Financial disclosure by SMEs listed on a semi-regulated market: evidence from the Euronext Free Market

It is interesting to note that the decline in the number of small company IPOs in the US since 2000 has been blamed on overregulation by the Sarbanes-Oxley act and the US Securities and Exchange Commission’s Regulation Fair Disclosure, which mandates that all publicly traded companies must disclose material information to all investors at the same time (e.g., Gao et al. 2012). The American JOBS act, which passed Congress on 27 March 2012, aims to ease the regulatory burden on companies seeking to raise capital, e.g., by allowing firms to remain nominally private so that they still face minimal disclosure standards and still raise up to 50 million USD annually in shares that are transferable (The Economist 2012).

The term limited liability company refers to société anonyme (SA) in France. These firms have limited liability and their shares may be offered for sale to the public. This is the most common legal form on the Free Market.

While firms listed on the regulated market must follow IFRS since 2005, those on the Free Market do not have this obligation. Nevertheless, we do not find any firm in our sample that voluntarily adopts IFRS. Firms on the Free Market are also not required to disclose a corporate governance and internal control report, or adapt their governance structures to the national Corporate Governance Code.

Indicators for equity markets other than the Free Market are based on statistics from the World Federation of Exchanges, available on its website (http://​www.​world-exchanges.​org/​). The statistics for the Free Market include the French and Belgian Free Market. The Belgian Free Market is small (see Sect. 6.4.1).

When we only consider Euronext Paris, we observe a similar proportion in terms of market capitalization.

Outside ownership might not only comprise ownership by retail investors, but also by institutions. Prior studies have found a positive association between institutional ownership and voluntary disclosure. Healy et al. (1999) and Bushee and Noe (2000) show that firms with higher disclosure rankings have greater institutional ownership. Ajinkya et al. (2005) find that institutional ownership increases disclosure. However, institutional investors have a low preference to invest in small stocks (e.g., O’Brien and Bhushan 1990; Hessel and Norman 1992). Therefore, we consider it unlikely that institutional ownership is an important determinant of disclosure in our SME setting.

Other important intermediaries that can reduce information asymmetry are institutional investors. Prior research shows that information is more quickly impounded in market prices in the presence of institutional investors. El-Gazzar (1998), for example, finds that the price reaction to earnings announcements is smaller for stocks with greater institutional ownership, consistent with institutional investors having incentive to search for private predisclosure information and triggering managers to voluntarily release predisclosure information. However, as institutional investors have a low preference to invest in small stocks (e.g., O’Brien and Bhushan 1990; Hessel and Norman 1992), we consider it unlikely that this intermediary has an important role in an SME setting.

If the financial year is different from the calendar year, we consider the financial year that overlaps most with the calendar year. If we refer to a certain year without further specification, this is the financial year. If we refer to the calendar year, this will be explicitly stated.

An additional reason for focusing on 2008 is that a large number of firms (31 firms or 19 % of our final sample for 2008) delisted from the Free Market on 1 July 2010 following the introduction of an annual fee of 2,500 euros. As financial information was in most cases removed from the corporate website as well as from the stock exchange website, analyses based on financial disclosure in 2009 would need to be done without these firms.

For 17 firms we cannot retrieve financial disclosure information since they were delisted in 2009 (13 firms went bankrupt, 2 active firms became dormant firms, 1 firm was dissolved because of a merger or take-over, 1 firm delisted after a voluntary repurchase offer). Relevant accounting data are not available for 13 firms in the Amadeus database. We cannot calculate the standard deviation of return on assets for two firms. This means that 84 % (164 firms) of our total population (196 firms) is included in our sample. To reduce the concern that the missing data bias our results, we replace missing accounting data with available data from another prior year. In addition, we also leave the standard deviation of return on assets (which is insignificant) out of our regression model. This increases our sample from 164 firms to 172 firms. Results are qualitatively very similar and are available from the authors upon request.

In 2008, only 9 of our 164 sample firms (5.49 %) disclosed information about performance in the first quarter, and only 6 firms (3.66 %) disclosed information about performance in the third quarter. In 2007, only 11 of our 143 sample firms (7.69 %) disclosed information about performance in the first quarter, and only 8 firms (5.59 %) disclosed information about performance in the third quarter.

Information is considered concrete when it relates to a situation or event that exists, has happened, or that one can reasonably expect will happen in the future and is specific enough to have a potential influence on the share price. Information is considered significant when one can expect that a rational investor probably will use this information to make an investment decision.

This index is constructed by analogy with the Euronext Brussels “Free Market grid” used by the stock exchange to judge the transparency of firms listed on the Belgian Free Market (see Sect. 6.4.1). This instrument was created in 2009 because of concerns about the reputation of the Free Market due to the low transparency of many firms.

Tests of unidimensionality check whether the items of a summated scale are strongly related such that they load highly on a single factor. As expected, the following items load highly on one factor: {Financial Calendar, Shareholder Meeting, Specific Contact Details}, which represents information that allows shareholders to exercise their rights; {Managers, Board of Directors, Shareholders}, which represents information about corporate governance; {Financial Statement, Audit Report, Management Report, Annual Report}, which represents documents about the previous financial period; and {First Quarter, Second Quarter}, which represents quarterly information.

We do not add a measure for systematic risk, as our estimates of Beta are noisy because of high illiquidity of many stocks. For the same reason, we also do not include stock returns and the standard deviation of stock returns as proxies for performance and the unpredictability of performance.

The following sectors (subsectors) are considered as being technology-based: software & computer services (computer services, Internet, software) and technology hardware & equipment (computer hardware, electronic office equipment, semiconductors, telecommunications equipment).

Aggregation of the NACE Rev. 2 code into a smaller set of larger groups is based on the following classifications: Intermediate SNA/ISIC Aggregation A*38 and High-Level SNA/ISIC Aggregation A*10/11.

Investors and firms on the regulated market of Euronext have a notification duty when the percentage ownership rises above or falls below certain thresholds. However, for the Free Market, no such declarations are required. Therefore, it is no surprise that detailed information on the ownership structure of firms on the Free Market is scarce, and Free Float is not available for all observations. When we add the Free Float dummies, differences in significance for some variables are caused by the reduced sample size and not because of the added variables.

The test is based on a sample that also includes firms listed on the Belgian Free Market (cf. Sect. 6.4.1). This allows us to consider both equity issuance and listing on the Belgian Free Market as instruments. At least two instruments are necessary to test whether instruments are exogenous. The F-statistics of the weak identification tests indicate that our instruments are relevant. The Hansen (1982) J-statistics indicate that our instruments are exogenous. The unreported results for the two-stage least squares regressions (available from the authors upon request) confirm the results of the OLS-regressions anyway. We also controlled for endogeneity using our normal sample with only equity issuance as an instrument. Again, this yields very similar results.

We also exclude those firms for which information on the free float is not available, but fulfill the other conditions to be listed on the regulated market with regard to market capitalization and track record.

A list of liquidity provider agreements is kindly provided by Frédéric de Laminne (NYSE Euronext Brussels).

The score on the investor relations index can also be considered as count data. We use the negative binomial regression as the count data model because the overdispersion alpha is significantly different from zero.

We use a sample that also includes firms listed on the Belgian Free Market (cf. Sect. 6.4.1), which allows us to consider both equity issuance and listing on the Belgian Free Market as instruments. At least two instruments are necessary to test whether instruments are exogenous. The F-statistics of the weak identification tests indicate that our instruments are relevant. The Hansen (1982) J-statistics indicate that our instruments are exogenous. Based on the Hausman (1978) endogeneity tests, we accept the null hypothesis of exogeneity. The unreported results for the two-stage least squares regressions (available from the authors upon request) confirm the results of the OLS regressions anyway. We also controlled for endogeneity using our normal sample with only equity issuance as an instrument. This gives very similar results.

The market-to-book ratio and equity offerings might be related to each other. Prior research shows that the market-to-book ratio might be a determinant of equity offerings (e.g., Pagano et al. 1998; Hovakimian et al. 2001) and that that equity offerings might have valuation effects (e.g., Mikkelson and Partch 1986; Masulis and Korwar 1986). Moreover, valuation might also be affected by institutional differences. In our case, firms on the Belgian Free Market have the listing requirement in Belgium to carry out a public offering and therefore publish a prospectus.