The oversight role of regulators: evidence from SEC comment letters in the IPO process

We thank the SEC for providing the EDGAR search volume data for research purposes. Following Drake, Roulstone, and Thornock (2015) and Lee, Ma, and Wang (2015), we apply some basic data screening to the SEC EDGAR serve log data. For example, we remove the request where the requester has just landed on the index page of a set of documents; we discard observations for which there is no match to a particular SEC filing based on the SEC master filing index; and we remove redundant clicks in the EDGAR log file related to clicks pertaining to the same financial filing and extension from the same IP address on a given day. In addition, we drop high-frequency search requests (e.g., those generated by automated web crawlers or robots).

Please refer to http://​www.​sec.​gov/​corpfin/​Article/​filing-review-process---corp-fin.​html for more details.

Published online on 25 June 2004: http://​www.​nytimes.​com/​2004/​06/​25/​.​.​.​/​25secure.​html.

For more details, please refer to: http://​www.​pwc.​com/​en_​US/​us/​technology/​publications/​assets/​pwc-technology-ipo-sec-comments-letter-trends.​pdf.

Comments and responses are available on the Commission’s website: http://​www.​sec.​gov.

Before the SEC decided to make comment letters publicly available, its policy was to release this correspondence only in response to Freedom of Information Act (FOIA) requests after the SEC completed its review, and only to the person making such request.

For example, in a comment letter issued to Western Gas Partners LP, the SEC questioned the company’s recognized revenues, expenses, and EBITDA (earnings before interest, taxes, depreciation and amortization): “Please tell us how you treat depreciation, cost of product, interest expense, … in the presentation of estimated revenues and expenses and minimum estimated EBITDA.” The SEC comments on EBITDA and risk factor disclosures in a comment letter addressed to Quest Energy Partners, L.P.: “We note Adjusted EBITDA excludes non-cash charges related to the mark to market of derivative instruments, which are recurring items. Please revise your disclosure to clearly address why these amounts have been excluded from the measure and why the resulting measure is more useful than net income in an investor’s understanding of your business. … We note on page 25 that you were not in compliance with your debt covenants as of March 31, 2007. Please discuss the impact this non-compliance would have had on your ability to make cash distributions to your unitholders. See also our related Risk Factor comment.” Similarly, the SEC highlights a risk factor in a comment letter to Thermon Group Holdings, Inc.: “Given the numerous director resignations you expect to occur after this offering …, please provide us your analysis of the materiality of the risk that investors in this offering will be unable to assess the qualifications and experience of a significant portion of your board since it appears they will not be appointed until after this offering.” The SEC may also directly criticize the IPO issuer’s proposed offer price. In a comment letter sent to Polypore International, Inc., the SEC states: “We also note that the $20 to $22 price range is well in excess of the $5.24 exercise price of your options issued during fiscal year 2006. Please expand your disclosures (herein or in your critical accounting policies, as appropriate) to provide the information for all equity instruments granted. …” In all four cases, the IPO issuers finally adjusted downward their proposed offer price (either in an S1-A or 424B) after they addressed the comments raised by the SEC.

We download IPO underwriter reputation rankings data from Jay Ritter’s website (https://​site.​warrington.​ufl.​edu/​ritter/​ipo-data/​).

Although untabulated due to space limitations, results are robust based on unsigned price change variables (e.g., |Δprice|, |Δprice|_down, and |Δprice|_up).

The process may comprise several comment letter exchanges between the SEC and the issuer. Each exchange may involve more than one letter and response from the firm. Normally, all letters and responses in each exchange cover the same or similar topics or issues. Audit Analytics provides a Comment Letter Conversation ID so that we can easily identify comment letters that belong to the same conversation. Here we use the date of the last letter in the first conversation between the SEC and the IPO firm to divide the waiting period into a pre-period and a post-period. All our results still hold if we use the date of the first comment letter in the first conversation received by the IPO firm.

The median percentage is 74.67% for general comment letters and 73.08% for accessed comment letters, respectively. The cutoff date for the pre-period and post-period is the date of the last letter in the first conversation. When we alternatively set the cutoff date as the date of the first letter in the first conversation, the median percentage is 20.90% for general comment letters and 26.60% for accessed letters. Our main results adopt only the former cutoff method, for the sake of brevity; the results are quantitatively similar if we use the latter cutoff method.

Some examples of IPOs in the no-comment letter group include: ZBB Energy Corporation, a company designing, developing, manufacturing, and distributing energy storage systems, which submitted its first SB-2 on Oct. 27, 2006; Masimo Corporation, a global medical technology company that develops, manufactures, and markets non-invasive patient monitoring products, which filed its S-1 on April, 17, 2007; Regency Energy Partners LP, an energy partnership that gathers, processes, markets and transports natural gas and natural gas liquids, which filed its initial registration statement (S-1) on Sept. 15, 2005. There is no change in offer price from the initial filing date to final offer date for Masimo Corporation and Regency Energy Partners LP, while ZBB Energy Corporation’s offer price increased from $2 on initial filing date to $6 on final offer date.

The sample size for gpct, acctpct, and legalpct variables is smaller due to data availability.

The average number of shares offered is 14.16 million, and $3.54 million is calculated as $0.25 per share multiplied by 14.16 million shares.

The one-sided p-values when comparing the coefficients on ncomlet with those on nacesedcl are 0.112, 0.021, and 0.050 for the models of |Δprice|, |Δprice|_down, and Δprice, respectively.

All results in Table 3 still hold if we use number of comment letter conversations instead of number of comment letters as our variables of interest.

For the sake of brevity, we only report univariate comparisons defining comment letter and no-comment letter groups based on all comment letter correspondences. Tests classifying the two groups based on comment letter correspondences accessed by investors exhibit very similar results.

Following Dechow et al. (2016), we use Audit Analytics field “LIST_CL_ISSUE_PHRASE” to identify revenue recognition–related comment letters. As reported in Table 2, on average, there are 9.395 comment letters between the SEC and each IPO with comment letters. Some 2% (0.188/9.395) of those comment letters relate to revenue recognition.

IPO firms’ revenue recognition discussions are normally disclosed under accounting policy sessions, which are mainly incorporated in MD&A, footnotes, and so on. IPO firms’ risk factor disclosures usually constitute a separate section with titles such as “Risk Factor” and “Risks.”

Brown, Tian, and Tucker (2015) construct a similar proxy based on Item 1A of firm 10-Ks to investigate the spillover effect of SEC comment letters on firms’ risk factor disclosures.

The set of control variables for the long-run performance tests are based on Compustat items and IPO founding dates, kindly provided by Jay Ritter, which are different from the set of control variables (mainly based on SDC items) in the price revision models in Eqs. (1) and (2).