Article analysis for Revenue Recognition Timing and Attributes of Reported Revenue: The Case of Software Industry’s Adoption of SOP 91-1 by Yuan Zhang Timing of revenue recognition is a crucial part in revenue recognition. According to US GAAP, revenue should be recognized when it is realized/realizable and earned (FASB, 1984, Para. 83). However, a number of software firms recognized revenue prior to product delivery or service performance in the past, which potentially violated one or both of the conditions of the revenue recognition principle.
In response, AICPA released Statement of Position (SOP) 91-1 in Dec. 1991, which stipulated that if collectability is probable, license revenue should be recognized upon delivery and service revenue should be recognized ratably over the service arrangement. The research question for this article is: How revenue recognition timing affects attributes of reported revenue?
This question is interesting because: 1) revenue recognition timing is important in financial reporting and standard setters have devoted much attention, 2) very limited empirical research examining revenue recognition timing has been conducted, 3) software revenue recognition is unique as transfer of rights is achieved by license rather than on-the-spot sale of products. The main hypotheses for this article and their intuitions are: 1) Early revenue recognition increases the timeliness of reported revenue.
Its intuition is: early revenue recognition better influences decisions by providing more timely information. 2) However, it will lead to greater uncertainty in reported revenue. Its intuition is: changes may not be foreseen at the time of contract signing. 3) Time-series predictability of revenue is lower under early revenue recognition. Its intuition is: early revenue recognition results in higher estimation error and therefore reduces the time-series predictability.
The author had chosen 122 firms’ data as sample based on 3 selective criteria: the firms have to engaged in software business; their financial statements were available for each fiscal year during 1991 to 1993; the firms directly reported the time, nature, and/or effect of adopting the SOP. The sample period had chosen from 1987 to 1997. The reason for choosing it is the analysis requires information from the statement of cash flows, which became available in 1987, and stop at 1997 to ensure the post-SOP firm-quarters had consistent definition of revenues. She classified the 122 firms into two groups: 29 EARLY firms and 93 CONTROL firms.
The EARLY group was assigned by the criteria that the firms reported negative cumulative adjustment, which shows they had previously applied early revenue recognition and had to change their revenue recognition policies upon adopting the SOP. The CONTROL group disclosed no material impact from adopting the SOP. In this sample, no firms reported positive cumulative effects. The author also provided descriptive statistics of three sets of constructs for the sample: size, growth, and performance in a reason that these constructs were expected to affect the attributes of reported revenue, thus mitigated bias.
In summary, the author examined EARLY firms and CONTROL firms for most of the variables and concluded for most of the variables, EARLY firms were not significantly different from CONTROL firms. In this section, the author described analysis of data and gave the process of regression analysis. The goal of accrual accounting is to account in the periods in which they occur for the effects on an entity of transactions and other events and circumstances. This statement indicated standard setters ‘emphasis on both relevance and reliability of accounting information.
However, there is a trade-off between relevance and reliability. Relevance may suffer when an accounting method is changed to gain reliability. In the software industry, since early revenue recognition has higher capacity to influence decisions by providing more timely information, it is expected to increase timeliness, and hence relevance, of reported revenue. Therefore, the author made a Hypothesis: Ceteris paribus, reported revenue under early revenue recognition is more timely in providing economic information than that under SOP 91-1.
An efficient stock market impounds value-relevant information of economic transactions in a quick and unbiased fashion. The author tried to find the relationship between reported revenue and stock return to prove the hypothesis by employing reverse regression methodology. After analysis of data, the author’s conclusion is that early revenue recognition increases the timeliness. In section 5, the author hypothesizes that the accounts receivable accruals map into the cash flow realizations to a lesser extent than those under SOP 91-1.
The author points out that the earlier the recognition, the more likely the customer is to not pay. Circumstances such as acceptance, commitment to pay and/or needs for customization may change prior to or upon delivery of the software or rendering of the service that may not be foreseen at the time of contract signing or earlier than as specified in SOP 91-1. In order to test his hypothesis, the author followed the Dechow and Dichev (2002), which examines the standard deviations of residuals from a regression of accounts receivable accruals on corresponding cash flow realizations.
The model uses information related to accounts receivable, sales cash collected during the periods, and errors in accounts receivable accruals (sresid). The author also broke down the firms by pre- and post-SOP and size. The author found that sresid is higher for EARLY firms than for CONTROL firms prior to SOP 91-1 and then reversed in post-SOP. Further, early revenue recognition decreases the extent to which accounts receivable accruals map into cash flow realizations and the reliability of revenue-related accounting information.
Section 6 of the paper examined the time series predictability of report revenue; that is how well reported revenue in financial statements predicts the ability of the firm to generate future revenue. The author concluded in section 5 that early revenue recognition resulted in higher estimation error in accounts receivables accruals. The author also expected that higher estimation error reduces the time-series predictability. Based on the expectation, the author focused in the absolute forecast error based on the Foster (1977) model to study the predictability of reported revenue of pre- and post-SOP period.
The author believed that size and revenue volatility of a firm might affect the result of the study; therefore, she placed control to those two variables in order to obtain an unbiased result. The result showed that the absolute forecast errors during pre-SOP period were generally higher for firms that recognize revenue early. Based on that, the author accepted her hypothesis that early revenue recognition reduced the predictability of reported revenue.
