Variance Analysis and Horizontal Analysis (#111)
/In this podcast episode, we discuss the problems with variance analysis and how to use horizontal analysis instead. Key points made are noted below.
The Nature of Variance Analysis
Variance analysis is using a bunch of different types of variance calculations to figure out why actual results vary from a standard that you set up at some point in the past. So, for example, the labor rate variance measures the difference between the actual rate paid and the standard rate that you’re using for a baseline. And, the material price variance measures the difference between the actual price paid for materials and the standard price. Obviously, these calculations both measure changes in the rate that you pay for something.
The other type of variance is a usage variance. For example, the labor efficiency variance measures the amount of labor used in comparison to the standard number of hours that you expected to use, and the material yield variance measures the amount of materials you used in comparison to the standard amount that you expected to use.
And there’re similar variances for overhead. You get the general idea.
Now, variance analysis started in the first half of the 1900s, when there were really large, high-volume assembly lines that ran pretty much the same product, over and over again, for a long time. And in those situations, variance analysis could be useful for finding out what when wrong, since the same production system would probably be in place the next month, or the next year, for that matter, and you could tweak the system to reduce the variances.
Why We Don’t Need Variance Analysis
The trouble is, production doesn’t work that way anymore, and yet we still have accountants running around calculating the same variances. My case is, that the bulk of variance analysis should go away. Here’re some points to consider.
The first issue is the increased use of just-in-time systems, where customer orders drive production. If there’re no orders, then there’s no production. This means that your labor variances could be wildly negative in comparison to whatever your standards may be, since the standards assume a specific amount of steady production.
The second issue is that variances are only calculated after the end of each month, and by the time the accounting staff sends the variance results back to management for review, the issues are so far back in the past that they’re irrelevant.
A really good variance analysis system would give feedback to the production staff within moments of something happening, so that they can act on it right away. Well, the accounting staff is not geared to do that. Accountants are better at aggregating financial information from a database, and summarizing it for consumption at much longer intervals.
Let’s try a third point. You go ahead and calculate all of these variances. What are they actually telling you? Not much at a gross level. For example, what if the labor efficiency variance is $50,000? Can I take some action based on this information? Of course not. I have to drill down further, and figure out exactly what happened, such as needing to use overtime to make a rush delivery to a customer. Because you have to drill down a long ways to obtain any information that you can actually act on, it takes even longer to get usable information into the hands of management.
And here’s a fourth point. Variances are based on a standard. Who sets the standards, and what do the standards mean? The real poster child for this issue is the material price variance, because the purchasing staff decides what the standard will be. Since the purchasing department is being judged on how well it can buy in comparison to the standard, you can bet that they try very hard to keep the standards extremely achievable.
Bottom line, standards are set to make the standard setters look good, or at least they try all kinds of politics to make the standards come out that way. And if all you’re using variance analysis for is to cover people’s butts, why use it at all?
The Need for Horizontal Analysis
So, those are some reasons why I haven’t used variance analysis for years. But, if I’m not using variance analysis, what do I use? I can’t very well just hand out financial statements with no explanation of what went wrong or right. Making recommendations to management is part of a controller’s or a CFO’s job, so you have to analyze operations somehow.
What I prefer to use is called horizontal analysis. It’s very simple. Just prepare financial statements that show results for the last 12 months on a rolling basis, so the most recent results appear, along with each of the preceding 11 months. You can do this with almost any accounting software, or you can build it on a spreadsheet. Either way, the information is easy to put together.
Horizontal analysis just means that you’re comparing results over the months, and looking for spikes or dips that you don’t expect. When you spot one, investigate it, usually by digging through the detail for that account in the general ledger. For example, rent expense should only change when there’s a change in the lease rate. So if I see a sudden jump in the expense, then I know a subtenant didn’t pay their rent, and I can take action on that.
Horizontal analysis has a bunch of advantages over traditional variance analysis. First, it doesn’t compare anything to standards, which is hokey anyways. Second, it’s based solely on financial information that’s entirely within your accounting database. That means all of your answers are in the general ledger, and not somewhere screwy, like the bill of materials or labor routings.
And third, everyone reads it. You should be sending these rolling-twelve month financials to the entire management team, so that everyone can see the same spikes and dips that you see. If you include the reasons for those items in the notes that go with the financials, then you’ll have done your duty; you’ve told management what’s going on with their company’s financial results.
Now, if the financial statements are extremely summarized, like having a single line item for general and administrative expenses, then either create a more detailed version of the financials, or else do the horizontal analysis at the individual account level in the general ledger, rather than using the financial statements.
And, by the way, you can certainly use horizontal analysis for all of the financial statements, not just the income statement. It’s just as useful for examining cash flows and the balance sheet.
So, in short, don’t even try to use old-fashioned variance analysis. You’ll just spend a lot of time creating incomprehensible information that no one will use. Instead, use horizontal analysis. It’s simple to use, and you can investigate problems more easily. And… that’s it.