Earnings management behavior of the IPO firms during pre-ipo, IPO, and post-ipo years: Evidence from the Casablanca Stock Exchange

Transcription

1 Earnings management behavior of the IPO firms during pre-ipo, IPO, and post-ipo years: Evidence from the Casablanca Stock Exchange Omar Farooq * Department of Business and Management, Aalborg University, Aalborg, Denmark Meryem Benali School of Business Administration, Al Akhawayn University, P.O. Box 104, Avenue Hassan II, Ifrane 53000, Morocco. * Correspondence Address: Department of Business and Management, Aalborg University, Fibigerstræde 4, Aalborg 9220, Denmark. Telephone: address: omar.farooq.awan@gmail.com

2 Earnings management behavior of the IPO firms during pre-ipo, IPO, and post-ipo years: Evidence from the Casablanca Stock Exchange Abstract This paper examines the pattern of earnings management around the IPO year in Morocco during the period between 2001 and Our results show that earnings management reaches its maximum level during the year of an IPO followed earnings management during the pre-ipo years and the post-ipo years. We argue that the need for making reputation as a profitable firm induces management to managing earnings during the pre-ipo years. This need to manage earnings is the most during the year of an IPO because announcing earnings that are lesser than the previous years would hurt the reputation of a firm in the eyes of potential investors/shareholder and would reduce the offering proceeds. However, as firm enters the post-ipo years, lower requirements of raising capital in immediate future along with greater scrutiny from stock market participants result in lower incentives to manage earnings. JEL classification: G34; G14 Keywords: Earnings Management; Initial Public Offering, Corporate Governance; Information Disclosure; Emerging Markets.

3 1. Introduction Do firms disclose their information properly when they are going for an IPO? Are incentives to manipulate information different during different stages of firms life? Answers to these questions are very important and have formed the basis for a vast amount of literature. Extensive amount of prior literature documents that varying degrees of earnings manipulation during the three stages of firms life, namely the pre-ipo, the IPO, and the post-ipo years. Aharony et al. (2000), for example, document significance presence of earnings management during the pre-ipo years. While, Teoh et al. (1998) document that the U.S. firms significantly manage their earnings during the IPO year. Similar behaviour of earnings management is documented during the post-ipo years (Kothari, 2001; Dechow and Schrand, 2004). An important result that emerges from this stream of literature is that the extent of earnings manipulation during different stages of firms life depends on the extent of incentives faced by management for manipulating earnings. This paper aims to document earnings management behavior of firms that went for an IPO in Morocco during the period between 2001 and Morocco is an interesting case to study earnings management behavior for a number of reasons. The Casablanca Stock Exchange, the main stock exchange of the country, is one of the fastest growing stock exchanges in the Middle East and North Africa (MENA) region. 1 Many observers argue that one of the major causes for the growth is the increased IPO activity in the Casablanca Stock Exchange. 2 Introduction of Maroc Telecom at the end of 2004 is considered by many as the single most important triggering event that brought significant number of investors to the market. 3 Increased investor interest in the market should result in more vigilant monitoring by regulatory authorities. However, this was not the case in Morocco. Belkahia (2005), for example, documents that Moroccan firms do not disclose 1 MASI and MADEX, major indices of the Casablanca Stock Exchange, recorded 71.14% and 77.66% respectively during MASI is the all share index of the Casablanca Stock Exchange, while MADEX is the index representing the most liquid shares at the Casablanca Stock Exchange. 2 In addition to IPOs, capital markets reforms initiated by Moroccan government have also contributed in increasing investors confidence in the Casablanca Stock Exchange. 3 Maroc Telecom s IPO was oversubscribed 18 times, with overseas institutional investors getting a 30% slice of the offering. This IPO allowed Morocco to show a strong potential in attracting investment.

4 information properly. 4 In another related study, Farooq and El Ouaabani (2008) argue that lack of independence contribute to ineffectiveness of the CDVM, the capital market regulator in Morocco. In the absence of proper monitoring and governance mechanisms, it is tempting for the firms, especially those that plan to go public, to manipulate their accounting statements. Criado-Jiménez et al. (2008), for example, document that the IPO firms attempt to hide unfavorable information. Overstating their earnings may allow them to raise excessive proceeds from the IPOs. In this paper, we aim to investigate whether the firms going for an IPO engage in earnings management or not in Morocco. Any evidence of earnings management during the year of an IPO should be taken as an indication that the firms do not disclose their information properly. As a result, investors may be buying the stocks that are already over valued. Using the data on discretionary accruals and absolute discretionary accruals, proxies for earnings management, we show that earnings management reaches its highest level during the IPO year, followed by slightly lower earnings management during the pre-ipo years. Our results are consistent with previous literature that considers management s desire to build firm s reputation as a profitable firm as one of the main reasons behind gradual increase in earnings manipulation (DuCharme et al., 2001). Better reputation helps firms to increase offering proceeds. Larger offering proceeds enhance the amount of fund available for investing and increase the capital resources of firms. Our results also show considerably lower earnings management during the post-ipo years. We argue that there is a less need on the part of management to engage in earnings manipulation during the post-ipo years relative to the pre-ipo and the IPO years (Aharony et al., 2000; Mikkelson et al., 1997). In addition, greater scrutiny by stock market agents also lowers the means available to management to engage in earnings manipulation during the post-ipo years. Our results have implications for investors and regulatory authorities operating in Morocco. We show that that firms going for an IPO do not disclose proper information to the investors, thus inducing investors to make a trade 4 He mentions that there is no information for investors about voting rights and that key executives do not disclose any information regarding their interest in any trade or other matters affecting the firms. He also documents the weakness of corporate governance in Morocco by mentioning that appropriate level of minority shareholders protection is missing in the country. One of the reasons for the ineffectiveness of firm- and country-level governance mechanisms is the immense power of individuals and families involved in business. According to one estimate the family groups contribute around 25% to 30% to the GDP. It also indicates that the top 12 families hold 30 % of the Moroccan economy.

5 on faulty information. We recommend that regulatory authorities should be more vigilant in monitoring the firms and investors should scrutinize the information provided by firms more rigorously. The remainder of the paper will proceed as follows: Section 2 briefly discusses motivation and background for this study. Section 3 discusses the data used in this study. Section 4 presents assessment of earnings management behavior of the IPO firms during the pre-ipo, IPO, and the post-ipo years. The paper concludes with Section Motivation and background Prior literature characterizes emerging markets with weak corporate governance mechanisms both at the firm-level and at the country-level. Claessens and Fan (2002), for example, document laxness of firm-level governance mechanisms by identifying the presence of pyramids and cross-shareholdings in firms headquartered in the emerging markets. On a similar note, Pistor and Xu (2003) highlight ineptness of country-level governance mechanisms by documenting failure of enforcement mechanisms in the emerging markets. Prior literature suggests that presence of family control, weak enforcement of investor protection laws, lax implementation of anti-director rights, and ineffectiveness of regulatory authorities have contributed to relative poor governance mechanisms in emerging markets. This paper argues that ineffective governance mechanisms have resulted in lower levels of information disclosure in the emerging markets. Leuz et al. (2003), for example, analyze disclosure quality of publicly traded firms in emerging markets and show that managers and insiders do not disclose true information about their firms. This paper argues that ineffective governance mechanisms not only effect disclosure quality of publicly traded firms but also induce private firms to evade effective disclosure of their value. This paper hypothesize that firms considering having an IPO also disclose faulty information, usually overly optimistic information, regarding their value. The reason for this is to induce potential investors to buy their stocks in the IPO. The premise that firms manipulate accounting information to influence decision of potential shareholders is central to a large body of literature in accounting and finance. Teoh et al. (1998) and

6 Roosenboom et al. (2003), for example, document significant earnings management by IPO firms. This paper acknowledges the presence of earnings management practices by firms considering having an IPO, but argues that presence of different incentives during the life cycle of a firm allows for different levels of earnings manipulation. For example, we argue that during the pre-ipo years, firms have the incentive to build a reputation as a profitable entity. In order to do so, managers are tempted to overstate accounting information. Aharony et al. (2000) document that accounts receivables of IPOs firms are abnormally high during the pre-ipo years. We argue that overstatement of accounting information is done with a hope to attract potential investors and establish a base for the sales of shares during the IPO. DuCharme et al. (2001) document that the higher pre-ipo discretionary accruals increase firm s initial value and its IPO proceeds. In addition, it is intuitive to argue that it is relatively hard to estimate fundamental value of private firms. Given that there are no shares available for trade, the interest of stock market participants in finding new information is minimal in these firms. As a result, most investors have to rely on the information disclosed in the financial statements to estimate fundamental value of private firms. We argue that, managers of those firms that are considering having an IPO are aware of this fact. Therefore, they are tempted to overstate their accounting performance. An important implication of the above stream of literature is that firms do engage in earnings manipulation during the pre-ipo years, but this earnings manipulation is not as much during the pre-ipo years as it is during the IPO year. The reason for this is simple: Lesser earnings than the previous years would hurt the reputation of a firm in the eyes of potential investors/shareholder. As a result, the incentive to overstate accounting information/earnings is the most during the year of an IPO. Extensive amount of literature documents firms manage their earnings upwards at the time of IPOs. Teoh et al. (1998), for example, document that the U.S. firms significantly manage their earnings during the year of an IPO. They show that firms generally use accruals to improve their financial statements during the IPO year. Similar results are reported by Roosenboom et al. (2003) for the Dutch firms. This strand of literature suggests that the desire of management to increase offering proceeds may be at the heart of such an opportunistic

7 behaviour. Increase in offering proceeds not only enhances the amount of fund available for investing but also increase the capital resources of firms. In addition, this strand of literature also argues that managers are under a considerable pressure to reach the forecasts made in the period of introduction in order to avoid the legal lawsuits by discontented investors (Teoh et al., 1998; Roosenboom et al., 2003). Therefore, managers may be induced to manage earnings upward during the IPO year. H1: Firms manage their earnings more during the IPO year in comparison to the pre- IPO years. This paper argues that there is a less need on the part of management to engage in earnings manipulation during the post-ipo years relative to the pre-ipo and the IPO years. Our arguments are consistent with Aharony et al. (2000) who document that account receivables are lower during the post-ipo years for the Chinese B-share firms in comparison to account receivables during the pre-ipo years. Ibbotson (1975) and Mikkelson et al. (1997) argue that the post-ipo earnings declines are bound to happen because of increased earnings management behavior during the pre-ipo and the IPO years. This strand of literature argues that if firms seek to boost earnings before selling stock, initial financial statements of the newly public firm will contain unusually high positive accruals In this scenario high earnings cannot be sustained, so earnings in the post-ipo years decline (Kothari, 2001; Dechow and Schrand, 2004; Ball and Shivakumar, 2008). Therefore, we may expect firms to have lower possibility of managing earnings in the years immediately after the IPO year. In addition to above arguments, we believe that lesser need of raising capital in immediate future after the IPO may also cause lower earnings management during the post-ipo years. This lesser need of immediate funds may induce managers to put less effort on manipulating earnings. Furthermore, one can argue that greater scrutiny by stock market agents, such as financial analysts and investors, also results in better disclosure and lower earnings manipulation during the post-ipo years.

8 H2: Firms manage their earnings the least during the post-ipo years in comparison to the pre-ipo years and the IOP year. 3. Data For the purpose of examining earnings management behavior of IPO firms, we identify all non-financial firms traded at the Casablanca Stock Exchange during the period between 2001 and Our sample consists of 41 non-financial firms. We classify the firms as those that went for IPO during the sample period and the firms that went for IPO before that. This classification will also allow us to compare earnings management behavior of IPO firms with more established firms, i.e. firms which were listed before IPO data Our sample contains 22 non-financial firms that went for IPO at the Casablanca Stock Exchange during the period between 2001 and Table 1 documents the frequency of IPO activity during each year. Our results show high IPOs activity during the last two years of our analysis period, i.e. 10 IPOs in 2006 and 8 IPOs in Our results show no IPO during 2002 and [Insert Table 1 here] 3.2 Choice of earnings management variable Prior studies use total accruals to detect earnings management. Healy (1985), for example, uses total accruals as a measure of earnings management, while De Angelo (1986) uses total accruals of the previous period as a proxy for the next period s earnings 5 The Casablanca Stock Exchange is the third oldest stock exchange in Africa. It was established in 1929 and currently has 14 members and 71 listed securities. The exchange is relatively modern, with an electronic trading system. There are also plans to implement a central scrip depository. It is also covered in a quarterly report on Arab stock exchanges. There are no restrictions on foreign investment on the Casablanca Stock Exchange, nor on foreign ownership of companies.

9 management. Both Healy (1985) and De Angelo (1986) assume that changes in nondiscretionary accruals are equal to zero between periods. Empirical tests prove that this assumption is far from reality (Kaplan, 1985). Further studies, therefore, developed models that distinguished between discretionary and non-discretionary component of accruals. Jones (1991) uses an estimate of the discretionary component of total accruals as a measure of earnings management. One of the limitations of Jones model is the assumption that earnings are non-discretionary. The modified Jones model was built to overcome this limitation. This paper uses the modified Jones model to come up with a proxy for earnings management. See Appendix-A for the detailed methodology to calculate earnings management variable. The accounting data used in the modified Jones model is retrieved from firms yearly financial statements. Firms yearly financial statements are available on the Casablanca Stock Exchange s and the CDVM s website. 6 The accounting variables used are: current assets, current liabilities, short term debt, cash, revenue, receivables, total assets, property plant and equipments, and depreciation. Table 2 documents the descriptive statistics for each of these variables during the pre-ipo, IPO, and the post-ipo years for our sample firms. The results show that the descriptive statistics for the variables used to compute earnings management are approximately the same magnitude during the pre-ipo and the IPO year. However, they decrease sharply during the post-ipo years due to, probably, increase in total assets of the firm. [Insert Table 2 here] Table 3, Panel A and Panel B, document the descriptive statistics of discretionary accruals and absolute discretionary accruals our proxies for earnings management for our sample firms during the pre-ipo, IPO, and the post-ipo years. Our results show that mean and median earnings management are highest during the year of an IPO followed by pre-ipo years and the post-ipo years. For example, our results in Table 3, Panel A, show that mean discretionary accruals in the year of an IPO are followed by the 6 The CDVM is abbreviation of Le Conseil Déantologique Des Valeurs Moblières. It is in charge of monitoring the Moroccan securities market.

10 pre-ipo mean discretionary accruals of and the post-ipo mean discretionary accruals of Similar results are observed for median discretionary accruals. Our results indicate that increased earnings management during the year of an IPO may be a result of firms desire to increase their offering proceeds. Consistent with our expectations, our results also suggest that firms manage their earnings more during the pre-ipo years relative to the post-ipo years, probably to develop their reputation as profitable firm during the pre-ipo years. The results for median discretionary accruals are qualitatively the same as the results for mean discretionary accruals. In addition, we also show that firms which went for an IPO prior to 2001 have the least earnings management. For example, our results in Table 3, Panel B, show mean absolute discretionary accruals of for firms which went for an IPO prior to These results indicate that firms which have been listed for a long period are less prone to manage earnings. It may be because older firms are more visible and thus attract more investors and analyst attention. More visibility diminishes the chances of any unscrupulous behavior by the management. [Insert Table 3 here] 4. Methodology The following section documents earning management behavior of IPO firms during the pre-ipo, IPO, and the post-ipo years. For the purpose of this paper, we will use two proxies for earnings management: (1) Discretionary accruals and (2) Absolute discretionary accruals. It is important to mention here that every firm has its own motivation to either manage its earnings upward or downward. For instance, firms seeking external funds tend to manage their earnings upward, while firms seeking tax reduction or government protection manage their earnings downward (Chao et al., 2004). Opposing directions of earnings management may result in canceling the effect of each other and therefore may show zero average earnings management in the sample (Krishnan, 2003). To overcome this limitation, we use absolute values of discretionary accruals along with discretionary accruals as a proxy for earnings management.

11 4.1 Univariate analysis In this section, we document the difference between mean and median earnings management for our sample firms during the pre-ipo, IPO, and the post-ipo years. Table 4 uses t-test to test the null hypothesis of no difference in mean earnings management during the pre-ipo, IPO, and the post-ipo years. Evidence from Table 4, Panel A, shows a significant difference in mean discretionary accruals between the IPO and the post-ipo years. For example, we show that discretionary accruals in the year of an IPO are basis points more than discretionary accruals in the post-ipo years. Similar results are reported for absolute discretionary accruals. For instance, our results in Table 4, Panel B, document a significant difference of basis points between absolute discretionary accruals in the year of an IPO and absolute discretionary accruals in the post-ipo years. Table 4, Panel B, also shows a significant difference of basis points between absolute discretionary accruals in the pre-ipo years and absolute discretionary accruals in the post-ipo years. These results are consistent with our hypothesis of highest earnings management in the year of an IPO followed by earnings management in the pre-ipo and the post-ipo years. As was expected, the results in Table 4 also show the least earnings management both discretionary accruals and absolute discretionary accruals for firms that went for an IPO prior to We argue that older firms have more visibility and thus are less prone to earnings manipulation by management. [Insert Table 4 here] Table 5 uses Mann-Whitney test to test the null hypothesis of no difference in median earnings management during the pre-ipo, IPO, and the post-ipo years. Evidence from Table 5 confirm our previous findings that earnings management during the year of an IPO and the pre-ipo year earnings management is larger than earnings management in the post-ipo years. For example, our results in Table 5, Panel A, show that discretionary accruals in the year of an IPO are basis points more than discretionary accruals in the post-ipo years, and discretionary accruals in the pre-ipo years are basis points more than discretionary accruals in the post-ipo years. Our results for absolute

12 discretionary accruals are qualitatively the same. For example, our results in Table 5, Panel B, show a significant difference of basis points between absolute discretionary accruals in the year of an IPO and absolute discretionary accruals in the post-ipo years. These results are consistent with our arguments regarding the highest level of earnings management in the year of an IPO followed by earnings management in the pre-ipo and the post-ipo years. In addition, as was the case before, we also show the least earnings management for firms that went for an IPO prior to [Insert Table 5 here] 4.2 Multivariate analysis We hypothesized in Section (2) that firms manage their earnings the most in the IPO year followed by the pre-ipo and the post-ipo years. In order to test this hypothesis, we estimate a regression with earnings management (EM) as a dependent variable and three dummy variables representing earnings management in the pre-ipo (PREDUM), IPO (IPODUM), and the post-ipo (POSTDUM) years as independent variable. It is important to mention here that there will be no perfect multocollinearity among dummy variables due to the presence of firms that went for IPO prior to We also include industry dummies (IDUM) and year dummies (YDUM) in our regression equation. For the purpose of completeness, we will report our results with and without industry and year dummies. Our basic regression equation takes the following form: EM α Yr Yr β PREDUM β 2 IPODUM 3 POSTDUM Ind YDUM IDUM ε 1 (1) Ind There may be concerns that results obtained from above regression analysis may be driven by some firm-specific characteristics. In order to address these concerns, we modified equation (1) by adding number of variables representing different aspects of firm-specific characteristics. For example, firm s size can be considered as a proxy for visibility of a firm. Smaller firms have less visibility to analysts, investors, and regulating authorities and thus have more opportunitoes to manage their earnings. Kim et al. (2003) show that smaller firms manage their earnings more than large firms. Therefore, we add

13 log of market value (SIZE) to capture the effect of higher visibility on earnings management. In addition, we also controlled for the effect that growth opportunities may have on earnings management by adding growth in firm s total assets (GROWTH). Prior literature suggests that managers of firms with growth opportunities tend to have more discretion in terms of accounting choices (Smith and Watts, 1992). Therefore, they may be tempted to manage earnings (Skinner and Sloan, 2002). Consistent with prior literature, we also controlled for revenues and leverage of a firm (Magrath and Weld, 2002; Sercu et al., 2006). Sales revenues (REVENUE) and total debt (LEVERAGE) were added in our regression equation for this purpose. Our regression takes the following form after addition of control variables: EM Yr 4 α β1predum β2 IPODUM 3POSTDUM SIZE 5 GROWTH 6 REVENUE β7 LEVERAGE Yr Ind YDUM IDUM ε Ind The results of above equations are presented in Table 6 and Table 7. Table 6 documents regression coefficients for discretionary accruals and Table 7 documents similar results for absolute discretionary accruals. Our results in Table 6, Panel A, show that earnings management was the highest in the IPO year followed by the earnings management in the pre-ipo and the post-ipo years. For example, our result from Equation (2), the most comprehensive equation, shows that earnings management in the IPO year is basis points more than earnings management in other years. Our results from the estimation of above regression equations also show that the coefficient of IPODUM is the highest followed by the coefficient of PREDUM and POSTDUM. It indicates that extent of earnings management was the highest in the IPO year followed by the earnings management in the pre-ipo and the post-ipo years. Our results in Table 6, Panel B, compliment our previous findings by showing a significant difference between the coefficient of IPODUM and POST DUM and the coefficients of PREDUM AND POSTDUM. [Insert Table 6 here] (2)

14 The results reported in Table 7 are qualitatively the same as results reported in Table 6. Our results confirm our previous findings of the highest in the IPO year followed by the earnings management in the pre-ipo and the post-ipo years. [Insert Table 7 here] 6. Conclusion This paper examines the earnings management behavior of Moroccan firms during the pre-ipo, IPO, and the post-ipo years during the period between 2000 and Our results show that Moroccan firms increase their earnings management considerably during the year of an IPO. We argue that higher earnings management during the IPO year is driven by the desire of management to induce investors to invest in their stocks. Our results also show that earnings management is higher in the pre-ipo years relative to the post-ipo years. It also shows the desire of management to gradually signal investors that they are profitable. These results are robust even after controlling for several firm-specific characteristics. Reference Aharony, J., Lee, J., and Wong, T. J., (2000). Financial Packaging of IPO Firms in China. Journal of Accounting Research, 38 (1), pp Ball, R. and Shivakumar, L., (2008). Earnings Quality at Initial Public Offerings. Journal of Accounting and Economics. Belkahia, R. (2005). The State of Corporate Governance in Morocco. Corporate Governance Trends, Center for International Private Enterprise. Chao, C-L, Kelsey, R. L., Horng, S-M, and Chiu, C-Y., (2004). Evidence of Earnings Management from the Measurement of the Deferred Tax Allowance Account. The Engineering Economist, 49 (1), pp

15 Claessens, S. and Fan, J. P. H., (2002). Corporate Governance in Asia: A Survey. International Review of Finance, 3 (2), pp Criado-Jiménez, I., Fernandez-Chulian, M., Husillos-Carques, F., and Larrinaga- Gonzalez, C., (2008). Compliance with Mandatory Environmental Reporting in Financial Statements: The Case of Spain ( ). Journal of Business Ethics, 79, pp De Angelo, L., (1986). Accounting Numbers as Market Valuation Substitutes: A Study of Management Buyouts of Public Stockholders. The Accounting Review, pp Dechow, P. M. and Schrand, C. M., (2004). Earnings Quality. Research Foundation of CFA Institute, Charlottesville, VA. DuCharme, L. L., Malatesta, P. H., and Stephan E. S., (2001). Earnings Management: IPO Valuation and Subsequent Performance. Journal of Accounting, Auditing &Finance, 16, pp Farooq, O. and and El Ouaabani, M., (2008). Relationship between regulatory authority s characteristics and information disclosure: Evidence from the Casablanca Stock Exchange. Working Paper, Al Akhawayn University in Ifrane. Healy, P., (1985). The Effect of Bonus Schemes on Accounting Decisions. Journal of Accounting and Economics, 7, pp Ibbotson, R. G., (1975). Price Performance of Common Stock New Issues. Journal of Financial Economics, 2, pp Jones, J. J., (1991). Earnings Management during Import Relief Investigations. Journal of Accounting Research, 29, pp

16 Kaplan, R., (1985). Evidence of the Effect on Bonus Schemes on Accounting Procedure and Accrual Decisions. Journal of Accounting and Economics, 7, pp Kim, Y., Liu, C., and Rhee, S. G., (2003). The Relationship of Earnings Management to Firm Size. Working Paper, University of Hawaii. Kothari, S. P., (2001). Capital Markets Research in Accounting. Journal of Accounting and Economics, pp Krishnan, G.V., (2003). Does Big-6 Auditor Industry Expertise Constrain Earnings Management? Accounting Horizons, 17, pp Leuz, C., Nanda, D., and Wysocki, P. D., (2003). Earnings Management and Investor Protection: An International Comparison. Journal of Financial Economics, 69. Mikkelson, W. H., Partch, M. M., and Shah, K., (1997). Ownership and Operating Performance of Companies that go Public. Journal of Financial Economics, 44, pp Pistor, K. and Xu, C., (2003). Incomplete Law. Journal of International Law and Politics. Roosenboom P., Goot, T., and Mertens, G., (2003). Earnings Management and Initial Public Offerings: Evidence from the Netherlands. The International Journal of Accounting, 38, pp Teoh, S. H., Welch, I., and Wong, T. J., (1998). Earnings Management and the Long-run Market Performance of Initial Public Offerings. Journal of Finance, 53 (6), pp

17 Appendix-A: Construction of earnings management variable We use modified Jones model to compute proxy for earnings management in this paper. The modified Jones calculates total accruals as follows: CCAt CCLt CCasht CSTDt Dept TA t (A) A t1 Where TA t is total accruals, CCA t is change in current assets at t, CCL t is change in current liabilities, CCash t is change in cash, CSTD t is change in short-term debt, Dep t is depreciation, and A t-1 is one period lagged total assets. According to modified Jones model, the total accruals are comprised of two components: Discretionary and non-discretionary component of total accruals. The nondiscretionary component or the inherent part of total accruals is not influenced by any managerial decision. It represents the accruals that are affected by the changing economic conditions of the firm. The non-discretionary component of accruals is estimated as follows: 1 CREV CREC PPE NDAt 1 2 t t 3 A t1 Where NDA t is non-discretionary accrual at t, CREV t is change in revenues at t, CREC t is change in net receivables at t, and PPE t is property, plant and equipment (fixed assets at t). The values of β 1, β 2, and β 3 are estimated by regressing total accruals against the inverse of one period lagged total asset, PPE, and CREV. 7 TA t 1 1 A t1 2 CREVt 3PPEt t Discretionary accruals are accruals resulting from direct manipulation of estimates by managers. The discretionary component of accruals (DA) is obtained by subtracting Equation (B) from Equation (A). DA TA NDA (C) t t t t (B) (B ) 7 The inverse of one period lagged asset explains the non-discretionary accruals related to firm s size. Change in revenues and account receivables are used as explanatory variables because they control for any earning management technique using revenues. While, property, plant and equipment are used to explain the portion of non-discretionary accruals related to depreciation expense.

18 Discretionary accrual and absolute discretionary accrual are used as a measure of earnings management in this paper.

19 Table 1: Descriptive statistics for IPOs This table documents frequency of IPO activity in the Casablanca Stock Exchange during our sample period. Our sample period comprise of the period between 2001 and Year No. of IPOs

20 Table 2: Descriptive statistics of variables used to estimate discretionary accruals This table documents descriptive statistics (mean) of the variables in the modified Jones model for firms that went for IPOs in the Casablanca Stock Exchange during the period between 2001 and The table also reports similar statistics for firms that went for IPO before Definition of each variable is provided below. CA t = Current assets at year (t) divided by total assets of pervious year (t-1) CL t = Current liabilities at year (t) divided by total assets of pervious year (t-1) Cash t = Cash flows at year (t) divided by total assets of pervious year (t-1) STD t = Short term debt at year (t) divided by total assets of pervious year (t-1) Dep t = Depreciation at year (t) divided by total assets of pervious year (t-1) Rev t = Total Revenues at year (t) divided by total assets of pervious year (t-1) Rec t = Short term receivables at year (t) divided by total assets of pervious year (t-1) PPE t = Property, plant and equipment at year (t) divided by total assets of pervious year (t-1) Pre-IPO IPO Post-IPO Post-IPO (Pre-2001) CA t CL t Cash t STD t Dep t Rev t Rec t PPE t

21 Table 3: Descriptive statistics for earnings management The table documents descriptive statistics for earnings management in our sample. We document descriptive statistics for earnings management during the pre-ipo, IPO, and the post-ipo years for firms that had IPOs during the period between 2001 and The sample consists of non-financial firms listed at the Casablanca Stock Exchange. The table also reports similar statistics for firms that went for IPO before We measure earnings management by discretionary accruals and absolute discretionary accruals. Panel A documents statistics for discretionary accruals, while Panel B documents similar statistics for absolute discretionary accruals. Panel A: Discretionary accruals Mean Standard Deviation Median No. of observations Pre-IPO IPO Post-IPO Post-IPO (Pre-2001) Panel B: Absolute discretionary accruals Mean Standard Deviation Median No. of observations Pre-IPO IPO Post-IPO Post-IPO (Pre-2001)

22 Table 4: Univariate analyses (parametric tests) for differences in mean earnings management: t-test The table documents difference in mean earnings management for our sample. We document difference in mean earnings management between the pre-ipo, IPO, and the post-ipo years. The sample consists of nonfinancial firms that went for an IPO at the Casablanca Stock Exchange during the period between 2001 and We measure earnings management by discretionary accruals and absolute discretionary accruals. Panel A documents statistics for discretionary accruals, while Panel B documents similar statistics for absolute discretionary accruals. We use t-test to test the hypothesis of no difference in earnings management between different classifications of our sample. The p-values are given in the parenthesis. Coefficients that are significant at 10% are shown in bold. Panel A: Difference in mean discretionary accruals: t-test Pre-IPO IPO Post-IPO Post-IPO (Pre-2001) Pre-IPO (1.000) IPO (0.397) (1.000) Post-IPO (0.216) (0.018) (1.000) Post-IPO (Pre-2001) () () (0.010) (1.000) Panel B: Difference in mean absolute discretionary accruals: t-test Pre-IPO IPO Post-IPO Post-IPO (Pre-2001) Pre-IPO (1.000) IPO (0.867) (1.000) Post-IPO (0.010) (0.008) (1.000) Post-IPO (Pre-2001) () () (0.164) (1.000)

23 Table 5: Univariate analyses (non-parametric tests) for earnings management: Mann-Whitney test The table documents difference in median of earnings management for our sample. We document difference in median earnings management between the pre-ipo, IPO, and the post-ipo years. The sample consists of non-financial firms that went for an IPO at the Casablanca Stock Exchange during the period between 2001 and We measure earnings management by discretionary accruals and absolute discretionary accruals. Panel A documents statistics for discretionary accruals, while Panel B documents similar statistics for absolute discretionary accruals. We use Mann-Whitney test to test the hypothesis of no difference in earnings management between different classifications of our sample. The p-values are given in the parenthesis. Coefficients that are significant at 10% are shown in bold. Panel A: Difference in median discretionary accruals: Mann-Whitney test Pre-IPO IPO Post-IPO Post-IPO (Pre-2001) Pre-IPO (1.000) IPO (0.625) (1.000) Post-IPO (0.051) (0.038) (1.000) Post-IPO (Pre-2001) () () (0.007) (1.000) Panel B: Difference in median absolute discretionary accruals: Mann-Whitney test Pre-IPO IPO Post-IPO Post-IPO (Pre-2001) Pre-IPO (1.000) IPO (0.568) (1.000) Post-IPO (0.012) (0.004) (1.000) Post-IPO (Pre-2001) () () (0.289) (1.000)

24 Table 6: Regression analysis for earnings management (discretionary accruals) and the pre-ipo, IPO, and the post-ipo years The table documents OLS regression coefficients for equation (1) and equation (2). The sample consists of non-financial firms that went for an IPO at the Casablanca Stock Exchange during the period between 2001 and We measure earnings management by discretionary accruals. Panel A documents OLS regression coefficients. Variables significant at 10% are followed by *, variable significant at 5% by **, and variable significant at 1% by ***. Panel B documents difference between regression coefficients. The p-values are given in the parenthesis. Coefficients that are significant at 10% are shown in bold. Panel A: Regression coefficients Discretionary accruals Equation (1) Equation (2) Pre-IPO 0.281*** 0.278*** 0.216*** 0.207*** IPO 0.409*** 0.332*** 0.352* 0.234* Post-IPO Size Debt Sales *** *** Growth ** ** Industry Dummies No Yes No Yes Year Dummies No Yes No Yes No. of observations R² Adjusted-R² F-value Significance of F-value Panel B: Difference between coefficients Discretionary accruals Equation (1) Equation (2) IPO Vs. Pre-IPO (0.863) (0.697) (0.515) (0.835) IPO Vs. Post-IPO (0.010) (0.056) (0.051) (0.095) Pre-IPO Vs. Post-IPO (0.001) (0.065) (0.038) (0.021)

25 Table 7: Regression analysis for earnings management (absolute discretionary accruals) and the pre- IPO, IPO, and the post-ipo years The table documents OLS regression coefficients for equation (1) and equation (2). The sample consists of non-financial firms that went for an IPO at the Casablanca Stock Exchange during the period between 2001 and We measure earnings management by absolute discretionary accruals. Panel A documents OLS regression coefficients. Variables significant at 10% are followed by *, variable significant at 5% by **, and variable significant at 1% by ***. Panel B documents difference between regression coefficients. The p-values are given in the parenthesis. Coefficients that are significant at 10% are shown in bold. Panel A: Regression coefficients Absolute discretionary accruals Equation (1) Equation (2) Pre-IPO 0.309*** 0.247*** 0.162** ** IPO 0.331** 0.261** 0.231* 0.189* Post-IPO Size ** ** Debt 0.008** 0.011** Sales ** ** Growth 0.088*** 0.085*** Industry Dummies No Yes No Yes Year Dummies No Yes No Yes No. of observations R² Adjusted-R² F-value Significance of F- value Panel B: Difference between coefficients Absolute discretionary accruals Equation (1) Equation (2) IPO Vs. Pre-IPO (0.863) (0.907) (0.515) (0.583) IPO Vs. Post-IPO (0.010) (0.021) (0.051) (0.060) Pre-IPO Vs. Post- IPO (0.002) 0.25 (0.002) (0.038) (0.040)