether the “Sydney Garrison Messes” is, strictly speaking, a legal entity or “another person” within the meaning of s.322 of the Criminal Code is somewhat unclear. However very little turns on this. I take the phrase to mean, in effect, ‘the members of the Sydney Garrison Messes as represented by CFMWS’. These are the people who suffered the loss; CFMWS is the entity which managed the Mess finances on behalf of these members and was the custodian of their funds.
[9] Although the timeframe for the charges is 14 July 2014 to 14 September 2016, I will also be referring to a “relevant period” which connects with the forensic accounting evidence. This period is 1 April 2015 to 29 September 2016 (explained in ‘Inventories’, below).
[10] Various aspects of the case are discussed under the following headings:
Overview, Mess operations – Par [11]
Overview, Mess bookkeeping – [19]
Financial Controls – [20]
The accused’s request for an audit – [42]
Documentation – completeness / integrity of the record – [46]
Cash registers – [61]
Policies and procedures – [71]
Cleaning services – [83]
Rental agreements – [90]
Training and use of forms – [91]
Cash found in Accused’s office – [98]
Unofficial receipts – [104]
Karmac account – [117]
Inventories – [118]
Forensic accounting evidence – [126]
Cost multipliers – [165]
Spoilage [ [191]
Bartenders – [209]
Conclusion – [211]
For ease of reference, certain acronyms denote:
DND – Department of National Defence
NPF – Non Public Funds
CFMWS – Canadian Forces Morale and Welfare Services
DRAM – Deputy Regional Accounting Manager
GL – General Ledger
DSR – Daily Sales Report
HR – Human Resources
Overview, Mess operations
[11] The Mess operations of the Sydney Garrison at Victoria Park are operated by an adjunct of the Canadian Armed Forces known as Canadian Forces Morale and Welfare Services, which is often referred to (and will be herein) as Non-Public Funds, or simply “NPF”. These programs provide support to military personnel. NPF is funded by memberships and revenue from operations, rather than by public finances.
[12] An accounts division of NPF, situated in Halifax, has provided accounting services to Mess operations in Sydney since at least 2014, and likewise to other military bases in the Atlantic region. They “keep the books”, so to speak. The person in charge there is the Deputy Regional Accounting Manager (DRAM). During the material time this person was George Hirshman.
[13] There were actually three Messes at the Sydney Garrison - a Junior Ranks Mess, a Sergeant and Warrant Officers’ Mess and an Officers’ Mess. Each had a regular Armed Forces member as a Mess President. Although physically separated they were managed by one person who reported to the Mess Presidents and who dialogued with a Mess Committee for each. During the material time this manager was the accused, David Mullins.
14] From a bookkeeping and operational perspective, it was necessary to track certain costs and revenues Mess-by-Mess, but essentially they were three aspects of one operation, sharing inventory and staff and equipment as needed. For certain things, such as purchase of equipment, authorization of the Base’s Commanding Officer was required. In addition to regular operation for members, the Mess Presidents were responsible for booking out the various facilities to third parties for weddings and other such events. The Mess manager handled the day to day operations, including the management of bar staff, who were themselves, according to policy, employees of NPF. From a revenue perspective, and for the purposes of this case, the bar function was the most important business of the Sydney Messes.
[15] The accused was hired to manage the Sydney Messes in the summer of 2014. Previously, Mr. Mullins earned an undergraduate degree and then an MIC designation (Management Institute of Canada). He spent 25 years with Public Works Canada as a senior planner, then worked as a Labour Standards investigator for N.S. He was director of a call centre in Sydney, comptroller for a start-up pharmaceutical firm and is now with a company doing work in freshwater filtration systems. Mr. Mullins was in the Military Reserves for 8 years from 2000 to 2008.
[16] Mr. Mullins said that in his various jobs he would have reviewed banking documents and records, and as comptroller performed accounting functions, supervised bill payments, etc. He also testified that the MIC course included instruction on accounting and bookkeeping.
[17] At one point in his testimony, concerning a possibility of missing funds, Mr. Mullins spoke to Hirshman about getting copies of deposit slips from the local Credit Union. He says Hirshman told him he “wasn’t an accountant and would not understand”. In response he told Hirshman that he had “a significant amount of financial planning and accounting experience.” I think the latter of these two statements is the correct one.
[18] Mr. Mullins acknowledges that bar sales was the primary business in the Messes. He knew that sales would have to be tracked. He says he knew, when he was hired as Mess manager, that there would be reporting requirements. His financial overseer at NPF from then until the date of his termination was George Hirshman, DRAM.
Overview, Mess bookkeeping
[19] Various records were kept at the Sydney Messes including daily sales reports (DSRs), invoices for purchases of liquor, etc. However, the accounting functions were ultimately handled by DRAM and others in his office. All financial transactions were to be reported there, and all relevant financial documents sent to them for recording and reconciliation. In terms of banking, revenues generated by the Sydney Messes were deposited into one ‘merged’ bank account. Other Mess operations from other locations in the Atlantic region did the same. This is less than ideal from an accounting perspective. It removes one layer of verification and reconciliation which would be available if the Sydney operations had their own dedicated account. However, DRAM had an internal accounting system (Oracle) which separately coded each Mess, in Sydney and elsewhere, thus permitting a breakdown of finances on a Mess-by Mess basis. DRAM employed clerks with responsibility to keep track of such things.
Financial Controls
[20] Much of what is discussed under this heading is clarified by what follows after.
[21] Defence claims that the evidence paints a picture of poor financial controls at NPF. It says this causes doubt about whether there was actual loss of inventory or money from the Sydney Messes, and provides no satisfactory explanation of the reason for any loss which may have occurred.
[22] The fact that the Sydney Messes did not have their own discrete bank account is regrettable. However, this did not deprive NPF of the ability to examine relevant financial transactions. While a true bank reconciliation would have given an additional layer of analysis and made the job of the forensic accountant, Roberta Sullivan, easier, her evidence still carries considerable weight.
[23] NPF’s internal accounting system, run on Oracle software, permitted receipts and deposits for the various Messes reporting to its office in Halifax to be tracked – to be allocated to particular Mess accounts. Three such accounts were maintained for Sydney, corresponding to the three Messes at Victoria Park. Although money received by NPF from various Messes in the region was deposited into one merged bank account, the GL / Oracle accounting system operated by NPF did distinguish one operation from another and did permit reconciliation of income and expenses for each individual Mess. The accuracy of these depended on proper attribution of financial records, which was the job of Darryl Beaumont. The Sydney Mess accounts were his particular responsibility.
[24] Mr. Beaumont spoke about various documents found in Ex#2 and Ex#3 and explained various features of the GL accounting system, including how various things were categorized and displayed, how Daily Sales Reports were inputted into the General Ledger from the original paper record, etc. He explained how petty cash was occasionally used to purchase liquor, and how this was reimbursed to the Mess. He said if dues or other things were paid in cash there “should have been a 602” (referencing an official form, discussed below). He would make some attempt to get missing documentation, after which it “went to the boss and was then his problem.”
[25] Beaumont spoke about the z-slips (sales data printed out by a cash register), the “departments” assigned to types of liquor, soft drinks, etc. Whatever z-slips he received would be matched to the deposit and to sales. Money received from sources other than the bar “would come with a 602 to justify”. In cross-examination he said “Sydney’s DSRs were supposed to come with the z-readings attached – if not I’d send Mr. Mullins an email and report to Mr. Hirshman – if this was a persistent issue it was out of my hands.” Here again, as with the lack of a separate bank account, cash register readings would give Ms. Sullivan an additional level of verification, but given the requirement for Mr. Mullins, as for any Mess manager, to provide NPF with a daily sales report, I do not think the occasional absence of such detracts significantly from the impact of Ms. Sullivan’s evidence.
[26] Beaumont could not see in Sch 1.6 of Ms. Sullivan’s report any transfer-out forms (603s) for free liquor, although he said such should have been filed with him. He said the obvious – that if no documentation is given to him, there is no input in the GL. He said the DSRs he received were “all captured in the GL, along with everything I’ve spoken about thus far”, suggesting that if he’d received a 603 regarding spoilage of inventory it would have been entered.
[27] The accused contends that accounting procedures at NPF were unreliable at the time of his hire and suggests that they didn’t get any better afterwards. He became concerned because he discovered a mistaken charge to Sydney for alcohol purchased by a Halifax Mess. He said that revenues were not being properly credited to the Sydney operation. He said he complained to Hirshman that “nothing would balance”, that his deposits “were not showing up.” He says he was told that their money was going into a general fund and that there was no way to track it. This is not what Hirshman told the court. He said that while there was no separate bank account, funds were tracked by their internal accounting system.
[28] Lindsey Reynolds had similar concerns, saying that she and the accused were frustrated at “not being able to see where money goes in, where it sits . . . if we deposit $9,000, where does it go?” She could not see how deposits made by Sydney “showed up in our bottom line . . . although we kept putting money in, that bottom line got less and less.” This coincides with Hirshman’s assertion that the Sydney Messes “were going into the red.”
[29] Hirshman acknowledged that there was a mistake made early in Mr. Mullins tenure – the misattribution to the Sydney Messes of liquor purchased by a Base in Halifax. He says this was identified and corrected. Having heard the evidence on this point, the idea that liquor may have been improperly ‘booked’ during the relevant period seems to be speculation. I here note that Mr. Beaumont had specific responsibility for Sydney during this period and this, to my mind, makes mis-coding of the accounting information sent to NPF even less likely. In cross-examination he said “I primarily did Sydney, unless I was off work.”
[30] The accused says that in his first year he met with the Mess presidents more than 20 times to discuss deposits and discrepancies. He says they were concerned that they could not see where his deposits were “showing up”. Given the absence of a discrete bank account, deposits would not “show up” directly on a bottom line bank balance. This may well explain the confusion. This does not mean that there was no accounting being done whatsoever.
[31] The accused says he was told by Hirshman during one meeting that the money went into a general fund, with no way to track it. The first part of that is undoubtedly true; the exasperation experienced by the Mess presidents, such as Ms. Reynolds, may have been a result of NPF’s accounting system and the lack of a separate bank account dedicated only to Sydney. Unquestionably, such an account would have lent transparency to the numbers. That said, the analysis and breakdowns of the financial data undertaken by Ms. Sullivan permits considerable clarity. There is no evidence of any actual mistake other than the one noted by Hirshman himself, which was caught. There is little to support the idea that data-entry practices were erroneous at any other time. I say this despite the evidence given by Trina MacDonald regarding the short time she was in Mr. Mullin’s position subsequent to his termination. She said that as of 2017, when she took over, the system “did not show accountability.” She seemed quite critical of lax accounting practices, but it is not clear to me that she was referring to NPF functions as opposed to local practices. If, for instance, practices followed by Mr. Mullins such as issuing unofficial and ambiguous receipts (discussed below) had continued, this would raise anyone’s eyebrows. Ms. MacDonald also criticized security measures at Sydney saying that until November of 2018 nobody checked who had keys, who was coming or going - “it was a free-for-all.” Despite this, there is nothing in the evidence to suggest that during the relevant period significant quantities of liquor were being pilfered from the stock rooms.
[32] The accused aimed fire at the lax accounting at the Messes at the time he was hired. He referenced a $2000 bill (not in evidence) from a local restaurant to the Officer’s Mess. However it is not clear that he was talking about something during the “relevant period”, nor how this would in any way affect the inventory discrepancy identified by Ms. Sullivan. He says “the money came out of the Officers Mess account”. As this could only be a GL account within NPF, this would have been known to NPF and would have been available to Ms. Sullivan if it occurred during the relevant period. Here I sensed an attempt at obfuscation, not raising doubt based on the evidence before the court.
[33] As to the involvement of the bartenders in maintaining records, it seems the only undocumented steps were those suggested by the accused himself. The bartenders had been NPF employees, paid their normal wage through NPF for the hours they worked, based on the hours they logged and which Mr. Mullins approved. This all changed under Mr. Mullins.
[34] Speaking of her experience working as a bartender during 2016, Trina MacDonald confirmed that Mr. Mullins paid her cash for this work. She said that bar sales were tracked and compared with stock counts at the beginning and end of the event in question. She explained that they would weigh the bottles, count the cash from the till, fill out the inventory sheet and other forms, and “put it all together in the office . . . so that the manager could verify” and do the bank deposit. She said “what we wrote on stock sheets should match with what comes out of the cash register”. She said readings from the cash register gave the sales totals. She said that when she was paid cash by Mr. Mullins for her shifts as bartender, he “told us not to tell anybody about it.”
[35] Lindsey Reynolds said that her role was not to “actively manage Mr. Mullins”, that she was not involved in day-to-day bookkeeping. She went on to say that bartender fees charged as such to outside parties in rental agreements should have been deposited to the Mess account and then payment made to the bartenders through NPF. She “never gave Mr. Mullins permission to pay people under the table.”
[36] While it is possible that some local officials did indeed know about, and tacitly approve of the idea of paying the bartenders in cash, it seems equally possible that they assumed the accused would clear this with NPF and work out any financial aspects as required.
[37] Ms. Sullivan agreed to the suggestion that sloppy record keeping could account for the inventory discrepancy she calculated. The point was not pursued, and Ms. Sullivan was quite scrupulous not to point fingers or attribute blame. That said, having heard all the evidence, it appears to me that one type of “sloppy record keeping” might be the failure to keep a daily sales report. If no DSR was done, or if one was prepared but then lost, and yet revenues were deposited into the NPF account as being from the Sydney Messes, it seems that a discrepancy of this nature would have been picked up by Mr. Beaumont and supporting documents requested. It seems highly unlikely that he would know that a deposit of money was made from Sydney yet carry on with no idea what it was for. Further, Ms. Sullivan said that “given the size of the bar compared to the size of the inventory, it would require very large errors or missing information to bring (the discrepancy I identified) outside the range I set out”, and “that for sales to be off by almost one-half of recorded sales was possible but unlikely.” If one takes at face value the accused’s professed attention to tracking inventory, the idea that local bookkeeping errors or mislaid DSRs account for the financial discrepancies has little currency. Nor do I think the perceived shortfall can be attributed to sloppy practice at the NPF office.
[38] Through all of this analysis, the accused must be taken to know how it would look to NPF if proceeds from liquor sales were being deposited without correlating sales reports. Just as z-readings, or paper tallies, act as a control on money remitted by bartenders to the Mess manager, so too do DSRs support deposits of money into the Mess accounts at NPF.
[39] In Ex.#11A one finds “a letter to Sydney Garrison Messes from 36 Canadian Brigade Group 26 August 2015 Col. G.B. Thomson –
2. Sydney Garrison Messes are in a negative cash flow situation. Some of the problems are related to the nature of Reserve Force training patterns, demographics, availability of Reserve Force personnel and the distance from NPF Accounting Services at CFB Halifax. Documentation submitted to the NPF Accounting Office must contain detailed account numbers identifying the correct GL and entity to charge expenditures and to record the collection of revenue. It is the responsibility of the Bar Supervisor and the Executive Committee to ensure this happens. A culture of mistrust has been allowed to grow since 1 November 2013 between the Executive Committees and the NPF Accounting Office and vice-versa; this needs to be resolved immediately. The Executive Committees are asking for an in-depth audit of their messes for the period 1 November 2013 to present; to help reverse this mistrust, the Canadian Forces Personnel Support Agency’s (CFPSA) Internal Audit and Review Division could be utilized.
[40] Nobody testified that the accused had knowledge of this letter, but it is reasonable to infer that the message was relayed to Mr. Mullins, including the dictum to submit documentation to NPF which “must contain detailed account numbers identifying the correct GL and entity to charge expenditures and to record the collection of revenue. It is the responsibility of the Bar Supervisor and the Executive Committee to ensure this happens.”
[41] As the Crown has argued, where a business operates on a cash basis it is next to impossible for external controls to guarantee financial integrity. Controls depend on the people who handle such cash; their honesty is one of the most important controls against loss of funds.
The Accused’s request for an audit
[42] According to the accused, friction with NPF in and around June of 2015 resulted in a request later that year for a forensic audit of the Sydney Mess accounts. The accused says this was his idea. The Mess Presidents apparently agreed with the suggestion. Hirshman declined, saying it was unnecessary and would be too expensive.
[43] Early in his tenure, there was an issue with improper allocation of expenses by Halifax Messes to the Sydney accounts. Lindsey Reynolds as President of the Jr. Ranks Mess put the request for such audit as “prior to his (the accused’s) investigation.” From Mr. Mullins evidence it seems he put forward the suggestion late in 2015, after concerns were expressed by Mr. Hirshman about a deposit.
[44] Defence argues this is not how someone who was stealing money would act. This point warrants consideration, but I also bear in mind that the Crown’s case does not rest upon the funds going missing at a specific time. It is possible that defalcation of funds occurred after the possibility of an audit was foreclosed. It is also possible that this was mere bravado, that the accused never believed the suggestion would be adopted.
[45] I do not give the accused’s request for a forensic audit the significance Defence argues for. Such words may have preceded any wrongdoing on the accused’s part; indeed, one could argue that the fact such an audit was denied may have emboldened the accused. Alternatively the request may have been pure bluster. While I find the request was made, this is not something which raises real doubt about the allegations. Neither, I might add, does it do anything to bolster the Crown’s case.
Documentation – completeness / integrity of the record
[46] A number of witnesses viewed, identified and explained specific examples of various documents. To have authenticated them all in this fashion would have taken an inordinate amount of time. In its final submission Defence conceded that the documents in evidence “speak for themselves” but it raised concerns about both the completeness and the integrity of the documentary record before the court.
[47] Although disclosure concerns were resolved during the trial, questioning of Hirshman and Beaumont was directed in part to whether Hirshman provided the military police investigators all financial records in the possession of NPF (and hence whether Ms. Sullivan worked with a complete record). However a review of the testimony convinces me that there is no missing or lost evidence and that nothing has been slanted against the accused.
[48] Mr. Beaumont was tasked with locating and transferring to his boss, Mr. Hirshman, “all the documents the office had for this matter – everything regarding the Sydney Reserves”. He said he collected this material from filing cabinets, where it had been stored, and also gave over data from their GL system. He said he “went over this quite extensively” so as to give “all the accounting records for the Sydney Garrison Messes for the period – he wanted everything I had.”
[49] Defence suggested in questioning that the complaints from Sydney about broken cash registers and a request for an audit of NPF’s bookkeeping may have turned Mr. Hirshman against the accused. Hirshman denied this, saying “it did not affect my opinion of him.”
[50] According to the accused he raised questions about some suspicious documentation (invoices with no record of payment) at the time of his hiring. He says Hirshman told him to destroy such records. Whether Hirshman should have sought accountability for earlier sloppy practices may be debated, but it does seem that he was trying to wipe the slate clean for Mr. Mullins to start afresh.
[51] Beaumont said that once documentation from the Messes was entered in the GL, the documents themselves were stored in filing cabinets in the NPF office. Hirshman said he gave the MPs all the documents he had relevant to the Sydney Messes, as well as the data from the GL. If he were to selectively edit this, he seems he would have to ensure this aligned with the GL, which presumably he could not alter. This makes highly improbable the idea that he would deliberately withhold documentation which might exonerate Mr. Mullins. I also regard as vanishing small the chance that Hirshman, suspecting a fraud, would engage in the same sort of behavior that he himself was investigating.
[52] Hirshman said, and I accept, that “whatever was pertinent to the investigation was turned over.” He said he spent four to five 10-hour days going through all the hard copies of documents he had on file, comparing to what had been reported in the GL, to ensure that they were not posted to the wrong account. In other words he gathered the “source documents” in file with NPF in Halifax and those taken from the accused’s office and compared these to what was shown in Oracle. He said he “absolutely” provided as complete a file as existed.
[53] Hirshman referenced Excel files in Ex.#3 which he created, showing all purchases and adjustments to inventory between the relevant dates, broken down by account, for the Sydney Messes. He said he pulled all the GL data, that he “encompassed everything from all accounts”. Although at first he seemed uncertain whether he had provided the sales data, as his testimony proceeded he regained a recollection of having done so and I am satisfied that he did. He said he “did not leave any Sydney-related entries out”. Later in cross he said that he specifically remembered giving sales data such as DSR’s to the MPs.
[54] Ms. Sullivan did not just perform an analysis of the digital records (the GL accounts kept through the Oracle software). She examined all the “source documents”, the paper records which were the source of the data inputted to the Oracle system. This provides a check against improper allocation of data to incorrect accounts (such as other Messes besides Sydney) within Oracle. As she put it “I worked backwards from what I actually had, not from what ‘should have been done’.”
[55] Sgt. Joey Wilson was accepted as an expert in the extraction and preservation of electronic data. Pursuant to a warrant, he extracted information from the computer used by the accused during the relevant period and seized on the day the accused was removed from his position. He explained how data was filtered, extracted and sent to investigators, who “flagged” certain files from a total of 54,000 as being of interest. He put these into a CD, Ex.#11. His evidence serves to support the authenticity of these electronic documents, introduced by the Crown as part of its case. Extracted were the accused’s emails, his Outlook ‘calendar’, invoices and other information concerning rentals of the Mess. Ex.#11A is a USB counterpart to #11, with a date correction.
[56] The accused claims to have kept ledgers which were not turned over to the military police and hence not disclosed to the Defence. He suggests these would have supported his testimony and served to exculpate him of wrongdoing. Hirshman was asked if he saw handwritten ledgers when searching Mr. Mullin’s office. He said he did not recall and would not have left such behind, that he “would absolutely have turned such over” to the MP’s, noting that Lt-Col Gallant was with him when he seized these documents.
[57] In this regard, Mr. Mullins testified that he kept four record books: an inventory book, payroll book, discussion book, (all in a drawer) and daytimer (which was on top of his desk at all times). He says he tracked events in two places: the calendar noted above and a Word document on his computer. He says he saw Hirshman leave with all these books the day he was fired. The idea that Hirshman would have taken these in full view of the accused and the Commanding Officer and then hidden them from investigators and the accused strikes me as preposterous. Mr. Mullins’ computer was also seized. How would Hirshman know that there would not be a reference on that hard drive to paper ledgers kept by the accused, some corresponding information which would point to the existence of written ledgers, or some person who could come forward to confirm their existence?
[58] As of September 15, 2016 Hirshman certainly had suspicions of wrongdoing by Mr. Mullins, and this certainly would not have endeared the accused to him. However I do not think that Hirshman would deliberately withhold documentation in his possession from the MPs, given the extreme animus this would require and the personal risks involved.
[59] In regard to 603 forms filed by the accused (as he said he did for large amounts of spoilage), Hirshman would know the importance and meaning of these. To withhold from the MPs under these circumstances would mean ipso facto that he was deliberately ‘framing’ the accused. He indicated that if a 603 document had been filed it would have been included in the record provided, which I accept.
[60] The DSRs were typed or hand-written records created by the accused as Mess manager. If liquor was sold at the bar but the sale not recorded this would result in an inventory discrepancy (as is discussed below). Absent cash register readings Ms. Sullivan could not confirm that all sales on a given night were recorded on a DSR. However, if cash registers were not producing these “z-readings”, Mr. Mullins would surely know the importance of keeping a hand-written record of bar sales. The bartenders reported to him in an employment and in an accounting sense. I conclude that any such tallies they were required to make would have been given to the accused. I conclude that there is nothing by way of missing records, except any which the accused himself chose to hide.
Cash registers
[61] There were undoubtedly periods of time when the Sydney Messes did not have working cash registers, despite the efforts of Mr. Mullins and others at the Base to rectify the situation. Mr. Hirshman was aware that there were issues when he “sat down” with personnel from Sydney in 2014 at the time when Mr. Mullins was hired. However NPF was not responsible for procuring such equipment. Hirshman recalled that new registers were purchased in 2015, but that they only gave proper readouts of sales (z-tapes) for a short time. He instructed Darryl Beaumont to follow up but “no time afterwards did we get the readings.”
[62] The accused said there were “constant problems” with the registers, even the new machines bought in 2015. As a result, registers had to be moved around between Messes.
[63] Beaumont said that “sometimes” the cash registers were being repaired but that it “did not happen very often” that no register was used. In cross-examination he said there was “the odd time” when a cash register was not working in Sydney. He could not remember “how Sydney would handle that, whether they kept a paper tab to keep track of it.”
[64] Mess committee minutes found in Ex.#11A from 22 September 15 show concern over the cash registers as well as the ice machines. Minutes of the Sgt and WO Mess committee of 11 February 16 read:
a. Mess Equipment: PMC gave a brief update on acquisition of the new mess equipment: Cash Registers, Ice machine, Fridge, dishwasher and draught beer machine. Sgt Capstick asked if or mess will be available for a mess committee meeting on 25 Feb 16. Update on progress regarding this issue to be presented at next General Mess Meeting, added to agenda.
[65] The Jr. Ranks Mess committee minute 25 Feb 16 reads:
1) Review of Old Business
- Cash Register was purchased in late Dec. Paid for by Marlant.
- Dish Washer, Ice Machine, Beer Fridge have been purchased and paid for by Marlant. Waiting for bids for the removal of current appliances and installation of new appliances. Hoping to be done by 31 Mar 16
[66] This and other evidence before me confirms that there were indeed problems with the cash registers and refrigeration equipment (the latter being relevant to “spoilage” which I deal with below).
[67] If Beaumont were to get no report at all of activity at the Messes on a given night, then he would be completely unaware whether cash registers were working or not. If there were no z-slip, no hand-kept record, no daily sales report and no deposit slip – then he might have a misleading picture of the frequency of cash register breakdown.
[68] It seems obvious that however frustrating or challenging it might be, non-functioning registers would have to be “handled” in some fashion. Tracking bar sales was a critical financial function. Mr. Mullins was well aware that this was his responsibility. In direct response to a question about the broken registers the accused said he would do a stock count after an event and “compare to the previous night’s sales.” This suggests that even if he had no z-readings there was indeed a sales record kept, if only a written slip by the bartender(s).
[69] One scanned document does indeed show a hand-written table of items to be completed by a bartender, including counts of liquor. It is obvious that individual sales could likewise be tabulated, if necessary. Indeed, Mr. Mullins testified that when the registers were not producing z-readings he instructed the bartenders to write down sales on slips of paper. He went on to say that “the only way I could correct this was to do an inventory the next day . . . which was normal procedure, by the way.” Elsewhere the accused said that he sent daily sales reports weekly to Mr. Beaumont, usually on Thursdays. This also suggests that he was able to track sales data without z-readings from the registers.
[70] I am left with the impression that either Mr. Mullins is exaggerating the extent to which the cash registers were not functioning, or that his own actions hid the extent of the problem from NPF (though not from local officials).
Policies and procedures
[71] Failure to abide by policies, including HR policy, which applied to NPF, is not itself an element of the alleged offences. It is not a crime to ignore or violate policy, but willful disregard of policies which governed bookkeeping and payroll practice is relevant to the mens rea which Crown must prove. Mr. Mullins’ questionable behavior in regard to the policy governing payment of bartenders raises particular concern.
[72] George Hirshman said that when rental fees were collected by the Mess, an official 602 receipt ought to have issued in exchange. He was referred to one particular rental agreement (Ex.#13) showing that $700 of the amount charged for this event was for the cost of bar stewards. He indicated that income earned for bartending was generally covered by salaries as part of that person’s employment, and that if money was collected from a renter to offset this wage expense, the proper procedure was to deposit the money with NPF and then make a request for payment to the individual bartender based on the hours worked.
[73] Hirshman said that any money received pursuant to a rental contract for the Sydney Messes should have been covered by an official receipt (602). He said there were “no circumstances” in which it was appropriate to pay bartenders in cash because they were NPF employees, paid salary by a national pay office, and all such remuneration had to be tracked for tax purposes.
[74] Jylene Ryan was acting Regional Manager for Human Resources for NPF towards the end of the relevant period. She explained that a Mess manager would input employee hours into NPF’s payroll system, and that all hours worked by bartenders should be tracked in this system. She said that “cash for work was not subject to deductions and not permissible under NPF policy”. From Ex.#9 she identified a specific policy for Mess administration and a “value and ethics code” for the public sector which governs all employees of NPF, both available on its website. She was unsure about training and orientation practices at the time the accused was hired; strictly speaking her evidence refers to the period 2016 forward.
[75] Mr. Mullins says he was not referred to any HR policies. He testified, “I was aware policies were there, but not whether they applied to me.” He says the only policy relevant to him would be hiring policy, and that he spoke to one Ellen MacKay in the HR department about this, specifically about “how to keep the bartenders by giving them sufficient hours.” I do not have the benefit of Ms. MacKay’s recollection of such conversations.
[76] In Ex.#9 one sees a copy of Mr. Mullins offer of probationary employment dated 20 Nov.14 (retroactive to October 13, 2014) which says “NPF Values and Ethics Policy attached.” This is not absolute proof that the policy was actually attached to the letter, but at the very least it is reasonable to infer that the letter was delivered to the accused, that it brought the existence and importance of the policy to the accused’s attention at a very early date in his tenure and that he would know it was relevant to him as Mess manager, the very position he was being offered.
[77] Major Doyle, chair of the Officers’ Mess at the time the accused was hired, said that he had seen the “mess administration” policy in Ex#9. He said that “we wanted all policies followed”, including policy around payment of employees and contractors, although he “did not instruct Mr. Mullins on such things” and that “it was up to NPF to ensure Mr. Mullins was following policies.”
[78] Mr. Mullins, for his part, said that his idea to pay cash to bartenders, as an enticement to work, had the approval of the Mess presidents and commanding officer. This seems less than certain, given the evidence of Major Doyle, and completely aside from local approval, it would be clear to the accused that approval for cash payment to bartenders would have to come from the HR department, NPF in Halifax, or both.
[79] While there may have been some confusion about responsibilities in certain areas, such as replacing broken equipment, the line of authority with regard to record keeping and accounting clearly went through the DRAM office of NPF. Major Doyle confirmed that he did not “interact” with Mr. Mullins in regard to such paperwork as daily sales reports, deposits, 602s, as such things were “not part of my role.”
[80] The accused himself had previously worked for twenty five years with another branch of the federal government, and then a number of years as a labour standards investigator for the province. I am incredulous of the idea that he would not recognize the importance of standards and would not understand that he was working in a policy-regulated environment.
[81] Whatever he may have learned from Ms. MacKay, he acknowledged that “she did not sign off on paying people cash.” Bar operations were the central business feature of the Messes, their primary source of revenue. While it is possible that some people at the local level tolerated the practice of paying cash, it is clear to me that the accused deliberately hid this practice from NPF.
[82] I note here, as elsewhere, that Mr. Mullins was very lax in regard to issuance of approved receipts (602s) for monies received by the Mess. Having professed to having no idea of the need for documenting receipt of money according to a pre-approved form until one year into his tenure, he admitted to not using 602s when he received money from third-party renters even after this. He said he would give the third party a receipt to show they paid for the bartenders, a receipt from his own private receipt book. He rationalized this by saying that “this money was not going through NPF, it was just a direct transaction between the renter and bartender.”
Cleaning services
[83] After Mr. Mullins was hired, difficulties arose in regard to cleaning of the Messes after private functions. Regular cleaning staff declined to perform this service. Although not part of his job description, the accused decided to take this on in order to make some extra money. A fee for cleaning was included in the negotiated price of any rental agreement. Mr. Mullins paid himself directly from the monies received from person or organization renting the Mess premises. He and his wife created a business name, D and H Cleaning Services, by which some of the payments were received.
[84] Helen Mullins gave evidence about this. She confirmed that she did do cleaning at the Messes along with the accused, was paid, and the money deposited into their New Waterford Credit Union account. She remembers cleaning 7 or 8 times. Nothing in her testimony assists with issues at trial. What she said serves neither to incriminate nor exculpate the accused.
[85] This arrangement was a violation of NPF policy which dictated that a contract for services be drawn up, the Mess collect all rental fees (which might include an amount for cleaning), remit this to NPF and request payment to the third party cleaner. Hirshman said “pursuant to any contract, across the board, any money received should be documented with a 602” except money received through the cash register. In cross-examination Hirshman acknowledged that the policy did not prohibit Mr. Mullins from also being a contracted cleaner, from ‘wearing two hats’ so to speak, he insisted that a contract for services should have been drawn up and that any money collected from rentals (which might include an amount for cleaning) remitted to NPF with a request for payment to the third-party cleaner. He says that when he learned that some employees were being paid cash for certain services rendered at the Messes he told the accused there would have to be a contract whereby NPF would pay them. He says he went through the “contract for services policy” with Mr. Mullins in Sydney. A template for such is found in Ex. #9.
[86] In defence of his off-the-books arrangement the accused said that Major Doyle had agreed to him doing the cleaning in the wake of the dispute between the Garrison and regular staff cleaners. He said “Halifax was aware I was doing the cleaning”. For his part, Major Doyle, who worked with Mr. Mullins in regard the Officers Mess, referring to a “contract for services template”, said that money received by a Mess manager for cleaning should have been deposited to the Mess account and that NPF would then pay the contractor. He recalls having a discussion with the accused about he and his wife doing cleaning, at which time he advised the accused to seek out other contractors. However he “did not get involved in this”, saying he did not “actively manage Mr. Mullins and had no involvement in the Mess bookkeeping.” Here, his evidence appears to contradict the accused’s.
[87] For her part, Lindsey Reynolds was aware that Mr. Mullins was cleaning but was unsure how these arrangements were made.
[88] Dayna Enguehard identified from Ex#5 an NSTU cheque in payment for use of the Sydney Mess, in particular an “additional fee” for cleaning written directly “to the person responsible”, i.e. the accused. Whereas the NSTU was billed once per year for its meetings held there, there was no bill rendered for this cleaning fee, and no receipt issued to her for it.
[89] It appears that the accused and his wife did indeed perform the service they took the money for. It appears others were aware of the fact he was cleaning at private functions for extra money and did not take steps to stop it. There is no indication of theft here, but Mr. Mullin’s failure to abide by NPF policy and the casual way he handled cash reveals something about his overall mindset and casts a shadow over certain aspects of the evidence. Even if the accused was given the impression by local officials that they were “OK” with the arrangement, he must be taken to know that NPF had its own reporting requirements. Those, he chose to ignore.
Rental Agreements
[90] As there was no prescribed form for rental contracts between third parties and a Mess, nothing is made of the fact that the accused, in conjunction with the Mess presidents, developed and utilized their own form of agreement. However the form of contract did not change the financial reporting requirements. There was no reason to think that the use of any particular form brought associated financial transactions outside NPF requirements. For instance, Hirshman agreed that there would be nothing wrong with charging out bartender services in a rental agreement, but said that any such funds received had to be “routed through” NPF.
Training / Use of forms
[91] Mr. Hirshman first met with Mr. Mullins in July of 2014, when Mr. Mullins had been hired. Hirshman had to travel from Halifax for this, and I accept his evidence that the specific purpose of his visit was to give guidance and instruction to the accused and the authorizing officers (regular forces officers with signing authority and operational authority over the Messes) about accounting procedures for the Mess operations. He gave them the Personnel Support Program (PSP) policy manual. He thereafter went to Sydney on a number of occasions to review the procedures, the chain of authority, documentation and the processing of funds. The only other formal training appears to be two days in Halifax at the NPF office, attended by the accused and commanding officer Col. Gallant, during which they did a cursory review of certain forms.
[92] For his part, the accused says that Hirshman mentioned the 601, 602 and 603 forms at their first meeting in Sydney but said “there was no discussion of how to use them.” He said he had to rely primarily on existing files to learn what to do. Elsewhere in his testimony he claimed that he was one year into his job before he learned what a 602 was for. This seems unlikely. Mr. Beaumont said, of the visit by Mr. Mullins and Col. Gallant to his office, that “there wasn’t a course per se – we did show them the forms”. While the training and orientation may not have been as complete as Hirshman suggests, the adequacy of such must be viewed in light of Mr. Mullin’s own background, in light of the training and experience he brought to the position. Indeed, Mr. Mullins testified that he was viewed favourably for this position of Mess manager because he was familiar with bar operations from prior experience. Coupled with what the accused already knew, it should have given him more than enough understanding of what was entailed in running a Bar, keeping proper records, awareness of possible skimming, ensuring the integrity of cash flow and sales, etc. He would have been aware of these basic requirements already; the ‘training’ would involve familiarization with NPF’s forms and reporting procedures. This was not complicated. He acknowledged having access to the NPF office by phone if he had questions about something.
[93] Hirshman says he gave the accused official receipt books. Each 602 form was individually numbered. Books of 602s were ‘signed out’ of the NPF office. These, and purchase orders (601s), were “the two key components where it was important (to) follow regulations regarding their use” according to Hirshman. This was also spelled out in Chapter 9 of the PSP manual. He said the requirements around 601’s were less strict, that invoices signed by officers could be used instead, but that the 602 “was one of the most important documents for us.” A “603” was to be completed where goods were transferred out of inventory for some reason. Most often this was simply a transfer between Messes. Given that Hirshman went to Sydney for the express purpose of training a new Mess manager, it is difficult to believe that he would have done less than explain the use of these forms to Mr. Mullins, in much the same way he explained it to the court at trial.
[94] Hirshman said there was no one form for a daily sales report (DSR), that each Base had its own. This was deemed acceptable so long as it gave NPF necessary information for its accounting system. DSRs were used to track revenues. Hirshman says he reviewed the use of DSR’s with Mr. Mullins as well, and that there were many follow-up calls. While Hirshman also explained to him how cash register tapes (z-readings) were used to track sales and deposits, this level of verification was sometimes absent at the Sydney Messes owing to problems with the cash registers, as discussed above.
[95] Colin Rankin succeeded Mr. Mullins in the position of Mess Manager. He was followed by Trina MacDonald who had worked as a bartender at the Sydney Messes in the summer of 2016. She became temporary manager there in March of 2017 (after the relevant period) and then permanent Supervisor some time later.
[96] Ms. MacDonald spoke critically of the lack of training given to her when she first assumed responsibility for the Mess. She says she was given little to none when she assumed temporary responsibility, but only given numbers to call if she had questions about reconciliations and the like. Her predecessor, Mr. Rankin, did go over things with her, but she “had expected more”. When she was hired permanently, a Mr. Raymond “came down for one day” to go over policies and procedures. This mirrors what Hirshman did for the accused when he was hired. Although she expressed surprise at the lack of training, she indicated in re-direct that it usually took only one call to NPF “to learn what I needed to know.” One aspect of this was completing 602s. She said it was “mainly by making calls to accounting did I get help to explain how to fill things out”.
[97] Defence appears to argue that failure to follow NPF procedures, especially around the use of forms, is not the fault of the accused but reflects a failure of oversight and direction. However, Ms. MacDonald’s evidence does not suggest to me that the training and direction the accused received was inadequate. She was ‘trained’ by someone other than Hirshman. She and the accused are different people who came to the job with different backgrounds and competencies.
Cash found in the Accused’s Office
[98] Mr. Hirshman testified that the “net cash” of Sydney Messes had decreased by $50,000 between March 2015 and September 2016. Some thought this was the fault of the NPF accounting system. This, plus concern over cash registers prompted Hirshman to visit Sydney in September of 2016. He went back again a short time later after a complaint was lodged. This is when he met with Mr. Mullins and seized various documents from in and around his office, and his work computer. This is when Mr. Mullins was relieved of his position.
[99] During Hirshman’s final visit to Sydney he searched the accused’s office and found $2,800 cash “locked in a filing cabinet totally away from his office.” The accused testified that this was proceeds from an event the previous Saturday (the money was found on a Monday). Although I accept that this would give little time to make a bank deposit, and while I make nothing of the location where it was found, the fact that Hirshman did not locate any accompanying documentation is very incriminating. Tellingly, there were no cash register receipts nor a daily sales report to correspond to this large amount of cash.
[100] Just two months previous, in August of 2016, Mr. Mullins and a Mess officer at the Base were told that a deposit of $4,700 into the NPF account was deemed suspicious because it was such a ‘round number’ (i.e. the unlikelihood that a bar function would generate revenue coming to such an amount) and because of the delay in depositing it. As well, “it was all new twenty dollar bills.” Referring to this, Mr. Mullins testified that Hirshman believed “there was a thief among the staff”. He explained that he had simply changed out all the five and ten dollar bills for twenties and went on to say that when they “balanced inventory” they weren’t off by much for the event, which had required 6 bartenders. This leads to the observation that the accused was aware that DRAM suspected that something was amiss at the Sydney Messes. Although the accused spoke of this when discussing the early period of his employment, and Hirshman spoke of it as having occurred in July of 2016, the same observation applies, i.e. that the DRAM had put the accused on notice about stealing, yet subsequently he was found with a large quantity of undocumented funds in his custody. Would an innocent but falsely-accused person not take greater care, knowing he was under the microscope, so to speak? I take the accused’s claim that the $2,800 represented proceeds from a recent wedding with a grain of salt. The absence of documentation suggests strongly that the accused had no intention of ever creating such.
[101] Hirshman deposited the amount in the Sydney Mess account as ‘bar sales’. Had he not done so the inventory discrepancy calculated by Ms. Sullivan would be even higher because sales revenues were compared to what was in inventory. Adding this $2,800 to bar sales reduces the observed discrepancy. As I see it, money that Mr. Mullins intended to steal was restored to the Mess account.
[102] Hirshman testified that as of September 2016 the Sydney Messes, which had been financially self-sufficient, were “going into the red” and that deposits were not being made in a timely fashion. This is substantive evidence of such, coming from him as DRAM. The unaccounted-for wad of cash he found when he went to Sydney supported his suspicions; taken together it makes for incriminating evidence against Mr. Mullins.
[103] While he testified about the lack of training he received, the accused described what was covered in his initial meeting with Mr. Hirshman in July of 2014. The accused says they mostly discussed purging existing files and the lack of supporting documentation for various things. As an example he said that upon taking over he found $1,400 cash in an envelope with no covering paperwork, no explanation as to source. He says Hirshman told him simply to deposit it because it concerned the previous Mess manager. Mr. Mullins says he disagreed with sweeping this matter under the rug. Yet an even larger quantity of undocumented cash was found in his possession in September of 2016. If he was so concerned by the carelessness of his predecessor, it is curious that he would subsequently engage in the same behavior.
Unofficial receipts
[104] Mr. Beaumont testified that a 602 form should have been prepared and issued any time money was received by a Mess, including for rental fees, whether the payment was in the form of cash or cheque. He said “the only receipt we use is a 602”, books of which were sent to the Messes from his office. Mr. Hirshman said that any time a Mess manager was handed money, an official form of receipt was to be written – “I informed him (the accused) of this personally when I met with him in July of 2014.” He said that unofficial receipts were not permitted because they gave NPF no record.
[105] The day Mr. Mullins was relieved of his job, George Hirshman found both NPF-approved receipts (602s) and a “blue-lined” (unofficial) receipt book. The latter was located behind the accused’s filing cabinet.
[106] Mr. Hirshman prepared a summary of these receipts, contained in Ex.#9. Where his summary shows a nil or blank, he explained that this meant he could find no corresponding 602. Only three have corresponding 602s - these relate to mess dues and bottle returns. Hirshman said he “compared the blue-lined receipt book with everything I had in front of me”. He said he went through NPF’s books for Sydney and that there was no indication that any of the others (other than the 3 noted) had been entered on a 602 – “names, dates, values – none had been entered on our system.” Altogether these totalled $4,835 as being received by the accused.
[107] The accused says the book seized by Hirshman was the only ‘unofficial’ receipt book he used.
[108] When shown the “unauthorized receipt book” from Ex.#9 Major Doyle said he had not seen these before and “was not aware that Mr. Mullins was using these types of documents”.
[109] Lindsey Reynolds, in 2016 a Master Corporal and president of the Jr. Ranks Mess since 2014, said that the unofficial receipts were a record – “if NPF did not want to accept it they could insist on using the proper form.”. However, it seems NPF were never given these in the first place. For her part she had no involvement with them whatsoever.
[110] The accused says he knew “third parties would be looking for a receipt” but that he was almost a year into his job before he found out that he was supposed to be using an official receipt to track monies received. He says only then did Mr. Beaumont advise him of the need for a 602. By his own admission, he went on to violate this advice.
[111] Mr. Mullins claims unawareness of an official form of receipt, even though he speaks about using another official form, a 603, upon the advice of the same Mr. Beaumont, to track inventory spoilage. Given his background, and the evidence of Hirshman that the orientation Mr. Mullins received upon hiring included use of the proper forms, I have difficulty accepting this profession of ignorance. The accused’s own testimony on this point is implausible – he remembers Hirshman mentioning 601, 602 and 603 forms during their first meeting in Sydney but says “there was no discussion of how to use them.” Given the evidence I heard from Hirshman, this seems highly unlikely.
[112] Mr. Mullins says that once he was apprised of the official form for receipts, he thereafter used them except for payments received from renters for any bartenders’ wages. However the evidence, including his cross-examination, shows that he used the unofficial receipts more widely.
[113] These unofficial receipts constitute very poor records. Even the accused could not tell, in retrospect, from the documents, what they were for or the precise date they were issued. A date such as “2016-24” seems suspiciously ambiguous. In regard to a receipt bearing the date “2016-25” he says “it would be the 25th day of the month” but cannot say what month. In regard to one which reads “09-2016” he testifies “I’d say this was written in September of 2016.”
[114] Shown one particular receipt Mr. Mullins says that “if it was a hall rental it was definitely deposited” and yet he could not ascertain from the self-same record whether it was for a hall payment or not. During questioning on his own “unofficial” receipts he was often unable to identify the person from whom the money was received, the date of the event, the date of the receipt, and whether it was for rental, cleaning or bartenders.
[115] A receipt issued simply for “cash $500” records nothing about what it was payment for. The accused says he distinctly remembered such an amount being put under his office door in an envelope. It beggars belief that he could never ascribe this to an actual event, yet at trial he still had no idea what it was for. The best he could say was that he thought it was for a wedding, that it could have been for hall rental or for bartenders. Where the receipts show no name for the person from whom it was received, the accused says he “might have been in a rush.” He admits that one cannot tell from the document what the receipt was for. This indicates that he never tracked such money and never intended to. It supports the Crown’s argument that he deliberately obscured the receipt of funds.
[116] Ms. Sullivan was unable to tie any of the unofficial receipts to other records. She could not even connect the one with a proper date to any known event or transaction.
Karmac account
[117] There was some considerable evidence at trial regarding an account at the New Waterford Credit Union, held by the accused and his wife, which was used for a car repair business they conducted for a time, and which they utilized for personal banking. This evidence, which includes the analysis done by Ms. Sullivan, neither incriminates the accused nor serves to exculpate him. It is equivocal, as Crown has conceded. The evidence of Helen Mullins about the account neither confirms nor dispels the possibility that the accused appropriated cash from the Sydney Mess operations, whether for cleaning or otherwise, and simply spent it.
Inventories
[118] In the discussion of Ms. Sullivan’s report I will refer to various stock-takings. These are physical counts of liquor, by the bottle and/or the ounce. These were done by visual inspection or, when done by the accused, by weighing. They usually involved more than one person, counting or supervising. On the evidence before me there is no reason to doubt the accuracy of any stock-taking, even though none was done by an independent outside agency.
[119] According to the accused, as corroborated by a former bartender, stock-taking was done on a day-to-day basis. These should pick up any discrepancies between reported sales and actual, measured diminution of the liquor supply.
[120] I note however that the maintenance of such records does not exculpate the accused, in the sense that it would still be possible for him to defalcate funds, stock-taking or not. Indeed, it is possible that someone might go through this exercise to create an appearance of propriety. Mr. Mullins did not retain records of his daily inventory counts.
[121] Mr. Mullins, at times, seemed to use the terms ‘audit’ and ‘inventory’ interchangeably. He said one had been done by Warrant Officer Slade in July of 2014, and mentioned a second just prior to Christmas of 2014, after which he says he was told to do these inventories himself. Specifically, he says that Hirshman told him in the spring of 2015 that if he wanted an audit to “do it on his own . . . same thing with inventory.” As noted, it appears this is just what Mr. Mullins did.
[122] There were three thorough stock-takings done, in October of 2013, March of 2015 and September of 2016. They were sometimes referred to as the “formal stock-takings”. I take it that these were more extensive and more careful than daily counts. They entailed supervision by personnel completely uninvolved with day to day Mess operations. The documents show multiple people ‘signing off’ on the inventory count-sheets.
[123] Hirshman testified that it was the accused’s responsibility to do an inventory when he started out in his position, but that the March 2015 inventory was “the only one I could determine he’d done.” Consequently, in his investigation of possible theft, “this is as far back as I went.”
[124] The March 2015 count, done late in that month for each of the three Sydney Messes, becomes the starting point for the period over which the alleged diminution of inventory (unaccounted for) may be analyzed. In other words, the opening inventory for the “relevant period” is taken as April 1, 2015. The September 2016 count constitutes the end point for Ms. Sullivan’s analysis and so the “relevant period” ends on September 29, 2016.
[125] The March 2015 count constitutes the starting point for the forensic analysis, but Ms. Sullivan explained how that “starting number” was itself derived from the 2013 count, utilizing cost multipliers. Since that count had been done “at retail” she employed the cost multiplier in effect for each individual Mess to get “the amount they had available to be sold, at cost” as of April 1st, 2015. The March 2015 stock counts were performed by the accused himself, with oversight. According to Ms. Sullivan this count was reflected in dollar terms “at retail cost”, or what the Mess was selling it for. By contrast, the count in September 2016 was done by Hirshman, with oversight, on a wholesale cost basis. Ms. Sullivan accounted for this difference in her analysis.
Forensic accounting evidence
[126] Roberta Sullivan, CPA, CA is employed by Forensic Accounting Management Group, a division of PSPC Canada. They provide services to various entities including the RCMP and Department of National Defence. She also maintains a certified forensic examiner designation. In this case she was retained by DND to analyze the accused’s personal banking record and also the sales and revenue records of the three Sydney Mess Halls at Victoria Park. Defence agreed that she was able to provide expert opinion evidence about the financial records kept by the Sydney Messes.
[127] Working with material supplied to her by Military Police, who obtained them from DRAM, she examined the general ledger maintained by NPF for the financial transactions of the Messes - sales, deposits, purchases, inventory, etc. - as well as loans and accounts of the accused. She also obtained original paper documents filed with NPF in Halifax by the Sydney Messes – these being the “source documents” from which NPF clerks made data entries into their internal accounting system.
[128] What I described above as the “relevant period” extends from 1 April 2015 to 29 September 2016. This is the period Ms. Sullivan examined, analyzed, and drew conclusions from.
[129] Records of financial transactions during the relevant period were maintained by NPF with Oracle accounting software. Documents were provided to Ms. Sullivan in electronic format. In addition, she was provided with original records, paper versions of documents similar to those provided electronically, which she also examined. She scanned all of these and constructed an inventory in Excel format as a reference tool. Both the data found on Oracle, which I will refer to as the GL (general ledger) and her scans of the original documents, were included on a USB tendered as Ex#3. She produced a written report which was tendered into evidence as Ex.#2, and testified in support of it.
[130] Her analysis was done separately on each of the three Sydney Messes, but concluded with summaries combining all three. This fit the manner in which they were operated. The accused managed all three, stock was often intermingled, at times cash registers were moved between messes, etc. No one mess operated independently, and so they were regarded, appropriately, as one large operation.
[131] The foregoing trove of data and documents form the underlying foundation for Ms. Sullivan’s opinion. Most of this evidentiary foundation was not specifically commented upon at trial, but I am satisfied that it was all carefully considered by the expert. Various document types were displayed and explained. All were available to the Defence for cross-examination purposes. It is available to the court to peruse, although I did not intend to undertake the daunting task of examining each of the thousands of documents. There was extensive evidence from Hirshman, Beaumont and others about the accounting system, record keeping, etc. The documents and data are in evidence. They “speak for themselves” in addition to supplying the expert with a firm factual foundation for her analysis.
[132] As noted elsewhere, Ms. Sullivan agreed that it was not possible for her to “tie in” bank records. However NPF’s accounting system, and the source documents both allow for a financial analysis. While a true bank reconciliation would add a layer to this analysis, Ms. Sullivan was able to provide an opinion based on the records which she had. I find that her conclusions are entitled to considerable weight.
[133] Stock taking, or stock counts, were critical to the forensic analysis. As noted above, these are counts of the liquor inventory. From the records provided Ms. Sullivan “isolated documents relevant to the March 2015 stock taking”. She did this on a mess-by-mess basis and then aggregated the results in a summary. The “count sheets” for the March 2015 inventory-taking were signed-off by the accused and one (or two) others. These counts were taken as accurate, although corrected for errors detectable from the documents themselves. Schedules 1.1.1, 1.1.2 and 1.1.3 - Warrant Officer / Sergeants Mess, Officer’s Mess and Jr. Ranks Mess respectively - were combined in Schedule 1.1 to show the aggregate amount. Ms. Sullivan took this as the opening inventory for the Sydney Messes as of April 1, 2015.
[134] Schedule 1.3 shows the calculation of purchases during the relevant period for all Messes (sub-schedules again break the amount down on a mess-by-mess basis). Schedule 1.4 shows aggregate sales and deposits during the relevant period, with similar sub-schedules. Both the cost of purchases in 1.3 and the sales in 1.4 were re-created from supporting documents.
[135] Schedule 1.2 shows the summary of all Messes of the inventory counts done on September 29, 2016. Again, the sub-schedules break these counts down on a mess-by-mess basis.
[136] Finally, as regards the study of inventory, Schedule 1.0 (also Table 6.0) shows the inventory count reconciliation for the Messes, with accompanying sub-schedules for each. This is the basis for the “variation to be explained”, or inventory discrepancy, based on the analysis of the source documents.
[137] Schedules 1.5 and 1.6 are the purchases and sales, respectively, as taken off the Oracle accounting system. They are the counterparts to Sch. 1.3 and 1.4., based on what showed in the GL as opposed to what was revealed by the source documents. These give rise to Table 6.1 and yield a slightly different “variation to be explained”, one which is about $5,000 lower than the figure resulting from the document-based analysis.
[138] The term “variation to be explained” does not itself prove wrongdoing on anybody’s part, much less suggest that there is any onus on the accused to explain it. This is but one piece of evidence, albeit a very important one.
[139] Ms. Sullivan observed that there were three different accounting systems at work: (i) the daily sales reports and other records kept by the Messes, (ii) the Oracle accounting system based on records sent to Halifax by the accused, and (iii) the banking system into which funds were ultimately deposited. She explained that “in a perfect analysis” the above-mentioned schedules would be tied to banking records, but as the Sydney Messes could not be isolated from the merged NPF account she “could not make the final tie-in”. While NPF should have been reconciling banking records regularly, the failure to do this did not preclude Ms. Sullivan from performing a forensic analysis. Despite this shortcoming, and despite the limitations clearly expressed, the records that were available to her did permit a valid analysis worthy of considerable weight.
[140] The most important aspect of the expert report is the inventory reconciliation. Ms. Sullivan focused only on liquor sales and purchases, even though a small amount of other product such as mix and snacks were sold across the bar. As noted, her reference points were two stock-takings the first done on March 27 and 30 of 2015 and the second on September 29, 2016. She recreated the count sheets for both from the records.
[141] Using cost multipliers in effect at 2015, one for each Mess, the amount at cost was obtained, permitting a comparison with the September 2016 counts which were done at cost by Hirshman.
[142] Ms. Sullivan took the 2015 counts as accurate, noting that they were done by multiple people, including the accused, although she did recalculate the math after noting some small errors (Sch.1.1 et seq).
[143] Ms. Sullivan reconciled inventory based on the Oracle records but also, because of possible errors in the GL, re-created with supporting documentation (liquor invoices, sales slips, etc.) to ensure all entries were in fact for liquor. I conclude Ms. Sullivan’s re-creation from source documents to be the more accurate. She created data sets based on the “hard copies” of relevant documents in NPF’s files which included such things as daily sales reports, z-tapes (where available) and deposit slips. She spoke to certain source documents displayed in court during her testimony and went through Schedules 1.3 and 1.4 in some detail to show how she utilized all the available documents to create them.
[144] Ms. Sullivan appeared to be very thorough. An example would be noting a $0.03 error in the physical count sheets from the March 2015 stock taking at one of the Messes. Similarly, for this stock taking, calculation and multiplication errors surrounding the retail cost of wine per oz. were detected and a $1,540 error corrected. She cleansed the records of mathematical and accounting errors.
[145] Purchases at cost during the relevant period were calculated from Nova Scotia Liquor Commission cash register receipts (cost of goods to Messes). This was added to existing inventory as of March 2015 (also reduced to cost basis) to give the total cost of goods available to be sold during the relevant period. Using Mr. Hirshman’s ‘cost multiplier’ she calculated the value of inventory at cost which ought to have been present as of September 29, 2016. This was compared to the inventory at cost determined by an actual stock count. The difference is a variation which she has referred to as an “inventory discrepancy”.
[146] Ms. Sullivan had evidence of purchases not recorded in Oracle but conversely Oracle showed purchases she was unable to tie to source documents. Hence there were “differences on both sides, going both ways.” The analysis from source documents yields a somewhat larger inventory discrepancy than the analysis done purely from the GL data; I regard the former as the more reliable (Table 3.1 vs. Table 3.0 of the report). One reason she performed this second, more pains-taking analysis was the fact that the use of Oracle changed during the relevant period.
scope limitations
[147] I will briefly comment on some limitations discussed in the report.
[148] The lack of sales data during the few days between the late-March stock-takings and April 1st, the beginning of the relevant period, is unimportant. Sales, if any, during these few days would be a negligible fraction of those in the relevant period.
[149] As noted above, records for revenue streams such as hall rentals or members dues were so incomplete that potential discrepancies in these areas could not be analyzed. As a result the important aspect of the report is for “things which were included in the stock-taking.”
[150] The movement of cash registers between the Sydney Messes, sometimes required if one or more was not working, is of no matter. The report treated all the Messes as one. What matters in the end is not the mis-attribution of a sale to the Jr. Ranks Mess as opposed to the Officers Mess, but the totality. As Ms. Sullivan said, “the summary is really the key.”
[151] Ms. Sullivan notes the unavailability of some “z” tapes from registers, the report generated by the register itself which details sales for any given day. “If I had z-tapes I could determine whether the daily sales reports agreed with what was rung through cash”. Bar sales were recorded on these DSRs and tied to a bank deposit slip (and sometimes to a “z” tape). The additional information the “z” tape would give would be a further check on accuracy, but this does not undermine the reliability of what the expert was able to do, utilizing the documents available to her.
[152] Ms. Sullivan was unable to determine whether the $4,835 in unofficial receipts were deposited and accounted for. She did not include this number in her analysis or her calculation of discrepancy. The existence of such documents raises the possibility that Mr. Mullins may have received/taken this money over and above anything suggested by the forensic report. I see this as another example of the careful approach Ms. Sullivan took to her assigned task. Another instance of this is the fact that she did not include in her calculations any increase in sales prices at the Messes during the final four months of the relevant period. By contrast, Mr. Hirshman had adjusted the inventory for the final stocktaking to account for such an increase. There well may have been such, but Ms. Sullivan appears to have taken a cautious approach. Her decision not to include this supposed price increase favours the accused – Ms. Sullivan testified that if she had “it would increase the discrepancy I calculated.”
[153] One accounting function which significantly affects the accuracy of the opinion is the “cost multiplier” employed by the expert. A cost multiplier approach facilitates inventory counts. It is less cumbersome than calculating actual mark-up on every product sold. The number Ms. Sullivan employed was calculated by Mr. Hirshman. Ms. Sullivan did not call into question the use of this accounting tool, indeed she seemed to recognize and accept it, but she could not vouch for the accuracy of the number she was given. I will return to this highly disputed aspect of the case below.
assumptions
[154] Ms. Sullivan assumed that about $2,000 of invoices showing dates in 2014 actually occurred in 2016 and were hence included in her calculations. If this assumption was incorrect it would operate in the accused’s favour by decreasing the inventory discrepancy by this same amount at cost. Given the great care shown by the expert in other facets of her analysis, I conclude that this was a reasonable assumption.
[155] The most significant assumption made was a breakage rate of 0%. This refers to liquor lost from inventory, not sold, due to spillage, breaking of bottles, spoiling due to heat, etc. Defence places great importance on this issue, warranting careful and somewhat lengthy consideration, and hence I will return to it below.
conclusions
[156] Depending on the data employed, whether Oracle-based or document-based, the inventory discrepancy determined in the report is either $12,778 or $17,842 (on a cost basis) respectively. Ms. Sullivan said the amount unaccounted for was “between” those numbers.
[157] Her formula for arriving at these sums was: the March 2015 stock-taking (adjusted for errors) + purchases during the relevant period – {sales x the cost multiplier} = closing inventory at cost. This was done for the three messes, and then consolidated. Subtracting the September stock-taking from the above result gives the “inventory discrepancy”. She concludes that “based on the documentation I was provided . . . the amount unaccounted for is between 12,778 and 17,842 dollars.” As she also put it “$12,778 is seemingly missing” based on the Oracle accounting records; “$17,842 is seemingly missing” based on the documents,.
[158] The above is summarized in Table 7.0 which shows the inventory discrepancy “at cost” as $12,778 (per GL) and $17,842 (per source documents). The variation “at retail” is $26,895 and $37,555, which would represent lost sales proceeds. This conversion from cost to retail value again utilizes the cost multiplier. As Ms. Sullivan put it, “if bottles are missing, cost is what matters; if cash is missing, retail value is what matters.”
[159] She concludes her summary by saying that because most bar sales were done on a cash basis it was not possible to determine the cause of the discrepancy. I take this to mean that she cannot say that the loss can be attributed definitively to any one of possible causes, i.e. pilfering of inventory, theft of cash, spoilage, gifts, etc. However she did testify that spoilage would have to be an extremely high to account for the discrepancy (I return to this below). She could not say whether liquor may have been removed, nor who, among the various people with access to the cash, might have stolen any particular sum. As to whether the loss was actual stock (bottles of liquor) or proceeds of sale of the stock, Ms. Sullivan said “I cannot tell how much of each, but this is what is missing based on what is unaccounted for.”
[159] In cross, Ms. Sullivan again acknowledged that she could not say whether the discrepancy is missing money or missing inventory, nor definitively say whether theft, or spoilage, or sloppy record-keeping might explain the discrepancy. Nor, if there was missing liquor or proceeds, could she say who is responsible. She also agreed that nothing was revealed from the accused’s personal finances which would show clearly that he was stealing.
[160] The fact that Ms. Sullivan cannot say whether inventory or cash or both are missing is not a weakness and does not undermine her conclusions. The product and the proceeds of sale of the product are equivalent for purpose of analysis. The evidence at trial indicates that what went missing was cash, given the security measures in place, practices at the bar, and the portability and relative ease of removal of dollar bills. $17,842 worth of liquor represents a goodly number of bottles.
[161] In her calculation based on accounting records, Ms. Sullivan included the cash found in the accused’s office as “sales recorded during the period”. In her calculation based on sales supported by documentation she did not. If this money had not been found and accounted for, this would have increased the discrepancy based on accounting records. The deposit of this money and its inclusion as sales – as Hirshman did - reduced the discrepancy which would otherwise have emerged from the GL accounts and served to put the accused in a better light.
[162] Because DRAM could not determine whether the $4,835 in unofficial receipts were deposited (nor could Ms. Sullivan given that only one of them bore a date) “there is a potential for an additional 4835 to be accounted for.” In other words, it appears the real discrepancy is higher than she is able to demonstrate in documentary fashion.
other aspects
[163] The Messes had other revenues besides bar sales, most notably membership dues and facility-rental fees. Ms. Sullivan was unable to determine whether such revenues were accurately recorded. At p.25 of her report she notes that “records available for these revenues were incomplete, therefore I was unable to determine if there were any discrepancies within these streams.” While her opinion gives no support to the idea that these revenues were “skimmed off” by the accused, it is fair to consider the simple fact that such records were not kept. Ms. Sullivan could not tie the rental agreements (Table 6.3) to specific deposits. She says at p.19 that “it does not appear that the rental fee revenue recorded is complete. It is not possible to determine the extent of the discrepancy”. She said that for nine of the events, she “could not tie a daily sales report to the date of the event . . . most rentals included bar service and one would expect to see bar sales on date of event recorded in the daily sales reports for each of these days.” Given that it was Mr. Mullin’s responsibility to keep such records, this is not flattering.
[164] The report also details the accused’s personal banking records (“Karmac” account, discussed above). The analysis of this account does not reveal any significant improvement in the accused’s financial position. I am unable to draw any inferences of guilt from any of this. However this does not make these account records exculpatory. If the accused stole cash from the Mess, he may simply have spent it rather than deposit it into this account or some other.
Cost Multipliers
[165] The “cost multiplier” devised and employed by Mr. Hirshman, and adopted by Ms. Sullivan in her forensic analysis, is one of two issues raised by the Defence which I will consider in greater detail. (The other was the matter of spoilage,
below).
[166] As I understand it a cost multiplier provides a means to convert a liquid amount to a dollar amount, either a cost amount (at the point liquor is purchased from the NSLC) or a retail amount (what it is sold for ‘across the bar’). The difference between these is the profit made by the Messes on liquor sales. Because a great variety of beer, wine and spirits are stocked and sold, and because the cost price and, to a lesser degree the sales price of these may change, it would be extremely difficult to determine the actual markup on every ounce or bottle or glass sold during the relevant period. A cost multiplier provides a useful shortcut, but the trade-off is that one gets an estimation.
[167] As Ms. Sullivan said, a cost multiplier is used to convert recorded sales to an estimate of their cost, rather than going through every sales invoice. She said “it would be almost impossible to use individual invoices because you can’t tie each bottle on hand to any particular invoice.” In cross-examination Hirshman said that by using a cost multiplier to measure inventory he “made some approximations” and that employing the device meant “cutting corners, leaving exact math behind.”
[168] As discussed above, comparing an actual liquid-amount stock count (as was done in March 2015 and September 2016) to purchase and sales data for each of the measured categories permits a comparison (reconciliation) of the figures obtained by each. The difference (discrepancy) is expressed in a dollar amount, on either the cost or retail price point. The former would represent loss to the Messes of monies it spent to acquire the stock. The latter would represent loss of profit on stock that was actually sold to customers at the bar. The method shows what expected sales should be, based on the cost of acquisition. Comparing this to actual sales figures yields a “variation” (discrepancy). The method can not distinguish between possible explanations for such discrepancy, i.e. whether (i) actual bottles of liquor were taken, (ii) whether sales proceeds (here these were exclusively in the form of cash) were stolen, or (iii) whether stock was lost due to spoilage, or from being gifted away, without any accounting of such.
[169] Proceeds of sale at the Mess bar were overwhelmingly for liquor. The price of each drink represented a certain and defined quantity of liquor (oz. of wine, spirits, beer) and can thus be converted to such. But because liquor is purchased at one price (from the NSLC or private licensed vendor) and sold at another (at the bar) a cost multiplier can be employed to convert from one to the other. Inventory can thus be expressed in units of alcohol (counted and/or weighed) or in dollar terms (at its cost price, or at its value at retail). This methodology permits a comparison of apples with apples, i.e. cost with cost, at both the beginning and end of the relevant period. The conversion to cost value, and vice versa, using a multiplier is thus an approximation. It is a short-cut to determining the actual mark-up in the price of every bottle of wine, spirits and beer sold at the Messes during the entire period. It avoids the almost impossible task of tracing each individual bottle of liquor back to an actual purchase at the NSLC.
[170] As Ms. Sullivan put it, there were three “official” (more complete and formal) stock-takings – in October 2013, March 2015 and September 2016. And so there are two accounting periods during which a cost multiplier approach to inventory was adopted, the first being the period October 2013 to March 2015. For this period, three different cost multipliers, one for each Sydney Mess, were used to calculate the inventory of liquor on-hand as of late-March of 2015. As noted, this is what Ms. Sullivan used as the opening inventory on April 1st, 2015, the starting date of the relevant period. These are noted in the first of the assumptions at s.2.4 of her report.
[171] The documents Ms. Sullivan reviewed describe these three multipliers as “CANEX approved multiplier until next stock-taking”. Mr. Beaumont and others indicated that CANEX is simply an arm or aspect of the operations of CFMWS. In his redirect Hirshman distinguished between cost multipliers utilized for the 2013 – 2015 period and the single one he computed for the relevant period. The former were “system-generated” whereas the latter was devised by him. Ms. Sullivan said that “for whatever reason” the March 2015 count was done “at retail”, subsequently converted to cost using the multiplier. At p. 7 of her report she states “the 2015 cost multipliers were applied to determine the amount at cost.” Hirshman said that the March 2015 stock counts were brought down to cost price so as “to make everything even”, which I take to mean to permit a comparison of cost value then vs. cost value later. He said in redirect that he was concerned about having a reliable starting point, an accurate figure “for what inventory was at a specific time” (i.e. late-March 2015) to “what it was at the time of my investigation” (i.e. September 2016).
[172] Defence questioning about the reliability of the cost multipliers focused on the period between the March 2015 and September 2016 inventory counts, and the comments which follow relate to that, although they pertain generally. During this, the relevant period, additions to liquor inventory (cost of purchases from the NSLC by the Messes) totalled $34,193, more than double the opening inventory of $16, 214, at cost, as of March 2015, making the reliability of Hirshman’s multiplier far more important. The fact that the earlier set of three cost multipliers was created by an agency which had responsibility for retail operations within CFMWS lends some credence to those values.
[173] Ms. Sullivan began her discussion of the report with the March 2015 “opening inventories” for each Mess. The detailed inventory count summary for all three Messes is at schedule 1.1, corrected for some small math errors. The subschedules show the count, by bottle and ounce, by each brand, and at their respective retail values according the selling price at the bar (with few exceptions, $3.50 for a bottle of beer, a 3 oz. glass of wine, and for a 1 oz. shot of liquor). She explained how the application of the cost multiplier gave the conversion from retail to cost values. She explained the rationale for different multipliers depending on how ‘high end’ a product was, and how price mark-up might be higher for such.
[174] The testimony most relevant to the reliability of the cost multiplier, and hence the reasonableness of Ms. Sullivan’s first-noted assumption, is that of Mr. Hirshman. I have considered his potential partiality in another context above. I do not think his calculation and utilization of a cost multiplier is affected by bias. Here again, what he did was open to scrutiny. I do not think Hirshman’s evidence is rendered less credible or reliable because of animus against the accused.
[175] Hirshman’s uncontradicted evidence is that sales prices at the bar of the Sydney Messes for standard servings of particular drinks was consistent across all three. He based this on inquiries – including inquiries of the accused - which, given his position as DRAM, provides a reliable basis for this information.
To derive his cost multiplier for the relevant period, Hirshman chose the most popular (by which I understand highest selling) brands of beer, wine and spirits and calculated the ratio (percentage) of cost price to sales price. This is done on a per-ounce or per-glass or per-bottle basis (as the case may be).
[176] Mr. Hirshman calculated and employed an ‘average’ cost multiplier of 47.51% for all types of liquor for the relevant period. When first questioned in direct about how he achieved this number, he said it was “a complicated calculation . . . off the top of my head I don’t recall where I got it”. It had been four years since his involvement in the investigation. While he had trouble articulating the concept at first, once he had been given an opportunity to refresh his memory by reference to the actual spreadsheet he had little trouble explaining it. He then said the matter “was not particularly complicated”, that he had used cost multipliers frequently, that the spreadsheet was based on a template already in existence, that it was something “used at work on a daily basis.” He said he only “created it” in the sense that he amalgamated the data from all three messes into one file. He said that the “worksheet . . . is where I get into the cost multiplier and where I derive the figure.”
[177] It thus appears that the Excel spreadsheet in Ex #12 is Hirshman’s design but that the program performs arithmetic functions automatically on inputted data. In this sense it operates much like a simple calculator but also sorts data and represents results comprehensively in a useful format.
[178] Hirshman said that because the examination went over 1.5 years, “reconstructing every purchase to get a complete average would require too much time” and so he used “the most popular brands of alcohol” to get the numbers. This enabled him to “come up with a general average which applied to all three clubs, because alcohol was the same price in every club”.
[179] The exhibit shows the merchandise he took to do the price comparison calculations for beer, spirits and wine. He used $3.50 as the selling price at the bar for a beer or a shot of liquor or a 3 oz glass of wine (standard servings). Using source documents he obtained the cost of these items, for instance a 1.15 litre bottle of rum, and broke it down by ounces to get the cost price of such, the “unit cost”. He accounted for the 15% HST. From these a gross profit amount and percentage can be calculated. This same process is applied to other items, and then an average taken, with the figure 47.51% being “the average of all the cost multipliers”. He said this method “was often used in accounting.”
[180] Hirshman then explained how the cost multiplier was used to re-value inventory. With an increase in bar price from $3.50 to $3.75 in June of 2016, and the corresponding increase in gross profit the cost multiplier was recalculated for the tail-end of the relevant period. Hence his cost multiplier was 47.51% from March 2015 to June of 2016, and 45.21% from then until September 2016. I note that inverting a cost multiplier gives an estimated profit margin. Here, a cost multiplier of 47.51% on liquor sales equates to a profit margin of 52.49% on such goods. (In Ex.#12 Hirshman adds $506.88 and $1121.17 for June to September “sales price increase”. Ms. Sullivan appears to have applied the 47.51% factor throughout the entire period).
[181] The latter part of the above exhibit shows the particular brands (the most popular, representing the most sales) which were employed to derive the 47.51% figure. It deals with purchases in 2015 and 2016, covering the relevant period. When asked to clarify where he obtained the NSLC prices for these brands – the amount the Messes paid to acquire the liquor, prior to mark-up – he said he used “actual invoices”. With reference to the exhibit he said “each value on here has supporting documentation – I did not pull any number out of the air.” Ms. Sullivan also says in her report that purchases shown at cost “are based on invoice and sales slips.”
[182] During the trial Defence argued that this evidence was expert opinion evidence and that Mr. Hirshman did not have the credentials to present it. I allowed the evidence, subject to weight. As I see it, the computation of a cost multiplier involves a series of basic steps, using basic arithmetic. Such concepts as cost, sales price, profit, average, etc. are matters of common parlance and understanding. It can hardly be said that deciding which brands of liquor to select as representative requires an expert; it only requires a look at bar sales or purchase invoices to see what product is selling the most. I consider the evidence is properly before me as a trier of fact. It is not the sort of evidence which requires vetting in a voir dire. Hirshman had an accounting background and work experience using this tool. Ms. Sullivan had encountered the use of cost multipliers when she worked as an accountant in the retail sector.
[183] As I understand the evidence, non-alcoholic products sold across the bar have little to no significance to this issue. By all accounts these represented a very small proportion of sales – Hirshman said it was “a minute amount of the total value of stock”. Ms. Sullivan said that snacks, water, etc. were not included in her report as “purchases”, as her analysis focused only on liquor. Defence argues that this may affect the reliability of the cost multiplier, but many varieties of liquor were excluded from the calculation as well.
[184] Hirshman noted that even if liquor were sold “off-book” at the Mess, this would not affect the reliability of the number.
[185] One sees in certain documents the occasional drink with a higher bar price (e.g. a certain Scotch at $4.50). Ms. Sullivan herself noted that some premium brands sold in the Officers Mess were somewhat more costly than the usual $3.50. But these are very much the exception, and understandably Hirshman did not employ these precisely because they were so unrepresentative. The prices, for whatever brand, appear to be consistent across the three Messes. While any cost multiplier, by definition, delivers an estimation of inventory, Hirshman’s has not been discredited.
[186] While not, by its very nature, a precise number capable of giving a precise result, I think that the 47.51% figure is reasonably reliable and is an adequate foundation for the forensic analysis which proceeds from it. I consider it capable of yielding a dependable estimation of an “inventory discrepancy” at the Sydney Messes as of September 30, 2016.
[187] With respect to any margin of error, Ms. Sullivan testified that the multiplier would need to be 70.3% to eliminate the discrepancy, based on the GL records. It would have to be 90.2% to eliminate the discrepancy, based on the source document analysis. As noted above, I consider the source document analysis to be the more reliable, and so I will apply this evidence to Table 3.0 of the report. There one sees that multiplying $41,723.56 by 90.2% yields $37,634.65 rather than $19,882.87 as the “normal percentage of sales costs”. This would reduce the “calculated value” of the inventory to virtually the same number determined by the actual stock count ($12,741.79), thus eliminating any “inventory variation” (discrepancy).
[188] One notes, first of all, how unrealistic it is to contemplate a cost multiplier of 90.2%, for this translates to a profit margin of barely 9.8% on sales of liquor. I think, without wanting to put too much personal experience on the line, that it is safe to say that bars and restaurants which sell liquor do so at a considerably higher mark-up than this. Mr. Mullins himself said that part of his mandate was to generate more income for the Sydney Messes. Sales of liquor, along with memberships, were a primary source of revenue for CFWMS, helping to sustain these social spaces for armed forces members.
[189] For instance, to take the standard drink price of $3.50 at the Sydney Messes, a 90.2% cost multiplier would say that that drink cost the Mess $3.15 to purchase at the NSLC. It makes no sense that a bar would set its prices so as to generate 35 cents profit on a liquor product which cost $3.15 to bring into inventory.
[190] Excel files found in Ex. #11A show a March 2016 inventory stock count, from which one may note the mark-up from cost price to bar price. This likewise lends support to the idea that a 90% cost multiplier is not realistic.
Spoilage
[191] Herein I use the word spoilage to mean any accidental and unintended loss of stock, including spillage. “Spoilage” seems the better choice of words because the accused contended in his evidence that significant amounts of liquor were lost due to spoilage. He attributed this primarily to problems with refrigeration, but also mentioned failure to properly seal up wine and spirits bottles which on three occasions resulted in some number of these being thrown out.
[192] Mr. Mullins and others also noted the usual loss when a bartender spills a drink or a drops a bottle prior to sale. Some sense of this, and the relative unimportance of it, comes from Trina MacDonald who worked as a bartender and said that “sometimes we might drop a bottle” but suggested that spillage was not a major problem, just “one or two bottles of beer or a shot or two” during one shift.
The Crown concedes that there would be some spillage during the relevant period, but the focus here is on the possibility that spoilage from malfunctioning coolers accounts for some significant portion of the inventory discrepancy.
[193] At some point, NPF stopped tracking spoilage of inventory at the various Mess halls under its supervision. This seemingly happened around the time they adopted the Oracle accounting software. Mr. Hirshman said that spoilage was not tracked in their general ledger; instead, NPF assumed a 2% to 5% spoilage factor and monitored any significant changes in gross profit which might warrant further inspection. He said the 2 to 5% figure “came from the experience of many military Messes across Canada.”
[194] As noted above, spoiled (lost) liquor would reduce the “goods available for resale” and hence the “calculated value of inventory at cost” in Table 3.0. Notably, Ms. Sullivan assumed a spoilage rate of zero. Hence any loss of inventory due to undocumented spoilage would reduce the discrepancy which emerges from her analysis. In addition, she agreed that the provision of free liquor at special events, if not recorded so as to show it was “taken out of inventory”, would similarly affect her conclusion. She noted however that if such had been reported it would have been captured in her report.
[195] Mr. Mullins testified at some length about the problems with the coolers. He said “power outages were common” and “caused horrendous problems”. He says he discussed this with the commanding officer. He said that he had no authority to call for repairs, that there was confusion over who should be responsible for repairs or replacements of the units. Lindsey Reynolds spoke about the need to clean the lines in some of the refrigerators.
[196] I have noted minutes of Mess meetings in the discussion of cash registers, above, which lend support to the fact that the Messes did indeed experience problems with the refrigeration equipment. In addition to the Mess minutes referenced there, I note another Officers’ Mess minute from the spring of 2016 (in Ex.#11) which indicates “We received an update from Capt. White stating that the fridges and the ice machines are in and functioning.”
[197] In addition the exhibit contains a quote from 27 November 2015 from Streamline Equipment for refrigeration repair, corroborating Mr. Mullins’ evidence on this point.
[198] Such problems may well – though not necessarily – result in loss of product which is kept cold. However there is a distinction between keeping a beverage cool to suit customer taste as opposed to loss of that product from inventory.
[199] I have no doubt that there were some problems with the refrigeration units, that repairs were required, and that there may have bureaucratic holdups. I also think that the accused is exaggerating the extent of the actual loss. He has given such contradictory and implausible evidence on this point that it diminishes his credibility on this issue, and indeed more generally.
[200] Firstly, Mr. Mullins must be taken to know how bad it would appear, and what suspicions might arise about him as the Mess manager if there were significant losses of inventory which were not properly accounted for. Given his background he would surely know the importance of documenting the loss of inventory, whether by spoilage or by giving it away for free. He would know that it would come down to him to explain and account for the resulting gaps in inventory. Indeed he testified to this very effect. He said that his first experience with spoiled product was in October of 2014 and that “in January 2015 we were losing a lot of product.” To this point, he says, he had no way to record spoilage, but that he spoke with Mr. Beaumont who advised him to file a form 603 to track such loss. Beaumont himself confirms this. After receiving a book of 603s from Beaumont “along with a photocopy of one used by another Garrison as a guide” he began to use them to track spoilage. He says “all I ever used 603s for was to record spillages.” He says he used 3 or 4 in 2014, at least 10 in 2015 and a fewer number in 2016.
[201] The context for all this is a concern about the profitability of the Messes. One can assume a heightened awareness of the need to document significant losses. At this point everyone was searching for answers as to why the Sydney Messes were losing ground. The finger of blame was being pointed at NPF. Mr. Mullins himself eventually called for a forensic audit.
[202] For his part, Hirshman did not recall any problems with refrigerators in Sydney, but he said he would expect any Mess to document significant drops in inventory for their own protection, given how this would reflect in their gross profit margin.
[203] Mr. Beaumont testified that he “told people if you throw something out keep a record of it anyway in case somebody comes back at you.” He agreed with the suggestion that there was no mechanism to capture spoilage after Oracle was adopted but added the qualifier “unless we were told about it.” In redirect he said that “there were cases where fridges broke or draft lines froze” but that he had no recollection of any large loss in Sydney. He said if there had been a large amount of spoilage an “MR (603) would be issued out and it would be tracked in Oracle, but I don’t recall any case where a large amount was lost.” He said something in the order of $300 to $400 should be noted. He went on to say that if alcohol had been provided for free a 603 “should be written up, on a per-event basis, and be put in Oracle like the rest of the data.” He looked at Sch.1.6 and did not see any such entries for the Sydney Messes during the relevant period.
[204] Mr. Mullins said “there was thousands of dollars of loss, reported directly to Halifax”. He said he “recorded all the losses” and sent documentation to Mr. Beaumont saying “he would have all these.” Strangely neither Hirshman nor Beaumont were aware of this. In order to believe the accused on this point I would have to accept that Beaumont declined or neglected to enter these losses in Oracle, even though they were reported to him in an approved form, and according to the advice he himself had given. I would be required to believe that the 603s filed during the relevant period were discarded. Why would Beaumont tell the accused to use 603s, and them throw them away when he received them? And I give no credence to the idea that these 603s were indeed filed but then deliberately withheld from investigators and Ms. Sullivan.
[205] When questioned on her assumption of 0% spoilage Ms. Sullivan said by way of illustration that if one assumed that a bottle of liquor was worth $40 (which she said was a “very high number”) that between 319 and 445 such bottles would have to be lost to account for the inventory discrepancy she calculated. She opined that this is very high given the size of the bar, representing one-half of inventory at any point in time. Nothing in the evidence suggests this kind of loss from routine bar operations; indeed, Ms. MacDonald’s evidence strongly suggests otherwise.
[206] Given the care taken by Sullivan, Beaumont and Hirshman, given what the accused himself most certainly realized about the accounting implications for large and undocumented losses of inventory, it seems quite clear to me that the 603s which the accused claims to have filed would reflect in the opening inventory for the relevant period and/or in the analysis of documents during the relevant period. That all these recorded losses, on forms designed specifically for use by NPF, would somehow disappear from the accounting records, or be withheld from the court, is simply unbelievable. While Ms. Sullivan’s assumption of 0% spoilage is not literally true – something all parties and she herself acknowledges – the accused’s evidence on this point does not call her report into serious question. If Ms. Sullivan had such 603 forms, authored by the accused, they would appear in her inventory and re-categorization of the source documents. The evidence does not disclose any statistically significant spoilage. Ms. Sullivan’s report is not undermined. Rather, it is the accused’s credibility which suffers.
[207 In a similar vein, both Hirshman and Beaumont said that liquor provided gratis at special events should have been tracked with a 603 form, which would have “reduced purchases” and shown this to be charged as an entertainment expense. Again, Mr. Mullins would surely understand all this and his accompanying responsibility. Again, there is no such documentation. He says free liquor was provided at the Lt. Governor’s Ball, at military functions, on “haggis night” and at the Mayor’s Levee. He testified that although there was no standard report he did report all these give-aways to Halifax “in writing”, and because he had to have a record “to keep my inventory books straight here in the Sydney office”. He says “I don’t think Hirshman or Beaumont kept any written record”. It is not clear why he would think this, but I have the sense that this is what he wanted the court to think.
[208] There probably was some amount of liquor provided free-of-charge from time to time, but this is relatively insignificant given the extent of the overall loss.
Bartenders as agents of loss
[209] The idea that losses from inventory or from cash proceeds of sale are the fault of bartenders invokes the concept of third-party suspects. This must be an air of reality to this. Trina MacDonald criticized security measures at the Messes, saying that until November of 2018 “nobody checked who was coming or going, or who had keys . . . it was a free-for-all”. This is but a mere suggestion that someone might have been stealing but provides nothing around which reasonable doubt might coalesce.
[210] While it is true that the various bartenders who worked at the Sydney Messes during the relevant period had access to stock and to cash, Mr. Mullins himself has testified that he performed an inventory check after each event. He says he was so scrupulous as to require a weighing of the bottles, because he didn’t trust visual estimation. The bartenders had to account nightly for their sales and Mr. Mullins would have the daily sales reports from the registers or from the bartenders. The evidence at trial indicates that the liquor was stored in a secure, locked area. If funds were being siphoned off, or liquor itself taken from the Messes, it seems the accused would have been the first to know.
Conclusion
[211] Given (a) the strong circumstantial evidence arising primarily from the forensic accounting report, (b) evidence of certain actions and behaviors of he accused and (c) credibility concerns emerging from the accused’s testimony, I am convinced beyond a reasonable doubt that Mr. Mullins stole over five thousand dollars from the Sydney Garrison Messes, and he is found guilty on count one.
[212] The second charge relates to the same conduct. Given the broad definition of theft, it here entails no fundamentally different legal elements than the first. Pursuant to the Supreme Court’s directions in Kienapple and Hammerling I am entering a stay of proceedings on count 2.
[213] It is impossible, as Crown itself acknowledges, to determine the exact amount which was stolen. However I think it is important to fix a minimum amount that has, on my grasp of the evidence, been proven beyond reasonable doubt.
[214] The evidence indicates that proceeds of sale, not actual inventory, was taken. This means I should consider either the $26,895 or the $37,555 amount from Table 7.0 of the expert report. As noted above I regard the latter figure (based on source document analysis) more reliable than the former (based solely on the general ledger analysis). That said, the expert opined that the observed discrepancy was somewhere between these two numbers, and so I will take the lower of them, $26,895.
[215] The expert report assumed a spillage / spoilage rate of zero. NPF generally makes an allowance as high as 5% of gross profit. While I think the accused has exaggerated the amount of loss by spoilage, there is evidence of some significant issues with refrigeration, and I will thus reduce the above amount by double that - 10% - giving $24,205.
[216] I will reduce the perceived loss by a further $1,000, a somewhat arbitrary figure which makes a generous allowance for liquor which may have been provided gratis (and hence lost to profits) during the relevant period. This gives $23,205.
[217] Ms. Sullivan did not include the $4,835 in unofficial receipts because from the information available she could not determine whether this had been deposited and accounted for. I am not constrained by accounting practices, and I conclude from all the evidence before me that the accused pocketed this sum. It should be added to the proven loss, giving $28, 040, which I will round to $28,000.
A. Peter Ross, PCJ