Letter on Stock Buybacks and Insiders’ Cashouts

Thank you for your December 18, 2018 letter regarding my research on the relationship between stock buybacks and corporate insiders’ stock cashouts—and for your leadership in urging the SEC to ensure that our rules protect investors when public companies buy back stock. I very much appreciate the opportunity to share further details on this work.

I first raised these concerns in a speech last June, when my Office released original research showing that corporate insiders cash out much more of their personal stock immediately after announcing a buyback than on an ordinary day. [1] If executives believe a buyback is the right thing to do, they should hold their stock over the long term. Instead, we found that many executives use buybacks to cash out. That creates the risk that insiders’ own interests-rather than the long-term needs of investors, employees, and communities-are driving buybacks.

The issue is more pressing than ever. Since January 2018, when the Tax Cuts and Jobs Act took effect, American public companies have announced a record $1 trillion in buybacks. [2] That’s all the more reason why the SEC should, as I proposed last year, hold an open comment period to revisit our rules governing buybacks-rules we haven’t examined since 2003.

In your letter, you asked me to address the possibility that my findings “could be coincidental because [a buyback] might coincide with periods when executives are permitted to sell their stocks.” The concern is that insiders, aware of a pending buyback, may be prohibited from trading until the event is public, so the selling we observe is driven by the lifting of that restriction. In response to your letter, my Office conducted additional analysis of buybacks and insider cashouts. Our findings show why this area deserves further attention:

• First, insiders sell more stock when they announce buybacks than on an ordinary day. Some firms likely restrict trading in advance of buybacks; in our sample, 38% of firms with insider sales after buyback announcements have no pre-announcement trading. However, as explained in more detail below, our findings are robust to controls for different levels of pre-announcement trading.
• Second, insider selling on buybacks is associated with worse long-term performance. It’s well known that some buybacks produce long-term stock-price increases while others lead only to a short-term price pop. [3] We show that, when executives unload significant amounts of stock upon announcing a buyback, they often benefit from short-term price pops at the expense of long-term investors. SEC rules do not address insiders’ incentives to pursue buybacks at the expense of buy-and-hold American investors.

It has been over a decade since the Commission last examined our rules governing buybacks. Since then, the growth of stock-based pal has given insiders reason to look for chances to liquidate their shares in public companies. [4] The evidence shows that buybacks give executives that chance-even when it doesn’t make long-run sense.

Our securities laws should encourage executives to pursue the kind of sustainable value that creates the stable jobs American families count on. But SEC rules governing buybacks do not distinguish between those that allow executives to cash out on short-term stock-price pops and those that reflect the company’s long-term needs. That’s why today I am renewing my call for the SEC to open a comment period to reexamine whether, and how, those rules allow corporate insiders to benefit from buybacks at the expense of ordinary investors.

I. Stock Buybacks and Corporate Cashouts: Further Evidence

An important and insightful question about my research has been raised by those who wonder whether buybacks lead to more insider cashouts because executives are often prohibited, or “blacked out,” from trading before a buyback. It might be the lifting of the prohibition on insiders’ freedom to sell, rather than the buyback, that is driving the selling we see, because there is “pent-up” insider interest in selling that can be addressed only after the buyback is announced.

To examine this possibility, my Office extracted data on all buybacks between January 2017 and the end of 2018. We then estimated the length of any pre-announcement trading prohibition by observing insider transactions in the period prior to the announcement. [5] Consistent with the possibility that such prohibitions apply during this period, 38% of the firms in our sample have no trading in the thirty days prior to the date the buyback is announced. However, and consistent with prior studies, [6] we see that a majority of firms conducting buybacks have insider transactions [7] during the eight days before the buyback is announced. [8]

Because different firms take different approaches to this issue, we empirically measure pre-announcement trading and control for those differences. Controlling for pre-announcement trading, we think, makes sense because a lack of pre-announcement trading may influence the level of post-announcement trading. However, we find that controlling for pre-announcement trading activity has little effect on the level of insider selling on the day a buyback is announced. [9] In other words: even after we account for differences in policies regarding pre-announcement trading, we still observe higher levels of insider selling on buybacks.

Because our estimates of trading restrictions are necessarily imprecise, we performed a second test. Since earnings releases usually involve this kind of restriction, we simply removed from our sample any buyback announced within twelve days of an earnings release. About 41% of the buybacks in our sample fall into this category. Even after removing these cases, we see statistically significantly higher levels of insider selling on the day a buyback is announced.

Even after accounting for important differences in firms’ approaches to insider trading before buybacks are announced, the evidence shows that, on average, executives sell far more stock when they announce a buyback than on an ordinary day. The implications of this evidence for the SEC’s work is debatable; the fact that many executives sell significant amounts of stock immediately after they announce a buyback is not.

II. Stock Buybacks and Executives’ Incentives

Another important question often raised about this research is the relationship between insider cashouts and post-buyback performance. The evidence you requested in your letter points to a troubling trend. When insiders sell upon announcing a buyback, long-term performance is worse. This raises the concern that insiders’ stock-based pay gives them incentives to pursue buybacks that maximize their pay-but do not make sense for long-term investors.

To examine this issue, we begin with data on all buybacks announced in 2017 and 2018. We then divide the level of insider selling into three groups based on the volume of insider sales and observe the abnormal [10] returns for the buybacks with the highest, lowest, and no insider sales for the ten-day period after the buyback announcement. Figure 1 describes the results:

Figure 1 shows that, when executives sell into a buyback, the buyback is more likely to produce a short-term stock-price pop rather than a long-term, sustainable value increase. The difference in performance between buybacks with executive cashouts and those without is meaningful: ninety days after the buyback announcements, firms with insider cashouts underperform the other firms we study by more than 8%. [11]

To be sure, this analysis does not show whether insiders’ sales cause lower long-run returns or whether insiders correctly anticipate that returns will be lower so sell opportunistically. But from the perspective of ordinary American investors saving for retirement, I cannot see why that distinction should matter. Whether insider sales cause the stock to fall or simply reflect insiders’ view that the buyback won’t add value in the long run, the opportunity to cash out stock-based pay gives executives reason to pursue buybacks that do not produce long-term value. Those incentives deserve attention from the SEC.

* * * *

The evidence your letter requested shows that insiders can use buybacks as a chance to cash out at high stock prices-at the expense of long-term investors. Yet SEC rules give a safe harbor to firms whose insiders sell when a buyback is announced. In a world where stock-based pay gives executives powerful incentives to seek opportunities to sell their shares, SEC rules on buybacks should do more to protect ordinary investors who save for the long run.

Although debate over these rules may seem technical or abstract, in my view, your letter reflects a fundamental principle underlying our markets. Our laws should encourage corporations to create the kind of long-term value that American families count on to build their futures. But outdated SEC rules give safe-harbor treatment to buybacks that do little more than give executives a chance to cash out. That’s why I am today renewing my call for an open comment period to revisit our rules to make sure they protect American companies, investors, and employees in light of today’s unprecedented volume of buybacks.

Thank you again for your letter—and for your work to ensure that SEC rules on buybacks protect the long-term interests of American investors and communities. Should you have any questions, or if you or your Staff would find further information helpful, please do not hesitate to contact me.

Very truly yours,

Robert J. Jackson, Jr.

Endnotes

1Commissioner Robert J. Jackson, Jr., Stock Buybacks and Corporate Cashouts (June II, 2018). Following my Office’s standard practice, 1 released my findings, along with a data appendix, on the day I reported these results. Commissioner Robert J. Jackson, Jr., Data Appendix to Stock Buybacks and Corporate Cashouts (June 11, 2018), available at https://www.sec.gov/files/speech-jackson-061118-data-appendix.pdf.(go back)

2See Bob Pisani, Stock Buybacks Hit a Record $1.1 Trillion, and the Year’s Not Over, CNBC TRADER TALK (December 2018); c;( Jesse M. Fried & Charles C.Y. Wang, Short-Termism and Capital Flows, 8 REV. CORP. FIN. STUD. 207 (2018) (contesting the degree to which figures of this kind reflect actual capital outflows relevant to long-term corporate investment).(go back)

3For a thoughtful review of the lengthy literature establishing this proposition, see Theo Vermaelen, Share Repurchases, 1 FOUNDATIONS & TRENDS IN FIN. 171 (2005).(go back)

4For insightful analysis describing this trend, see David F. Larcker & Brian Tayan, CEO Compensation Data Spotlight, STANFORD BUSINESS SCHOOL CORPORATE GOVERNANCE RESEARCH INITIATIVE 8 (2017) (describing increased CEO equity ownership in a sample of 4,000 public companies).(go back)

5The finance literature has estimated the length of such periods, see, e.g., J. Carr Bettis, Jeffrey L. Coles & Michael L. Lemmon, Corporate Policies Restricting Trading By insiders, 57 J. FIN. ECON. 191 (2000) (documenting that most earnings-related blackout periods last between two and twelve trading days). Nevertheless, because corporate policies of this kind frequently (and properly) change in response to market dynamics, we chose to empirically estimate these windows based on data on insider transactions in 2017 and 2018.(go back)

6D. Scott Lee, Wayne H. Mikkelson, & M. Megan Partch, Managers’ Trading Around Stock Repurchases, 47 J. FIN. 1947 (1992); see also Ilona Babenko, Yuri Tserlukevich & Alexander Vedrashko, The Credibility of Open Market Share Repurchase Signaling, 47 J. FIN. & Q. ANAL 1059 (2012).(go back)

7As I noted when I initially raised concerns in this area last year, it remains especially important to “be clear: this trading is not necessarily illegal.” Jackson supra note 1, at text accompanying notes 24-25. Instead, we observe insider transactions in the company’s stock solely to identify corporate policies restricting such trading.(go back)

8To the degree that transactions we observe are pursuant to prearranged trading plans, such trading may not be dispositive with respect to the existence of a blackout period-although in that case there would be less concern about “pent up” insider interest in selling. Still, to address the possibility that our data include such trades, using standard methods from the finance literature, see Lauren Cohen, Christopher Malloy, & Lukasz Pomorski, Decoding Inside Information, 67 J. FIN. 1009 (2012), we also identify and remove “routine” insider trades, such as those providing liquidity immediately after the vesting of stock-based pay. For two reasons, we follow the finance literature and identify such trades statistically rather than through disclosures. First, such disclosures are voluntary, raising the selection issues that come with voluntary disclosure, M. Todd Henderson, Voluntary Disclosures Regarding Insiders’ Rule I0b5-1 Trading Plans, HARV. L. SCH. F. ON CORP. GOV. & FIN. REG. (Aug. 26, 2008).

Second, and more importantly, I share the bipartisan concern, reflected in a bill recently introduced by your Office, that insider trading pursuant to plans under Rule 10b5-1 is associated with unusual insider profits. See Sen. Chris Van Hollen, Van Hollen, Fischer Introduce Bipartisan Bill to Increase Transparency in Corporate Trading (Feb. 27, 2019); Pete Schroeder, REUTERS POLITICS, U.S House Panel’s Top Democrat, Republican Seek Executive Trading Oversight (Jan. 18, 2019). Academic research identified this concern long ago, but neither Congress nor the SEC has addressed it yet. See Alan D. Jagolinzer, SEC Rule 10b5-1 and Insiders’ Strategic Trade, 55 MGMT. SCI. iv (Oct. 2008); M. Todd Henderson, Alan D. Jagolinzer, & Karl A. Muller, Offensive Disclosure: How Voluntary Disclosure Can Increase Returns from Insider Trading, 103 GEO. L. J. 1275 (2015).(go back)

9For example, the average buyback in our sample has total insider net selling upon announcement of approximately $3.824 million. After controlling for the degree of pre-announcement net selling, the residual average on the date of the buyback announcement is about $3.786 million.(go back)

10We calculate abnormal returns by subtracting factor portfolio returns from the individual firm’s returns following Eugene F. Fama & Kenneth R. French, The Cross-Section of Expected Stock Returns, 47 J. FIN. 427 (1992) and Mark M. Carhart, On Persistence in Mutual Fund Performance, 52 J. FIN. 57 (1997). We estimate factor portfolio exposures over a one-year period prior to the buyback announcement, with a 30-day gap between our estimation period and our event period. Contemporaneously with the release of this letter, my Office has publicly released the data we used to conduct this analysis as well as a Data Appendix describing our methodology. To address the important questions raised by those concerned about the effect of trading windows on these findings, the dataset described in Figure 1 excludes any buybacks announced within twelve days of an earnings release. In our Data Appendix, we conduct additional analysis using data going back to 2004, the last time the Commission revisited its rules in this area, and show that the results in Figure 1 are robust in that dataset.(go back)

11In the Data Appendix, we subject the finding described in Figure 1 to several further tests for robustness. For example, we extend the dataset to buybacks going back to 2004; we use coarsened exact matching, Stefano M. Iacus, Gary King & Giuseppe Porro, Causal Inference Without Balance Checking: Coarsened Exact Matching, 20 POL. ANALYSIS I (20I 2), to address potential selection issues; and we extend the post-buyback performance period from the 90 days in Figure I to over 200 days. (In light of important and insightful finance scholarship explaining the problems with multi-year factor pricing, S.P. Kothari & Jerold B. Warner, Measuring Long-Horizon Security Price Performance, 43 J. FIN. ECON. 301 (1997), we do not extend further than one trading year.) Our findings, which are consistent with longstanding literature showing that the market is sensitive to signals insiders send when trading around buybacks, see Lee et al., supra note 6; Babenko et al., supra note 6, are largely unchanged.(go back)