Aggregate insider trading and market returns: The role of transparency

• DOI: http://doi.org/10.1111/jbfa.12357
• Published date: 01 March 2019
• Author: Francois Brochet
• Date: 01 March 2019

DOI: 10.1111/jbfa.12357

Aggregate insider trading and market returns:

The role of transparency

Francois Brochet

BostonUniversity

Correspondence

FrancoisBrochet, Associate Professor, Boston

University,595 Commonwealth Avenue,Boston,

MA,USA.

Email:fbrochet@bu.edu

Abstract

Using a sample of countries that require timely disclosures of insider

trades, I investigate the effect of country-levelinstitutions that pro-

mote transparency on the extent to which aggregate insider trades

predict marketreturns. I find that financial information transparency

mitigates the predictive content of aggregate insider trades when

markets are more likely to deviate from fundamentals (i.e., during

marketfads), and when there is greater co-movement in stock prices.

Incontrast, there is some evidence that governance and investor pro-

tection mitigate the association between aggregate insider trades

and future earnings surprises. Hence, holding constant the timely

disclosures of insider trades, other capital market institutions play

complementary roles in mitigating the informational frictions that

give rise to the predictive content of aggregate insider trades.

KEYWORDS

fundamentals, insider trading, investor protection, investor

sentiment, market fads, synchronicity,transparency

1INTRODUCTION

This study examines the predictive content of aggregate trading by corporate insiders in a large international sample

of equity markets. Prior research has found that insiders’ net purchasing activity aggregated at the marketlevel is pos-

itively associated with subsequent market returns using US data from the 1980s and 1990s (Jiang & Zaman, 2010;

Lakonishok & Lee, 2001; Seyhun,1988, 1992), and more recent data from China (Zhu, Wang, & Yang, 2014). This phe-

nomenon arises when corporate insiders trade on information signals that are not strictly idiosyncratic to their firm.

Yet,this empirical regularity may seem puzzling in today's capital markets, given the timely dissemination of both infor-

mation that moves market returns and insider trade disclosures around the globe. Indeed, over the last decade, an

increasing number of countries and stock exchanges have been mandating the disclosure and prompt dissemination

of corporate insider transactions. However,the US and China differ greatly from each other in terms of present day

capital market institutions, and the US has undergone significant changes since the 1990s. Hence, the first question I

seek to address is whether country-levelinstitutions affect the extent to which corporate insider trades predict market

returns, and if so, which institutions?

US corporate insiders have also been shown to trade in aggregate on superior knowledge of future fundamentals

and stock mispricing (Jiang & Zaman, 2010; Seyhun, 1992). The distinction between those two sources of information

336 c

2018 John Wiley & Sons Ltd wileyonlinelibrary.com/journal/jbfa JBus Fin Acc. 2019;46:336–369.

BROCHET 337

advantage is far from trivial, due to their different implications for market fairness and efficiency. Institutions that

promote the timely incorporation of information in stock prices must balance the incentive alignment and signal-

ing benefits of insider trading with insiders’ rent-seeking behavior. My second question is whether market-level

institutions affect the extent to which aggregate insider trades reflect fundamentals or perceived mispricing driven

by market sentiment or other frictions (i.e., ‘fads’).1Among the countries that now require corporate insiders to

disclose their equity transactions, the disclosure requirements for insiders’ equity transactions are homogeneous

(e.g., similar timeliness and content).2In and of itself, this convergence suggests a move towards ‘best practice’ in

capital market transparency (Martynova & Renneboog, 2011), i.e., the availabilityof information about publicly listed

companies to outside investors (Bushman, Piotroski, & Smith, 2004). But those countries otherwise differ along

regulatory and informational institutions that are likely to affect the averageinvestor's ability to infer insiders’ private

information in a timely manner. I exploit country-level variations in those institutions to develophypotheses on the

cross-country variation in the predictive content of aggregate insider trades for future fundamentals and marketfads.

Prior research documents the effect of firm- and country-level transparency on the informativeness of firm-level

insider trades, with most of the evidence based on US data. In brief, insider trades are less informative when firm-

level financial transparencyis higher (Frankel and Li, 2004; Huddart and Ke, 2007) or when firms have stronger gover-

nance (Dai, Fu,Kang, & Lee, 2016). Country-level evidence is sparser. Fidrmuc, Korczak, and Korczak (2013) document

a positive (negative) association between investorprotection and the informativeness of insider purchases (sales) in a

cross-country panel, while Gebka et al. (2017) find that insider trading is more profitable in European countries with

stronger investorprotection and enforced regulatory rules. Notwithstanding the limited evidence on firm-level trades,

I expect a cross-country analysis of aggregate insider trades to offer a novel perspective.The information content of

firm-level insider trades is a function of country-,firm-, insider- and transaction-specific factors (e.g., Hillier, Korczak,

& Korczak, 2015). In contrast, while it is arguably a coarse signal, aggregate insider tradinghas the advantage of teas-

ing out idiosyncratic differences in agency frictions, liquidity and portfolio rebalancing needs, etc., and thus should be

affected more directly bycountry-level informational transparency. Furthermore, Gebka et al. (2017) show that insider

tradesgenerate no significant alpha in most European countries. This raises the question of whether corporate insiders

trade on correlated instead of purely idiosyncratic signals in those economies.

It stands to reason that market-levelfinancial information transparency could, theoretically, mitigate the predictive

content of insider trades both for future cash flows and during market fads. With more timely and reliable access to

financial information, market participants can directly infer insiders’ information from those alternative sources or

react more promptly to insider trade disclosures. Many transparency constructs are designed to reveal non-public

information both at the firm- and aggregate level. For example, sell-side analysts’ recommendations inform industry

and market returns (Howe, Unlu, & Yan, 2009; Kadan, Madureira,Wang, & Zach, 2012). Likewise, high-quality finan-

cial reporting standards and auditors facilitate comparability across firms (Francis,Pinnuck, & Watanabe, 2013; Yip &

Young, 2012). Ex ante, these transparency mechanisms should preempt the predictive content of aggregate insider

trades, whether insiders’ trades reflect future fundamentals or if market prices deviate from fundamentals due to

forces beyond the scope of corporateinsiders’ responsibility (e.g., because of investor sentiment).

However, the null hypothesis may not be rejected for several reasons. For one, because prior studies find that

aggregate insider trades have predictive content for both returns and fundamentals in the US (Jiang & Zaman, 2010;

Seyhun, 1992) – arguably a high-transparency market.Second, other market characteristics such as the degree of gov-

ernment intervention in the economy mayhave a first-order effect on the extent to which insiders trade on correlated

1Seyhun (1992) uses the term ‘fads’ in reference to movements in market returns unexplainedby observable economic fundamentals (such as aggregate

earnings,GPD, etc.) Given the difficulty of controlling for all ‘fundamentals’ that may explain variation in market returns, especially in a cross-country setting,

Iuse the term ‘fad’ more specifically to refer to country-quarters where investor sentiment (as defined in Baker and Wurgler, 2006) appears to be abnormally

highor low. Empirically, I identify fads using proxies from prior literature (Baker,Wurgler, & Yuan,2012).

2I consider the disclosure requirements to be homogenous for this study.That is, while differences remain across countries, they should have no bearing on

myinferences, given my research design. For example, the median transaction reporting lag varies from 0 days in the Netherlands to 5 daysin France, butsince

I aggregate trades by quarter,those differences are inconsequential. Likewise, there is some variation across countries in what defines a reporting insider,

but – consistent with prior research – I focus on top executivesand directors, all of whom are subject to trading requirements in my sample. I also control for

aggregatechanges in reported institutional ownership in all regressions.

338 BROCHET

information. Third, to the extent that insiders trade much more often in countries with greater transparency (such

as the US or the UK), aggregate insider trades may be a low-power signal in low-transparency countries. Hence, the

effect of transparency on the predictive content of aggregatei nsidertrades is an empirical question.

A broader view of transparency includes corporate governance and investorprotection. Because governance pro-

visions are put in place more specifically to prevent insiders from taking advantage of their private information about

near-term cash flows, they are more likely to mitigate the predictivecontent of insider trades for future fundamentals

than for fads. However,even if designed at the country-level, governance transparencytargets firm-level agency costs,

consistent with Firdmuc et al. (2013) and Gebka et al. (2017). Hence its effect on aggregate insider trades is likelyto be

weaker than that of financial information transparency,if not absent.

Using a dataset of equity transactions compiled in large part by a data vendor (Director Deals) and supplemented

by data obtained from individual stock exchanges,I first document that, on average, proxies for net insider purchases

aggregated by country and calendar quarter,are positively associated with the next country-quarter stock return. On

average,a one-standard deviation increase in aggregate net insider purchases is associated with 1% higher excess mar-

ketreturns in the next quarter. Measured separately,purchases and sales have positive and negative predictive content

for next quarter's return, respectively.

Next, I examine the effect of transparency.To measure financial information transparency,I use four country-level

variables: the adoption of high quality accounting standards, analyst coverage, the proportion of firms audited by Big

Four auditors, and earnings quality (Bhattacharya, Daouk, & Welker,2003; Leuz, Nanda, & Wysocki, 2003). Because

those variables are highly correlated with each other,and because they are unlikely to affect insider trades in isolation,

I create a financial transparency score by grouping the four variables. I use the same method to create a governance

transparency score, based on anti-self-dealing (Djankov,La Porta, Lopez-de-Salines, & Shleifer, 2008), the degree of

(perceived) insider trading restriction (Du & Wei, 2004), the extent of timing restrictions on corporateinsider trades

around earnings announcements, and whether the country allows for class-action lawsuits (Leuz, 2010).

The main findings indicate that financial information transparency mitigates the predictive content of aggregate

insider trades when investorsentiment is abnormally high or low (i.e., during fads) and when stock prices are more syn-

chronous. In contrast, higher investorprotection appears to mitigate the association between aggregate insider trades

and future earnings surprises, albeit weakly.Collectively, the results suggest that aggregate insider trades’ predictive

content is primarily driven by fads in countries with lower financial information transparency.

Robustness tests using instrument variables, additional controls for country-level institutions, and the passage of

the Sarbanes-Oxley Act (SOX) in the US continue to indicate that country-level transparencymitigates the predictive

content of aggregate insider trades.These results alleviate concerns of endogeneity or omitted capital market frictions

driving the observed results.3

This paper makes several contributions to the literature. I extendthe work of Chowdhury, Howe, and Lin (1993),

Jiang and Zaman (2010) and Zhu et al. (2014), Lakonishokand Lee (2001), Seyhun (1988, 1992) by showing that capital-

marketinstitutions have a significant effect on the predictive ability of aggregate insider trades for market returns. This

paper offers significant belief revision by showing that the association between aggregate insider trades and market

returns in the US is much weaker after the passage of SOX. I attribute this difference to the improved transparency

brought about by SOX, including the significant change in disclosure requirements for insider trades.

This is the first study to document the effect of country-level institutions on the degree to which insider trades

reflect information about future fundamentals versus market fads. By showing how country-level governance and

financial information transparency affect each channel, the paper offers significant insights beyond single-country

studies. Insofar as regulators require the timely disclosure of insider transactions to promote market integrity and

3Iexamine whether the passage of the Sarbanes-Oxley Act (SOX) – a significant shock to transparency – had an effect on the predictive content of US aggre-

gate insider trades. Using monthly data, I find a significant reduction in the predictive content of US aggregate insider trades around SOX.Results based on

a difference-in-difference design with the UK as a control sample confirm that the decrease is not due to contemporaneous and unobservable changes in

global equity markets.Hence, my results are consistent with prior findings based on pre-SOX data but shed further light on the effect of cross-sectional and

time-seriesvariation in transparency on the predictive content of aggregate insider trades.