Executive compensation in less regulated markets: the impact of debt monitoring

Abstract

This paper shows that in the lightly regulated Alternative Investment Market (AIM) voluntary corporate board structures might not reduce agency costs between shareholder and executive directors. In this less regulated market, we find that the extent of debt affects executive pay. In addition, the theoretical determinants of executive pay affect CEO and other executives’ pay and incentives differently in this market. We find no evidence that debt levels affect CEO pay in a matched sample of Main Market firms. Our results suggest that debtholders could be better monitors of executive directors’ actions, in comparison to voluntary governance committees in less regulated markets.

KEYWORDS:

• Debt monitoring
• corporate governance
• board structures
• executive compensation
• executive incentives
• less regulated markets

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Helena Pinto http://orcid.org/0000-0002-7931-8219

Notes

1 Doidge et al. (2009) and Chhaochharia and Grinstein (2007) show that regulated markets impose significant listing and corporate governance costs.

2 In Europe the Capital Requirements Directive (CRD) IV covering prudential rules for banks, building societies and investment firms, states that from 2014 there is a cap on the variable pay to staff.

3 Yermack (1995) and Cambini, Rondi, and De Masi (2015) also show that regulation can influence executive pay.

4 Alternatively, the light approach to regulation can result in unscrupulous managers being able to capture boards and influence their remuneration packages extracting higher pay levels (Bebchuk and Fried 2003; Cyert, Kang, and Kumar 2002; Hu and Kumar 2004).

5 A number of studies have examined executive compensation in emerging economies where the governance regulations and law enforcement are weaker than developed countries (see e.g. Conyon and He 2011; Gallego and Larrain 2012; Liang, Renneboog, and Sun 2015). However, these lightly regulated environments apply to all firms in these emerging markets. We are motivated to examine executive compensation in less regulated markets in developed countries as AIM firms make a choice to list in this less regulated market rather than the Main Market. This is despite in many cases being of similar quality to Main Market firms (Nielsson 2013) and many meeting the listing requirements of the Main Market (Doukas and Hoque 2016).

6 AIM companies are encouraged by the LSE to adhere to the Quoted Companies’ Alliance (QCA) guidelines but there is no mandatory requirement to do so (LSE 2010). The guidelines do state that the board should have at least two independent directors, the company should have separate remuneration (only with independent directors); audit and nominations’ boards and the role of chair and CEO should be separate. Appendix A presents some of the guidelines of the QCA in more detail.

7 Snell and O’Brien (2008) analysed the 100 largest AIM companies and found that although the large majority comply with some aspects of the combined corporate governance code, only 3% of those companies fully adopted it. In addition, Mallin and Ow-Yong (2010) reported different levels of information disclosure for AIM companies than Main Market firms.

8 There have been a series of important corporate governance guidelines emerging from UK policy reports that were translated into principles and provisions in the UK Corporate Governance Combined code (Cadbury 1992; Greenbury 1995; Hampel 1998; Higgs 2003; Walker 2009). Main Market companies are not obliged to follow the combined code, but they will have to carefully explain to shareholders when they decide not to comply. Shareholders can then discuss their position and through their voting power take action to influence the board. This requirement to ‘comply or explain’ does not apply to AIM companies.

9 In Appendix 2, we present a table with some of the central results presented in the previous literature for the UK Main Market. The literature on executive pay and incentives is extensive and thus the table presents the main results relevant for this study.

10 One of the reasons for the current interest in less regulated markets is the fall in the number of listings in the highly regulated markets (including the LSE Main Market, the US and the Japanese markets), driven mainly by the low new list rate, and the high delist rate. A number of stock exchanges have tried to create markets for smaller firms with less regulation but have failed (for example, Germany’s Neuer Markt and France’s Nouveau Marché). The success of AIM market in attracting a number of listings by growing and smaller firms is the light regulatory regime (Arcot, Black, and Owen 2007; Nielsson 2013).

11 We thank a referee for this suggestion.

12 The results found in regulated markets are mixed. Yermack (1995) found that leverage had no effect on option grants, but Lewellen, Loderer, and Martin (1987) found a positive relation between leverage and option grants. Fernandes et al. (2013) find a positive relation between leverage and total pay. Ortiz-Molina (2007) suggests a negative effect of leverage on equity like pay. Coles, Naveel, and Naveel (2006) show that leverage has a negative effect on the CEO wealth incentives (delta) but a positive effect on the CEO risk incentive (vega), albeit weakly significant at a 10% significance level.

13 We are aware that some non-executive directors may not be independent. According to the UK corporate governance code, companies should indicate in their annual reports if a director is independent. An independent director will not have, or had in the past, significant relations with company, i.e. a shareholder; a shareholder representative; auditor; employee; supplier or significant customer; family member; cross directorships; or call options on the company’s shares. This provision, does not apply to AIM companies and thus it is difficult to establish if non-executive directors are indeed independent, also many studies do use the percentage of non-executive directors as a proxy for board independence (Ozkan 2011).

14 Other specifications of the remuneration board variable used in robustness tests are a dummy variable which indicates the existence of a remuneration committee (not considering its composition as in Benito and Conyon 1999), and the percentage of non-executive directors on the remuneration committee (Conyon and Peck 1998).

15 There is also the argument that the separation of the roles of CEO and Chair can lead to additional agency costs and that the separation of the role does not really have an impact in company value (Brickley, Coles, and Jarrell 1997).

16 We thank a referee for this suggestion.

17 There is considerable debate in the literature on the types and measurement of variables to be included and the form of the estimating equation (e.g. current or lagged measures).

18 For example, in 2005 the largest company in our sample (using total assets as the measure of size) was 8168 times larger than the smallest company.

19 Similarly, Arcot, Black, and Owen (2007) show that on average 20% of the shares being traded on the AIM are owned by directors, employees, and members of the founding family of the AIM firm.

20 The CEO change dummy controls for the fact that new CEO might not have been in service for the all year and/or might have been given abnormally high grants to take the role.

21 In 2006, there were 1634 companies listed in the AIM and thus our initial sample represents 37% of the whole market.

22 For robustness in additional tests, we adjusted our sample so that only companies alive for at least 5 years were included and so that only companies alive through the 6 years were included. Our overall results do not change.

23 A two-sample Kolmogorov-Smirnov test suggests that we cannot reject the null hypothesis that the AIM and our sample are from the same distribution.

24 We follow the standard approach to valuing executive options (Jensen and Murphy 1990; Fernandes et al. 2013).

25 The same approach appears in Knopf, Nam, and Thornton (2002) and Coles, Naveel, and Naveel (2006).

26 Our control sample includes 176 firms listed in the Main Market. To be included in the sample the firms had to be alive for at least 4 years and compensation and financial information had to be available for this period.

27 As an example, Ozkan (2011) considered pay in UK companies with an average market capitalisation of £1828 m.

28 Not tabulated for brevity.

29 These last two results show that there are a significant percentage of firms that do not adhere to the corporate governance guidelines of the QCA (see Appendix 1).

30 Conyon and Murphy (2000) and Fernandes et al. (2013) show that US CEO receive significantly larger pay packages than UK CEO. Fernandes et al. (2013) argue that the difference in pay can be explained by the risk premium which US CEO receive since their pay packages are more heavily weighted towards performance related pay and thus are riskier than UK CEO pay packages

31 Ozkan (2011) also reports much larger pay packages, the average total pay for larger UK companies in 2005 (reported in 1999 constant terms) is £700,000. This is three times larger than the average values we found for our sample in 2005 (£211,914 – not tabulated).

32 Although the large proportion of total pay comes from salary, companies also use variable/performance related forms of pay. In 2005, salary represented 61% of total pay and in 2013 46%, with options contributing 34% of total pay (as opposed to 28% in 2005). These values are not tabulated.

33 In some cases, the company only discloses the total amount of cash paid to the board, including both executive and non-executive directors in this figure. For these cases, our per capita of cash is the total cash paid to the board divided by the total number of board directors. Thus our per capita figures for cash pay and total pay, could understate the amount paid to executive directors

34 Not tabulated.

35 Results not tabulated show that in 2008 only a small proportion of the options were in the money and the delta of the portfolio of options is small.

36 Similarly, Marshall, Kemmitt, and Pinto (2013) report an average share ownership by AIM CEO and other executives of 13.7% in 2006, the average CEO share ownership was 9.9% (2.96% for the other executives).

37 Since the delta of shares is equal to one the Conyon and Murphy measure for shareholdings is simply the CEO share ownership.

38 In Table 2 options grants and share awards have for every year a median of zero indicating that many companies do not grant options or shares.

39 Gregg, Jewell, and Tonks (2012) also considered the dynamics of executive compensation using GMM methods.

40 The literature has long recognised that the different elements of pay can result in different incentives (Jensen and Murphy 1990).

41 We also include a year of end change dummy and a CEO change dummy.

42 Boschen and Smith (1995), Main, Bruce, and Buck (1996), Conyon and Murphy (2000) and Boschen et al. (2003) all suggest the importance of modelling the dynamics of executive compensation. Models that only consider the contemporaneous response of compensation to firm performance implicitly assume that the long run relation between compensation and performance is the same as contemporaneous relation. In other words, a pay increase resulting from an unexpected good performance remains in the level of pay indefinitely (Boschen and Smith 1995).

43 The lagged levels are valid instruments since they are uncorrelated with the error term Δϵ_it of the model in first-differences.

44 Although, we believe that it is important to consider the dynamics of pay and thus the GMM methodology should be applied, for robustness and comparability we also run our regressions using a fixed effect regression model since this is the most widely used research method in prior executive pay studies.

45 The Sargan test does not indicate correlation between the instruments and the error term. Also, the serial correlation tests, AR(1) and AR(2) that the instruments are valid.

46 For brevity, we do not present the fixed effects panel data regression for the matching sample of Main market firms. The results are available on request from the authors.

47 In other words, if the wealth incentive is significantly larger, than our estimate in Table 2, then the monetary change resulting from a change of 1% in leverage will also become much larger.

48 For robustness we repeated all our tests using an alternatives definition of leverage (debt over book equity). Our results do not change.

49 The effect of board size on the risk incentive held by other executives is an exception. Executives other than the CEO of companies with larger boards have larger risk incentives.

50 The results on board size are difficult to interpret though since it is not clear from the literature if large or smaller boards represent better governance.

51 These results are consistent with the fixed effects methodology.

52 Due to data constraints we cannot test in model 2 the effect of remuneration committee on CEO pay. We nevertheless discuss the results previously found in the literature for the Main Market.

53 We do not have data on the remuneration committees of the companies in our matched Main Market sample.

54 We interpret the results on board size and CEO/Chair with caution due to number of observations.

55 Managers could want to reduce firm’s risk even if this results in lower pay packages. Similarly, Berk, Stanton, and Zechner (2010) suggest that employees will accept temporary pay cuts if this helps avoid bankruptcy.

56 The data for Altman Z score is from Fame database. Since we were not able to find the Altman score for some of our sample companies in these tests our sample is reduced to 1824 firm/year observations.

57 We do include CEO share of ownership in our analysis, but we omit outsiders’ ownership. Clearly, companies in which the CEO does not own any company shares can still have concentrated ownership.

58 This information is not available for all companies in our sample, and this reduced the number of company/years to 1759.

59 CEO option and share awards are still significant at a 10% significance level.

60 Tenure is the number of years in company and these data is from the annual reports.

61 We acknowledge that tenure can also be a proxy for managerial entrenchment but to the extent that we control for managerial share ownership we do not see this as a major concern.

62 In our sample 50% of the companies’ board is comprised of non-executive directors and around 70% of the companies have a full independent remuneration committee.