Spreadsheets Are the Devil

According to revenuerecognition.com:

92% of Public Companies’ Revenue Processes at Risk for Compliance Failures and Restatements

New study by RevenueRecognition.com and IDC reveals widespread reliance on spreadsheets, which greatly increases the likelihood of control weakness and accounting errors.

I was going to read the full report, but I didn’t want to give some unknown company all of my personal information for the privilege. If every field on your registration form is required, you don’t need to put the asterisks by each one. Just say that every field is required. If someone has access to the report and wants to give it to me without me giving them my phone number, I’ll be happy to read it.

My first reaction to this article was “Yawn”. My second reaction was that their numbers are wrong. The headline should have read “8% of Public Companies Lie About Using Spreadsheets”. Maybe they didn’t lie, maybe they just didn’t understand the question. Or maybe the surveyors asked the wrong person. Ask the CFO of General Motors whether spreadsheets are used to create journal entries. He will say in a very dignified voice “Our journal entry creation procedures are fully compliant with Sarbanes-Oxley”, or some such thing. Now ask the two-year accounting clerk that actually makes the journal entry and he’ll say “Of course we do. Now get out of my office.”

First a word about SOX. (Note that I don’t work for a public company and am not a Sarbanes-Oxley expert). SOX is the unholy bastard child of the Enron scandal. A lot of voters lost a lot of money when Enron tanked and our brave men and women on capitol hill came to the rescue. I don’t believe that anyone is above the law, including the FASB or any other accounting standards body, but just as I cringe when Senator Stevens legislates my Internet, I cringe when non-accountant politicians meddle in my business.

There are two overriding concepts that control my thinking regarding Enron, SOX, and fiduciary responsibility in general. First, the people who lost their life savings when Enron went bankrupt were speculators, not investors. Investors properly diversify their portfolios so that one company/industry/decision cannot wipe them out. Many, if not most, of the people who were hurt by Enron probably didn’t know they were speculators, but ignorance is not a defense in my opinion. Does that make me a heartless prick? Yeah, probably. I lost everything I had invested in Inacom when they went bankrupt. I didn’t end up on the dole because Inacom was less than 5% of my portfolio. Why wasn’t Congress passing laws when that happened? Okay, enough kicking the Enron-homeless when they’re down.

The second concept, and I’m not sure if anyone realizes this, is that Enron didn’t fail because of spreadsheet errors. No, it’s true. Look it up. In fact, Enron didn’t fail because of accounting errors at all, whether spreadsheet related or not. Enron failed because of fraud. We already have laws against fraud and breach of fiduciary duty and Enron executives and directors should be prosecuted under those laws. Some of them are. If SOX was in place before the Enron collapse, it would have prevented exactly nothing. Enron execs would still have defrauded the public by recognizing revenue inappropriately and by hiding losses off balance sheet.

The article states

Surprisingly, public companies with more than $200 million in revenue are substantially more reliant than the overall sample (61% vs. 52%) on spreadsheets for basic accounting entries.

Surprising to whom? First, I have to admit that I’m a little unclear on what “reliant … on spreadsheets for basic accounting entries” means, which I’ll address later. I, for one, am not surprised that larger companies use spreadsheet more than small companies, but I am a little surprised that they admit to it. The fact is, no accounting software program can cover every type of business, every type of business in an industry, nor seemingly identical businesses. There are a lot of variables in accounting, and there are many accounting software packages that do their job well. Yet not one of them does it perfectly.

When I worked for a general contractor in the construction industry, we used Timberline. It is a truly magnificent accounting package that was built specifically for the construction industry. And while they have been branching out to become a useful solution for many subcontractors, it was developed for general contractors. And yet, it didn’t account for our flexible spending account. Guess which program I used to create journal entries for that account. That was for a mid-size general contractor in a mid-size city in the middle of America. I think a $200 million company might have a few complex issues that their accounting software doesn’t handle exactly right.

What does it mean to rely on spreadsheets for basic accounting entries? Does that mean that they use Excel to compute amounts that make up the accounting entry? Does it mean that they use Excel to determine in which accounting period a transaction belongs? Are they using Excel in some way that makes determinations about generally accepted accounting procedures? Accounting systems are not black boxes. You don’t feed source documents into one end and get financial statements out the other. The process involves human beings that make errors and it always will. If you take away spreadsheets, you will be forcing users to make their errors using some other technology. Give them a pencil and a notepad and see how they do. Accountability and proper internal controls are how you produce reliable financial statements, not legislation. And certainly not by maligning the noble spreadsheet who, despite her flaws, has preformed admirably lo these many years.

Posted in Uncategorized

21 thoughts on “Spreadsheets Are the Devil

  1. For the last 6 months or so I’ve seen serveral reports from companies like Gartner and highly respective international auditing firm with the same message: “Don’t rely on Excel when it comes to the numbers!”

    Why does all these alerting reports come in a short period?

    Well the only answer I could came up with is that there exist a so called ‘cut throat’-competition within the BI-market (BI = Business Intelligence, or in more ordinary terms so call decision support systems – DSS).

    Imagine that people from SAS or Cognac try to sell their $$$-systems to companies who use the unexpensive and simple tool Excel and also claim that they are pleased with it…

    As long as people are involved in any process it can go wrong no matter of what tools are in use.

    Next time I will tell You a true story where an inexpensive Excel-solution ‘outdated’ a SAP-system ;)

    Kind regards,
    Dennis

  2. Charley Kyd (ExcelUser) has written about possible responses for Excel users to use when hearing the “spreadsheet hell” terror that some SOX-conscious people declare is not only rampant, but will only get worse.

    Recently I read about the proposal that only IT should control data, and never use Excel.

    And just to be clear: some people lost retirement funds in Enron, not because they (unwisely) invested, but because those controlling retirement funds invested in Enron without knowledge or approval of the retirees. My mother-in-law was one of them.

  3. Further to what Rich said re Enron, back before Enron crashed companies could offer little other than their own stock in their 401(k) plans, and employees’ investment options weren’t necessarily confidential. There was a lot of preasure at some companies to load up on the company stock.

    Now on to spreadsheets. It’s simple. Before companies had spreadsheets, any programs that fed accounting entries were documented. Apply the same rules to spreadsheets: no one gets to use them until their documented and someone else checks and signs off that they do in fact function as documented.

    There really are software engineering best practices, and it’s really true that nearly all spreadsheet users are ignorant of them, and that’s not a good thing.

  4. – Harald,

    When the present NAD (5 years) expire by the end of this month I will tell the story about Excel vs SAP :)

    As for the follow up on the conference I’ve asked some friends if they would like to share their impressions. My health didn’t allow me to participiate which I sincerely regret :( :(

    Speaking about MS Outlook, I switched from MS Outlook to PocoMail and since then never had any issues with corrupted files etc.

    – fzz
    From a strictly professional point of view it’s indeed simple.

    However, it’s important that the corporates take officialy a responsibility like the case I showed in the following post:
    http://www.dailydoseofexcel.com/archives/2005/12/10/relevance-lost-the-raise-and-fall-of-large-workbooks/

    Kind regards,
    Dennis

  5. As far as spreadsheet errors are concerned I think regardless of the process/program, 999 times out of 1000, the error is generated by some end user. Whether they be under-educated about using the spreadsheet or were just entering data in a hurry, human involvement = errors. I agree with Dick, regardless of what laws you impose Enrons are still going to happen. (Unless, of course, we prove that we will actually prosecute these white-collar criminals with the laws we already have in affect.)

    Die_Another_Day

  6. We spend lots of time looking at the spreadsheet error/quality story over at http://www.eusprig.org, which despite the name has an active worldwide community.
    Some research did suggest that 92% of commercial spreadsheets had material errors. Others got less exciting results. Some people even suggest that errors in spreadsheets are totally in line with the rest of human endeavour, ie we mess up 3-5% of the time. google Ray Panko and spreadsheet.
    Spreadsheets maybe the devil, but that is not the spreadsheets fault it is more likely a problem of poor quality management and unrealistic expectations. Both of which would appear whatever the technology, as Dick says. Spreadsheets also happen to be utterly mission critical for most organisations. Much to the disappointment of the BI vendors Dennis mentions.
    There are also studies that show spreadsheets as the leading edge of development, with many being working prototypes before conversion to other technology.

    Dennis the UK XL User Conference was very good, I’m thinking of going to the next one in the US in October. That one will include some Access stuff too.

    Cheers
    Simon

  7. I’ve seen this “spreadsheets are the devil” at my mid-sized public company. In my opion, this attitude is driven by the unholy alliance of big-time software vendors, company IT departments hell-bent on controlling every user transaction, and the large audit firms that are making a killing off of Sarbanes-Oxley compliance.

  8. Great post Dick….I agree totally…

    a) “Bad” people = “Bad” Governance = “Fraud” = Closed companies

    b) Spreadsheets/ERP’s/Accounting softwares have no role to play….

    c) A person/company that wants to commit fraud will commit fraud -SOX compliant or otherwise

    d) To err is human….in Spreadsheets/in Erps … anywhere

    Dennis…Looking forward for the Excel V/s SAP story…I am sure many readers will relate

  9. Just a note regarding “If SOX was in place before the Enron collapse, it would have prevented exactly nothing.”

    I am not saying SOX is a God given answer to all cases of financial reporting errors, but I want to make clear that preventing fraud is a big part of it.

    I guess the implementation of SOX is different from company to company, but basically all companies will follow guidelines set up by the big auditing firms. And preventing fraud through strict segregation of duties and approval/counter-approval is huge part of SOX.

    A lot of what Enron managed to do actually COULD have been prevented by SOX. That is because it is mandatory to be audited by an external party (as well as do internal audits on a regular basis) – and the external auditor do also audit the fraud angle *very heavily*.

    Regarding spreadsheets… Well, it is causing me a lot of pain right now, so I’ll be back and write a comment after I read the linked article a little better.

  10. Spreadsheet or paper and pencil? Yes, spreadsheets can be relied on too heavily, and people do make mistakes. But, as productivity is pushed, employees have to identify ways to meet those demands, or watch their jobs go somewhere else. The large accounting systems often fail to meet individual users needs, either because of SOX, funding, lack of functionality, or training. It is sometimes faster, easier, to put together a spreadsheet to support a manual function, or task. What we need is better training for spreadsheets, including how to make sure a spreadsheets works! In the end, any computer program large or small, does what the programmer and user tells it to.

  11. fzz –

    “back before Enron crashed companies could offer little other than their own stock in their 401(k) plans”

    Not true. I worked for a Fortune 100 company until five years ago. Our 401(k) plan included around a dozen funds, one linked to the company stock, of course, but the rest offering a range of risk/return for the participants. The 401(k) plan was managed by one of the large financial services companies that handles individual and group plans like this. The opportunity for fraud was very slim in this arrangement.

  12. Johan wrote: “That is because it is mandatory to be audited by an external party (as well as do internal audits on a regular basis)…”

    Unless, of course, the auditing/consulting firm is part of the problem as well – ala Arthur Anderson’s role in the Enron debacle.

  13. Nothing like a good ole DK rant to kickstart the morning ;). I agree with pretty much everything that Dick has to say; some of the responses, not so much. Especially, the company stock comment in 401k plans. If the fidicuiaries of plans that only offered company stock haven’t been sued, I will be amazed. I think there were plenty of studies in the 80s and 90s that showed average fund offerings were around 5 and then grew to 12 or 15 in more recent years. If a plan only offered one and that one was company stock, that is horrible decision making by the plan sponsor.

    Also, about human error and Excel. I agree it happens and it will keep happening, but mindless drones that plug information into Excel without stepping back and ‘analyzing’ the outcome of calculations, need better bosses or better training. Controller, accounting managers, accounting supervisors, whatever title you want to give them have to have the knowledge and take the time to actually review what they’ve done to ensure it passes the ‘smell test’. A little healthy skepticism is always good when you look at things.

  14. DK:

    I agree with Z–nice freakin’ rant.

    I agree with you that non-diversified portfolios should be approached with open eyes to the negative possibilities. I have little sympathy for folks who loaded up on Enron, or never diversified when it NNG and Houston Natural Gas merged to become Enron because “our family has owned the stock forever”. (And you should have little sympathy for me when LVLT tanks.)

    I also agree with you that Enron failed due to fraud. But what was fraud’s cause? Greed and really terrible business decisions. These guys wouldn’t have needed to resort to the fraud if they would have stayed within the company’s competencies. When that industry de-regulated in the mid 1990’s, many utilities started buying all kinds of crap with no idea how it worked–just to “grow the business”. So preced fraud a healthy dose of greed and ignorance, and viola, you have major implosions of Fortune 500 companies and Big 8/6/5/4 accounting firms. Then Congress–that’s right, the same body that deregulated utilities–comes up with SOX as the corporate messiah. And so the hits just keep on coming . . .

    Now you have me all worked up.

    DA

  15. Sarbanes-Oxley was initially greeted with hollow laughter here in the Square Mile: all British financial institutions must comply with the Financial Services Act, which specifically states that we have to have adequate internal management controls, and those control systems have to be documented, audited, and respected in daily use.

    It’s like we already knew what was coming from SOX, and had been doing it for the last fiftenn years… But the laughing stopped when we realised just how detailed, bureaucratic and burdensome Sarbanes-Oxley turned out to be. The costs in management time, and the administrative overhead – never mind the lawyers’ fees – are making every company that ever did business in the USA (or ever offered shares, bonds or depositary certificates to a US investor) look for ways of extricating themselves from this dysfunctional jurisdiction. Take a look at how many companies are floating in London this year, when they all used to issue their IPOs in New York.

    Worst of all, the long, long lists of rules and filings have turned out to be another gadarene rush into the swamp of ‘prescriptive’ audits – a game for clever lawyers and accountants, who satisfy the letter of the law in a lengthening list of ever-more-detailed and complex regulations, and sign off the audit with a statement of compliance with all relevant legislation.

    Here in Britain, we sign off accounts as ‘True and Fair’; ‘principles-based’ audit, and I know which type I prefer. So when I document a trading spreadsheet, and put in place the audit controls, I can’t just cover my backside by stating that it satisfies the detailed wording of a specific regulation: I have nowhere to hide when the FSA conduct their audit, nor in a court of law, unless the controls I put in place are ‘Adequate’, transparent and effective, and a full and determined attempt to comply with the spirit and principles of the law.

  16. Has anyone else noticed the ads on the bottom?

    ” Oracle Tools Extend your Oracle System. Meet Sarbox Requirements.”
    ” Sarbanes-Oxley Solutions Promote best practices for Sarbanes Oxley compliance -Free white paper.”
    ” Spreadsheet Professional Spreadsheet review software for Sarbanes Oxley. Free download. “

    SOX is tossed around in my company as a reason to make any CYA decision. Need something done twice? Call it SOX and you’re golden.

  17. A one truth-version of company data remains such an elusive goal because it mainly consists in making individuals agree on data models. Compared to this huge undertaking, discussing the shortcomings of Excel really seems futile, even if it is often not the best choice for a business application.

  18. I would draw your attention to this interesting paper by Hans Pottel of Innogenetics, that demonstrates flaws in the way Excel implements various statistical calculations.

    http://www.mis.coventry.ac.uk/~nhunt/pottel.pdf

    In the General Remarks section he writes,Quote:
    Another important remark is that although many people have voiced their concerns about the quality of Excel’s statistical computing, nothing has changed. Microsoft has never responded to comments on this issue. Consequently, the statistical flaws reported in Excel 97 worksheet functions and the Analysis Toolpak are still present in Excel 2000 and Excel XP. This, of course, is most unfortunate.
    The final section summarises the implications:Quote:
    Will these problems affect you?
    If you are using Excel for simple data analysis, on relatively easy data sets, it is most unlikely you will have any problems. The impact of the poorer algorithms used by Excel will be more visible on relatively ‘not so easy’ data sets. If you are dealing with very large numbers, scaling and/or centering your numbers will solve the problem. Note that you should not use Excel’s STDEV function to scale your data, in case of large numbers. In most cases, centering the data will suffice to generate the correct results. Some of the difficulties mentioned in this text can be overcome by using a good third-party add-in. These add-ins will usually provide the user with many more statistical tools, including non-parametric hypothesis testing, which is completely absent in Excel.
    It remains to be seen whether any of these flaws are addressed in Excel 2007.

Leave a Reply

Your email address will not be published. Required fields are marked *