It is human nature to want to look good to others. Add to that the financial incentives that managers of private companies have to make their companies look good, and you have a recipe that sometimes results in manipulation of financial data and sometimes outright “cooking the books.” Research by Wan Wongsunwai, an assistant professor of accounting information and management at the Kellogg School of Management, shows that oversight by the best venture capitalists (VCs) can substantially improve financial reporting in the crucial period surrounding the initial public offering (IPO) of equity stock. But the presence of a VC alone is not enough to reduce the risk of dishonest financial reporting—the VC must have the skills needed to exercise influence over managerial behavior.
There are widely divergent views of VCs, Wongsunwai notes. Previous academic papers have tended to portray them as very useful intermediaries who channel funds from those who have them to those who need them. However, some businesspeople refer to VCs as “vulture capitalists,” believing they seize opportunities and exploit weaknesses they detect in people they are dealing with. Wongsunwai’s study attempted to reconcile these opposing views of VCs.
Unlike other studies examining the relationship between VCs and company financial reporting, Wongsunwai’s work distinguishes between VCs who are able to meaningfully influence management behavior (“high-quality VCs”) and those who lack that ability. However, trying to measure the quality of VCs is not straightforward, Wongsunwai says. For example, previous research attempted to use financial performance as a measure of VC quality, but that approach was complicated by the difficulty researchers faced in obtaining accurate information. He says that is because VCs are often required to disclose this information only to their investors, not to outsiders. To overcome this problem, he distilled a variety of quality measures into two: past investment history and syndication activity.
Trying to measure the quality of VCs is not straightforward, Wongsunwai says.
Peering into the Past
Past investment history reflects the success of past deals—for example, whether the VC invested in Google. Syndication activity looks at how many times a given VC attracts others as investors in a project and suggests how he or she is perceived by peers. “The idea is, VCs will know who the better ones are amongst themselves,” Wongsunwai explains. “If we see a successful VC going to syndicate often with another VC, chances are they think that other VC is a quality VC,” he says. On the other hand, if a VC tends to do things on his or her own, it is often because other VCs avoid joint ventures because of his or her poor reputation.
The advantage these measures have over financial performance is that the data are readily available to researchers. “We can see which companies VCs invest in because they don’t mind telling us,” Wongsunwai says. “We can also see who they have been working with because if one party doesn’t tell you, another party is willing.”
Wongsunwai focused on how VCs impact two kinds of earnings management: real earnings manipulation and accrual-based manipulation.
Real earnings manipulation describes activities firms undertake to influence the numbers that they are required to report. For example, a company nearing the end of a quarter may realize that it is going to fall short of what the market expects. Perhaps investors forecast that the company would deliver $100 of profit in the coming quarter, with sales of $500 and expenses of $400. Prior to the end of the quarter, the company realizes that although its sales will be $500, its expenses will be $420, leaving it with profit of only $80. In this situation, the company might look for expenses it could cut or postpone. It might decide that a marketing campaign costing $20 and scheduled to start two weeks before the end of the quarter should be postponed two weeks so that it is not part of the fourth-quarter report. Thus, the earnings figure will rise to the expected $100.
According to Wongsunwai, “The problem with this type of behavior is that it has implications for long-term future performance. If I cut down on my marketing today, I am likely to make fewer sales tomorrow.” He went on to explain that companies manipulate figures in this way, despite the negative impact on long-term performance, because the need to avoid disappointing investors is seen to outweigh the future loss of profits that may derive from those actions.
The other type of earnings management, accrual-based manipulation, “is really cooking the books,” Wongsunwai says. “In our example, you go ahead and do the marketing activities as planned, but you misreport it. Instead of saying that you spent $20 on marketing, you use some accounting tricks to make it look like it is only $5. There are many ways companies can do this kind of thing.”
Wongsunwai used Thomson Financial’s Venture Economics database to identify VC-backed companies that conducted IPOs between 1990 and 2004, together with the names of the VCs who provided first-round financing. For each VC-backed company that went public, he identified from Securities Data Corporation a matching non-VC-backed company that conducted an IPO in the United States in the same year, was in the same industry, and had the closest return on assets in the year of the IPO. This yielded 1,226 IPO companies. He then collected financial statement data on these companies from Compustat, as well as data compiled by Audit Analytics on financial restatements announced between January 2001 and December 2008.
Wongsunwai examined earning management activities over four periods of time relative to the IPOs. Phase 1 was the pre-IPO phase, consisting of four quarters ending on the fiscal quarter-end date immediately preceding the IPO date. Phase 2 was the period starting with the fiscal quarter-end immediately preceding the IPO date and ending on the fiscal quarter-end immediately preceding the lockup expiration date. Phases 3 and 4 were two subsequent periods of four quarters each.
Quality That Shows
Wongsunwai found that companies backed by higher-quality VCs engaged in less aggressive financial reporting, as reflected in lower abnormal accruals and lower real activities manipulation (Figures 1 and 2), and a lower rate of subsequent financial restatements. The difference in abnormal accruals was driven by the period immediately preceding the expiration of IPO lockups. (When attracting VCs, many entrepreneurs commit to a “lockup” period, a length of time they will hold their own shares of the stock after the IPO, to assure investors that the company is a legitimate operation.)
Figure 1. Average abnormal accruals for three groups of IPO companies in four time periods.
The phases are defined as follows: (1) the pre-IPO phase consisting of four quarters ending on the fiscal quarter-end date immediately preceding the IPO date, (2) the period starting with the fiscal quarter-end immediately preceding the IPO date and ending on the fiscal quarter-end immediately preceding the lockup expiration date, (3) and (4) two subsequent periods of four quarters each.
Figure 2. Average aggregate measures of real earnings management for three groups of IPO companies in four time periods
Moreover, higher-quality VCs drove the lower abnormal accruals in VC-backed companies. Lower-quality VC-backed and non-VC-backed IPO companies had statistically indistinguishable levels of abnormal accruals in all periods surrounding the IPO and lockup expiration. “These findings suggest that lower-tier VCs are unable or unwilling to monitor their companies so as to constrain the aggressive reporting of financial results to potential new investors,” Wongsunwai reports in a working paper. “Overall, the evidence points to higher-quality VCs monitoring their portfolio companies more effectively.”
The existing literature generally asserts that real activities manipulation and accrual-based manipulation tend to be used interchangeably (i.e., companies try one approach and, if it doesn’t work, try the other). However, Wongsunwai’s research suggests that when high-quality VCs are involved in a company, not only is one type not being used as a substitute for the other, but neither type is observed. “I ascribe that to the fact that top-tiered VCs with strong skills are able to constrain the human-nature tendency to overstate performance,” he says.
“This research is about why people do the things they do,” Wongsunwai explains. In the case of business managers, for example, the reason they manipulate earnings may be stock options. “If people running a company have their wealth tied to the price of the company stock, they will resort to a lot of ways to keep the stock prices as high as they can,” he says. In the case of VCs, this research looks at a point in time at which VC monitoring incentives might be weakened. If the companies are trying to inflate prices, “the VCs might be tempted to let their monitoring role lapse a little bit,” Wongsunwai points out, because if as a result stock prices go up, the VC benefits as well. However, the highest-quality VCs have a counter-incentive: their reputation. Wongsunwai concludes, “If VCs don’t have the skills needed to influence managerial decisions, or if they choose not to apply their skills at the time their companies are going public, then we will see negative consequences.”
Related reading on Kellogg Insight