Financial accounting is an information system. This was true already many years (maybe centuries) before we came up with the term information system and started using it in connection with computer technology.
Information because it processes data and gives its users useful information—for example, about whether their “business” sufficiently assesses invested funds or whether their company is unduly over-indebted despite having a fairly high bank account balance.
System because it is composed of simple foundational elements (accounts) and the organized links between them, which can form a relatively complex and complicated construction. This seeming complicatedness discourages many people, so they simply don’t like accounting. But what we don’t understand is what is complicated. What we understand is simple. And the same applies to understanding accounting.
Additionally, I would argue that financial accounting is a basic business intelligence (BI) system. It describes the entire organization and integrates data from various areas—assets, stocks, funds, claims and obligations, resources and reserves, costs, revenues and more. Accounting undoubtedly provides the familiar “one version of the truth,” even the version that we present to shareholders and the authorities, and we pay taxes and dividends based on it. Another important feature of BI is data historization and a consistent view of past periods, where accounting yet again meets all the criteria. Our company may have changed completely over the years—we may have changed half our computer systems, changed strategies twice or even shifted to a new field of business—but the accounting outputs will still have the same structure and explanatory power. If I want to know the structure of long-term claims overdue for the past five years, I will get completely reliable information thanks to accounting, even though today‘s companies have three times as much turnover, have undergone a digital transformation and instead of selling foreign products would rather provide their own services.
Of course, compared to other BI systems, accounting also has its limitations. That is why the field of business intelligence with its predecessor MIS (management information systems) was created. The basic limitations are related to dimensionality, granularity and the variety of figures available. Accounting figures are always financial, that is they express values denominated in a specific currency (in Czech crowns, euros, dollars, etc.), but they don’t allow you to work with the number of products sold, with kilograms of raw materials consumed, kilometres travelled, etc. Everything is converted to money. Even qualitative indicators such as risk levels, otherwise determined by a percentage, are converted to money in the form of financial reserves.
Dimensionality generally indicates which criteria can be used to further subdivide the figures being observed. There are not any problems with the fundamental dimension (time) in accounting. Accounting easily works even with different parallel timelines (the date of the accounting differs from the accounting period, the accounting period does not always exactly correspond to the date of taxable supply, etc.). However, when it comes to other dimensions, we do not have many options—the fundamental dimension corresponds to the account defined in the chart of accounts (which is also the basic “codebook” of each accounting system), accounts can be used for different generic or specific classifications of the figures being observed. Furthermore, it typically uses subdivisions into projects (e.g., when comparing the direct costs and revenues of project XY with its contributions to covering the company’s overheads) and cost centres (e.g., how much the marketing centre costs us). Dimensions in accounting also include various subdivisions into tax-deductible and non-tax-deductible items, but with this, the dimensionality of our strictly accounting BI practically ends.
Granularity determines the maximum level of detail that can be observed in a given system. The granularity of our universe is said to correspond to Planck’s length of 10^(-33) meters. My fitness tracker records all the data it measures every five seconds. Banking BI/DWH (a BI system supported by a data warehouse) granularity displays the details of individual banking clients, each of their accounts and each individual financial transaction (even payments made by credit card that were denied). Accounting usually works with a lot of aggregated data. Financial figures add up to big units—for example, I can know the total amount of claims, but I do not know who the specific debtors are; I can see sales totals for last month, but I do not see that most of them took place in the last week, etc.
Of course, it is possible to work with and eliminate all these limitations. The easiest way is dimensionality. By suitably constructing the chart of accounts and especially its hierarchization, it is possible to broaden the figures observed and their level of detail. It is possible to supply additional sorting features beyond traditional projects and cost centres. This is also easy in case of granularity limitations. It is not necessary to aggregate data. We instead work with the details at the level of individual accounting entries. Limitations at the level of figures may seem necessary, but in fact, it is possible to go even deeper. Accountants are familiar with the concept of operational accounting records, which allow them to work with non-financial “natural” units, of which the more curious include units such as the weight gain of a herd of livestock.
Thus, in many ways accounting can serve as basic BI, but it would not be good to make accounting into something it is not and fail to give real BI/DWH systems the moment in the sun that they deserve. Instead, let’s focus on the strongest part of accounting.
This has already been mentioned in the introduction of “systematicity”. Analogies are misleading, but allow me to once again take a step in the direction of physics. A physical system is described by the state of its elements and changes occur between these states. An accounting system is described by the state (balance) of its accounts (e.g., in one account, the balance expresses the sum of money available in Czech crowns; in another account, the balance expresses the sum of all payroll costs since the beginning of the calendar year). Furthermore, the law of energy conservation applies in physics. And believe me that accounting has something similar. The total sum of assets is always equal to the liabilities. With a loan (liability), I can increase the amount of money in my bank account (asset). These free assets can be transformed into other (material) assets, and the resulting profit is either distributed to shareholders (paying account dividends decreases the volume of assets and liabilities) or stays in the company (no change). Nobody knows exactly how the law of energy conservation is implemented in physics, but in accounting, A=L is ensured by double entries to the sides known as debit (the asset side) and credit (the liability side). Any change in the system is recorded as a change on both sides in the same amount (or a gain in one asset and a loss of the same amount for other assets, and likewise for liabilities). This gives us basic system control—if the sum of the assets is not equal to the liabilities, we know immediately that something is wrong.
Perhaps even more than equality of assets and liabilities, accounting excels in the principle of transactions. All changes in the system are done only by means of entries. An entry makes a change to the balance of one account and the opposite change in exactly the same amount to the balance of another account. Nothing but an entry can change the state of your account. The accounting-entry level of detail thus determines the smallest possible detail that we can observe in the accounting system (i.e. it determines the aforementioned granularity).
I dare say that some modern technologies have borrowed this principle from accounting, such as relational databases (an analogy to debit/credit in the world of databases is INSERT/UPDATE/DELETE, and none of the other data in the tables changes—I’m not counting DROP and TRUNCATE, which corresponds to something like “shred accounting books”). And block-chain technology is directly based on accounting practices.
Finally, back to business intelligence. Again, I really don’t want to say that financial accounting can compete with robust BI/DWH systems. I am just saying that accounting is something like an “elementary” BI system that also represents the official version of the truth for many “business objects” (such as the sums of assets and liabilities, profit and loss, claims and obligations) and thus represents a “standard” for other company BI systems. For every accurate BI/DWH system, we perform reconciliations to the general ledger, that is to say, compare whether the sum of very granular data in total corresponds to the accounting, and if not, then whether we understand all the differences and whether we can justify these differences (e.g., accounting uses a different exchange rate than a DWH). However, if the reconciliation differences are too large and unjustifiable, it is clear that the DWH has unambiguously bad data (despite all the records having passed all DQ checks). Accounting reconciliation gives us basic assurance regarding the completeness and accuracy of the data in the DWH.
To finish on a lighter note, let me make one last reference to physics (which I otherwise barely understand) by paraphrasing one of its giants. If I know anything about information systems and business intelligence, it’s only because I am standing on the shoulders of a T account 🙂
Author: Petr Hájek
Information Management Advisor
—
Note: A careful reader with some accounting knowledge will certainly have noticed that the above text contains a lot of simplifications. Firstly, accounting here is explained purely as financial accounting (and thus we ignore the concepts of, for example, management accounting). Also, it is not explicitly stated that we are talking about double-entry accounting (as opposed to single-entry accounting, where the debit/credit principle does not apply). Furthermore, I have not mentioned, for example, legal or regulatory constraints, which do not allow sufficiently flexible or creative ways of adapting accounting to the needs of business intelligence.