Corporate financial practices of the auto parts industry

Table of Contents

1. Executive Summary

2. Introduction

3. Dividend and Dividend Policy
3.1. Lintner’s Model
3.2. Modigliani-Miller Theorem (MM; Miller & Modigliani, 1961)
3.3. Information Asymmetry & Signalling
3.4. The Free Cash Flow Hypothesis
3.5. Conclusion

4. Mergers & Acquisitions (M&A)
4.1. Overview ofM&A Activity
4.2. A more in-depth look
4.3. Created synergetic value of one example deal

5. Ownership Structure
5.1. Ownership distribution
5.2. Ownership distribution and performance
5.3. Ownership identity and performance

6. Interrelationships
6.1. Dividend Policy and M&A
6.2. M&A and Ownership Structure
6.3. Dividend Policy and Ownership Structure

7. Conclusion

8. Bibliography
Appendix A: Summarized dividend data
Appendix B: Relationship between DPS and Share Price
Appendix C: DPR vs. DPR 5 years ago
Appendix D: DPS Changes
Appendix E: Delay EPS to DPS
Appendix F: Correlation FCF to DPS
Appendix G: M&A Industry Data
Appendix H: Detailed information on deals
Appendix I: Target Financials
Appendix J: Ownership Structure
Appendix K: Ownership Identity
Appendix L: Performance difference in Ownership Identity
Appendix M: Performance difference in Ownership Structure
Appendix N: Interrelationship M&A and Dividend
Appendix O: Interrelationship Ownership Structure and M&A
Appendix P: Dividends and Ownership Structure
Appendix Q: Complete dataset used for analysis of dividends

1. Executive Summary

Research Purpose: This report aims to evaluate the validity of corporate financial practices in the under-researched auto parts industry, additionally considering the ownership structure of companies because of a division in share distribution. Recent economic struggles have impacted the automobile industry, resulting in stagnating sales on a global scale (Haugh et al, 2010). This affects the auto parts industry, a direct supplier to automobile companies, aftermarket and exchange parts. Rising raw material costs have increased the pressure on auto parts manufacturers, as buyers disallow price increases. Market stimulus, mostly from Asia-Pacific, is forecast to increase total parts market by 20% by 2015 (Datamonitor, 2011). Research Design: This project discusses major corporate finance theories, focusing on data gathered from 547 companies through the ThomsonOne database (less when complete information was not available) aiming to provide a representative data set for analysis. Findings: Dividend payout, although ubiquitous, varies greatly within the industry, with few companies adhering to MM's irrelevance principle. Dividends fulfil the role of information transmission and the reduction of free cash flow. M&A, mostly conglomerate or horizontal, occur in waves and merger rumours lead to share price increases. Ownership distribution varies greatly, which does not affect company performance in contrast to ownership identity. No interrelationships were found between the three concept areas.

Research limitations: There are shortcomings inherent to this analysis. The sample size used (n=547) is limited by the scope of the accessed databases and only covers a maximum of 13% of the total auto parts market, with a size of $478.9bn (Datamonitor, 2011). Therefore, this report may not consider all industry trends. The exclusion of private companies, who only partially make their data available, may have impacted results. Due to limited resources, statistically inferior methods were applied which lead to the loss of analytical power. Nevertheless, this analysis seeks to provide considerable insight into the auto parts industry. Research applications: This analysis provides valuable theoretical tools for management as well as practical guidance for managers, which could enable better decision-making. Originalitv/Value: This report breaks new ground on the under-researched application of corporate finance principles in this industry. Further studies should cover a larger sample and database and draw linkages between the automobile and auto parts industry and investigating common corporate finance practices. Lastly further studies may focus on strategies employed during takeovers, especially linking ownership structure to defensive strategies.

2. Introduction

The auto parts industry^[1], total market size $478.9bn (Datamonitor, 2011), consists of crash repair, wear & tear parts, mechanical parts, consumables & electronics and service parts^[2] (see Figure 1).

Mechanical Parts 19%

Figure 1: Elements of the Auto Parts Industry. Source: Datamonitor

illustration not visible in this excerpt

The distribution of sales varies significantly around the world, with the Americas taking the majority share, followed by Europe and Asia-Pacific (see Figure 2).

Figure 2: Sales ofAuto Parts byArea. Source: Datamonitor

illustration not visible in this excerpt

The auto parts industry is heavily fragmented and highly competitive, with market leaders holding just over 2% of global revenues (see Figure 3). Even with the sample size of this report (n=547), only 13% of total revenues in the industry are represented.

illustration not visible in this excerpt

Figure 3: MarketShares. Source: Datamonitor.

3. Dividend and Dividend Policy

Dividends^[3] and dividend policy^[4] are an unsolved problem in finance. As Allen and Michaely (1995) point out, 'much more empirical research on the subject of dividends is required before a consensus can be reached.' The following analysis will outline key contributions and investigate whether the sample obtained is able to add value to the dividend debate.

3.1. Lintner's Model

Lintner (1956) provided the first empirical study on the topic of dividend and dividend policy. Firstly, he observed that companies tend to set long-term dividends-to-payout ratios (DPR^[5] ). Secondly, Lintner noticed that dividend policy is not altered until it has been validated that higher levels of earnings are tenable, thus concluding that there is a delay in the change of DPR after earnings increase.

Drawing on companies with sufficient data for analysis (n=233)^[6], current DPR was compared to that of 5 years prior. There was a significant difference in the ratios between 5 years ago (M=28.89, SD=18.16) and most current (m=22.9, SD=19.25; t(232)=-5.074, p<0.05), showing that on average, companies decreased their DPR.

Investigating Lintner's second observation, it was examined whether earnings^[7] were more closely associated with dividends^[8] of the same or later time periods. If the difference between earnings and dividends within same time periods is larger than the difference in following ones8^[9], then a delay in dividend increases following earnings increases has occurred. In the sample from the auto parts industry (n=14)^[10], this was not encountered. The difference between EPS and DPS from same time periods was larger than following time periods (17.51 >15.76 and 64.85 > 22.3).

In conclusion, neither of Lintner's findings were obtained in the sample drawn. This may be an effect of the recent financial crisis, where earnings have decreased (Haugh et al, 2010). However, DPS has not decreased, suggesting that companies prefer to keep the amount of dividends steady, as opposed to DPR.

3.2. Modigliani-Miller Theorem (MM; Miller & Modigliani, 1961)

The Modigliani-Miller theorem is one of the most influential concepts in corporate finance. The theory, also called the 'irrelevance principle' (Dybvig, 1991), states that in perfect and complete financial markets^[11], a firm's value is unaffected by its dividend policy. MM argue that investors can create their own preferred income stream by the use of homemade dividends^[12]. Moreover, considering taxation, investors will form clienteles with preferences for particular levels of dividend yields; altering dividend levels only leads to a change in shareholder clientele.

Contrary to traditional models, which argue that share price is the sum of all future dividend payments (Hiller et al, 2010), the irrelevance principle states that companies need not pay dividends at all. In fact, 31% of companies in the auto parts industries (n=415)^[13] did not pay dividends in 2010. However, dividends of the remaining 69% are substantial and payout is widely spread (Range: 0-2.64, M=0.146, SD=0.32) (see Figure 4).

If shareholders can make use of homemade dividends, share price and DPS should not be correlated. Using a Pearson's product-moment correlation coefficient to assess the relationship between DPS and share price^[14], there was a strong positive correlation in the sample of the auto parts industry (n=500)^[15] between the two variables (r=0.946, p<0.05).

A scatterplot summarizes the results (see Figure 5). Increases in share price were correlated to increases in DPS, thus providing evidence contrary to MM's predictions.

The clientele effect assumes there should be no major changes to DPS over long-time periods. However, in the auto parts industry (n=214)^[16], this is not the case. On average, DPS grew by 5% over a period of five, 4.5% over a period of three and 79% over a period of one year (see Appendix D). Conversely, some companies pay out more constant dividends. In the same sample, 40% of all companies had small or no change to their dividends over a five-year, 30% over a three-year and 22% over a one-year period (see Figure 6). Although overall DPS fluctuates, some companies smooth dividends over a longer period of time and adhere to the clientele effect (Marsh & Merton, 1987).

illustration not visible in this excerpt

Figure 6: Dividend Changes over a period ofl, 3 and 5years.

In conclusion, the analysis of the auto parts industry was able to show that the irrelevance proposition provides mixed results, possibly due to the underlying settings of MM, who assume perfect and complete financial markets. Other models discussed hereafter, such as information asymmetry and signalling, may better account for the mixed findings.

3.3. Information Asymmetry & Signalling

Deviations from MM's dividend proposition can only be obtained if underlying settings are violated. Building on the notion of 'asymmetric information'^[17], Bhattacharya (1979) aims to explain the reasons for dividends although they are tax disadvantaged^[18]. He reasons that dividends carry information effects, signalling expected cash flows in an imperfect- information setting.

As a measure of signalling effect, Miller & Rock (1985) argue that the bid-ask spread is significantly higher one day prior to a dividend announcement than on other days. However, it is beyond the scope of this project to investigate these findings^[19].

3.4. The Free Cash Flow Hypothesis

So far, the literature mentioned assumed that managers are perfect agents of investors. Others, such as Easterbrook (1984), argue that managers are imperfect agents, questioning how managers' interests may be aligned to those of shareholders. Although presumably irrelevant (MM), dividends are ubiquitous. Easterbrook (1984) suggests that dividends enable firms to remain on the capital market, where monitoring of managers is available at a lower price than agency cost^[20]. A dividend reduces free cash flow (FCF) and reduces management's scope to increase its personal utility.

Hence, if FCF is high, dividends are high in order to reduce cash flow available at manager's discretion. Using a Pearson's product-moment correlation coefficient to assess the relationship between FCF and DPS in the sample of the auto parts industry, there was no correlation between the two variables (r=0.028, n=471, ns)^[21], contrary to Easterbrook's theory. This may be due to other functions that dividends fulfil, discussed in Part 4.3.

3.3. Conclusion

The theories described above are the main concepts in the area of dividends and dividend policy. Some companies have stable DPRs, others (31%) do not pay out dividends at all and few (~40%) companies provide stable payments. Overall, dividends are substantial and may have other reasons not covered by MM, such as providing information or reducing free cash flow. Although many additional concepts have recently been developed, it is outside the scope of this report to evaluate these.

4. Mergers & Acquisitions (M&A)

Most commonly, M&A's motive 'is simply, that the purchasing firm considers the acquisition to be a profitable investment' (Pautler, 2001), for example due to created synergies^[22]. M&A can be horizontal (both firms in the same industry), vertical (firms in different stages of production) or conglomerate (unrelated firms) (Hiller et al, 2010). However, some argue mergers may not add value at all (Lubatkin, 1989).

4.1. Overview of M&A Activity

In the last 20 years, 2463 deals of M&A in the auto parts industry were recorded. The number of deals per year fluctuates greatly and usually occurs in 'waves' (Figure 7). Auster & Sirower (2002) claim this is a persistent finding across many industries, the nature of which is unknown^[23]. I put a comment by the footnote

illustration not visible in this excerpt

Figure 7: M&A Activity peryear

Taking a closer look at the type of M&A deals (n=2463), the majority of all M&A's were either horizontal or conglomerate, with only a minority being vertical (Figure 8). Firms more readily invested in similar or completely unrelated companies than in companies of different production stages.

illustration not visible in this excerpt

Figure 8: Type ofM&A activity

4.2. A more in-depth look

The average deal size (n=123)^[24] was $299.64m with a large spread (Range: 0.12m to 4,903.59m, SD: 644.04). Figure 9 shows the dispersion ofM&A deals in more detail.

illustration not visible in this excerpt

Figure 9: spread of deal values.

In order to convince target companies' shareholders to accept an offer, acquiring companies often pay a merger premium^[25]. Using Netter et al's (2011) methodology^[26], premiums in the sample size of the auto parts industry were quite significant - 17% on average, with a very high variability (SD=0.5).

Analogous to the signalling effect of dividends^[27], Muth (1961) argues that expected increases in value may lead to present share price increases. An example of this is brought forward by Nielsen & Melicher (1973), who contend that merger rumours, due to the market expectation of a premium, lead to an increase in share price. This was reflected in the data obtained from the auto parts industry. Setting the share price at announcement to 100%, the data shows an increase in historical stock prices predating the sale. Across all deals, there was an average increase of 6% (83% to 89%) in the share price from 4 weeks prior to sale as compared to 1 day prior to sale (see Figure 10).

illustration not visible in this excerpt

Figure 10: Development oftarget company share price.

Many academics have long argued the reasons behind companies becoming a target for acquirers. The most consistent finding is the underperformance of target companies versus acquiring companies (Sorensen, 2000). Looking at profitability of acquired companies in the auto parts industry, targets overall were operating at a loss^[28] (n=189, M=-3.2)^[29]. However, there is a very wide spread (SD=49.13) and, due to the limitations of earnings as profitability measure^[30], other measures such as net income (M=74.18m USD, SD=546.29) do not support this hypothesis. Therefore, other proposed theories (see Davis & Stout, 1992) may more appropriately describe why companies become a target.

4.3. Created synergetic value of one example deal

Due to the highly fragmented nature of the industry^[31], it is difficult to select a representative case of synergetic value creation. However, as an example, Faurecia acquired Emissions Control Technologies^[32]. Looking for synergy, Faurecia's share price was plotted over a 1-year period, starting 4 weeks prior to the merger^[33] and compared against the S&P 500 (setting share price at Oct-2 to 1). Over this time period, Faurecia outperformed the S&P 500 (Faurecia: 1.44, S&P 500: 1.165) (see Figure 11). However, such a measure of synergy is flawed as other factors such as other mergers, macro-economic changes or merger-unrelated performance increases may have impacted share price.

illustration not visible in this excerpt

Figure 11: Relative Share prices of Faurecia and S&P 500.

5. Ownership Structure

As far back as Adam Smith (1776), it has been noted that 'negligence and profusion [...] must prevail [...] in the management of [...] a company'. Jensen & Meckling (1976) describe the agency relationship^[34] and argue that there is reason to believe that agents may not always act in the principal's best interest. The term 'agency cost' defines the expenditure or loss incurred due to the agency relationship^[35].

5.1. Ownership distribution

On the basis of 511 companies, Demsetz & Lehn (1985) conclude that ownership varies widely for two reasons; (a) a large dispersion of ownership increases the risk of managers shirking; (b) a smaller dispersion of ownership increases monitoring of shareholders. The authors reason that ownership structure is endogenous and a firm's capital structure is tangent upon its own strategy.

Using variations of Demsetz & Lehn's measures A5 and A2 0^[36], the distribution of ownership of the 10 biggest auto parts companies varies greatly, A5 averaging 37.8% (SD=21.71), A10 averaging 46.5% (SD=19.28) and total shares on market averaging 74.% (SD=17.86).

5.2. Ownership distribution and performance

Academics widely agree that ownership structure may have an important role in determining a firm's performance: Leech & Leahy (1991) argue that in wide ownership dispersion, no individual or group has the voting power or incentive to exercise control over agents and enforce profit maximisation; Oswald & Jahera (1991) maintain that large shareholders are able to pressure managers to increase company performance; and Thomson & Pedersen (2000) argue that company performance is a bell-shaped function of the ownership share of the largest owner.

In order to scrutinize Thomson & Pedersen's (2000) hypothesis, companies in the auto parts industry were grouped into either 'minority shareholders only' (n=6) or 'companies with both minority and majority shareholders' (n=4) (see Figure 12 and 13 for examples).

illustration not visible in this excerpt

Figure 12:An example ofa company with minorityshareholders only (here: Visteon)

illustration not visible in this excerpt

Figure 13: An example ofa company with both minority and majority shareholders (here: Federal Mogul)

Comparing the average net margins^[37] of both groups over a three-year period, there was no significant difference found (Mminorityoniy=2.4%, Mmajority&minority—2.9%), showing that ownership structure was not an endogenous factor (Demsetz & Lehn, 1985). There are numerous reasons for the insignificance in comparison to previous findings, the most obvious being an inadequate sample size due to limited resources.

5.3. Ownership identity and performance

Thomson & Pedersen (2000) also contend that the identity of owners is equally important as ownership structure, discovering^[38] that company shareholder value increases with ownership share if the largest owner is an institutional investor^[39].

[...]

---

^[1] defined as 'Manufacturers and distributors ofnew and replacement partsfor motorcycles and automobiles, such as engines, carburettors and batteries. Excludes producers oftires, which are classified under Tires.' (Industry Classification Benchmark code 3355) (http://www.scribd.com/doc/36837305/ICB-classification)

^[2] The Datamonitor report does not cover auto parts that supply assembly parts within the OEM (original equipment manufacturer) scheme.

^[3] the payments a corporation makes to its shareholders (Sheffrin, 2003)

^[4] the amount of dividend that a company decides to payout to shareholders (Hiller et al, 2010)

^[5] the percentage of earnings paid out to shareholders as dividends (Hiller et al, 2010)

^[6] only 233 companies out of the 547 total provided data on DPR over a 5-year period

^[7] in this case earnings per share, or EPS

^[8] in this case dividends per share, or DPS

^[9] e.g. ifEPS5yearsago - DPS5yearsago > EPS 5yearsago - DPS3yearsago

^[10] companies with earnings growths over the last 5 years and with sufficient data for analysis

^[11] e.g. with no taxes, agency costs, asymmetric information, efficient market etc.

^[12] investment income from the sales of shares held by a shareholder (MM)

^[13] with sales above 50 million

^[14] most current DPS was compared to average share price over the last 12 months

^[15] filtering companies which did not have data on 12 months historical share price available left 500 companies for analysis

^[16] excluding companies that don't pay dividends

^[17] defined by Miller & Rock (1985) as the manager knowing more than the outside investor about the true state of the firm's current earnings.

^[18] in paying out dividends, companies pay corporate tax and shareholders pay income tax. Capital gain is only taxed once.

^[19] As data was not readily available.

^[20] See Jensen & Mekhling (1976), described further in Part 3.

^[21] Only 471 companies provided information ofFCF.

^[22] defined as the shareholder value created above that of the sum of the two individual companies (Chatterjee, 1986). Synergies have many sources, such as efficiencies, financial and tax benefits and market power effects (Pautler, 2001).

^[23] Although some authors (e.g. Kallunki et al, 2009) have argued that market conditions and technology levels are key drivers

^[24] only 123 deals provided complete information

^[25] an offer which exceeds the value of the target company prior to the merger (Nielsen & Melicher, 1973)

^[26] Price per Share acquired minus share price 4 weeks prior to share price announcement

^[27] as described above; see Bhattacharya (1979)

^[28] as measured in earnings per share

^[29] deals in the last 20 years which provided information on performance

^[30] e.g. differing accounting standards

^[31] see above; the four biggest companies own less than 2% of market shares

^[32] finalized on February 10th 2010 for the total value of 407.81m USD; announced on November 2nd 2009.

^[33] due to market expectations (Nettler et al, 2011)

^[34] the contract under which a principal (shareholder) engages another person, the agent (managers), to perform some service on their behalf, which involves the delegation of authority to the agent (Jensen & Meckling, 1976)

^[35] e.g. ; monitoring (by the principal), bonding expenditure (by the agent) and the residual loss (reduction in welfare).

^[36] A5': the percentage owned by the 5 biggest shareholders; 'A10': the percentage owned by the 10 biggest shareholders; 'total shares on market'

^[37] as a measure of performance

^[38] on the basis of a study of the 435 largest overall European companies

^[39] However, the authors only employ a short-term horizon. It may well be that institutional investors trim companies for short-term profits