Hidden Ownership and Firm Performance: Evidence from Thailand’s Initial Public Offering Firms

Abstract

Previous studies have overlooked hidden ownership in their analysis, which could result in biased findings. This research utilizes unique data sources to uncover hidden ownership patterns and employs ordinary least square regression to investigate the relationship between hidden ownership and firm performance. The findings indicate that hidden ownership affects a firm’s performance, but not in the same manner as previously thought. Firms with hidden ownership actually perform better than those without. These results contradict the belief that hidden ownership leads to wealth expropriation from minority shareholders and negatively impacts a firm’s performance. The study also remains robust after accounting for potential endogeneity using an instrumental variable approach. The findings provide policy implications and contribute to the ownership and firm performance literatures.

1. Introduction

Ownership structure and corporate governance are core areas of corporate governance. However, many puzzles remain in empirical studies. One area that has been excluded from ownership structure studies is hidden ownership. Consequently, previous ownership calculations may yield biased results. Specifically, in Thailand, ownership structures are highly concentrated but not easily observable, as owners have incentives to split their shares and mask their identities through foreign offshore companies. The motivations may include self-dealing benefits, tax avoidance, flexibility to buy or sell, price manipulation, foreign quota reserves, and hostile takeover risks (Chernykh 2008).

Thailand’s emerging market offers a unique environment for studying the relationships between ownership structure, corporate governance, and firm value. Most Thai firms are heavily concentrated in the hands of their founding families (Wiwattanakantang 2001; Khanthavit et al. 2003). Most family businesses in Thailand have conflicts of interest between the majority and minority owners. Ineffective governance practices and weak market regulations in emerging countries may worsen the problems of agencies. In particular, when the timing of an initial public offering (IPO) is favorable, private companies transition to the public market. However, the company’s ultimate owners want control. They most likely use nominations from local or offshore companies to mask their identities. Additionally, a significant degree of information asymmetry allows owners to behave according to their desires (Kim et al. 2004). It would be interesting to examine how IPO firms act when they have a hidden ownership structure since the ultimate owners’ objectives are likely ambiguous.

Identifying the true ownership of a company can be a complex and challenging task, as many ultimate owners use tactics such as splitting shares and utilizing offshore companies to conceal their identities. This makes it difficult to accurately aggregate the propositions of ultimate owners in empirical studies, as some studies may not include this hidden ownership in their tests. However, this paper has access to unique data sources that can aid in identifying the chain of ownership for a company’s shares. By analyzing these distinctive data sources, a better understanding of ownership control patterns can be gained. The aim of this study is to investigate the relationship between hidden ownership and firm performance. The sample used for this study includes all firms that went public on the stock exchange of Thailand (SET) between 2011 and 2015. The results show that hidden ownership is positively related to firm performance and that firms with hidden ownership outperform those without hidden ownership. This suggests that hidden ownership may not be detrimental to a firm’s value but may, in fact, increase it. These findings contradict the theory that hidden ownership may cause wealth expropriation from minority shareholders to the ultimate owner, negatively affecting a firm’s performance (return on equity (ROE), return on asset (ROA), and Tobin’s Q).

Furthermore, additional analysis shows that firms with hidden ownership have a greater proportion of foreign institutional investors. As institutional investors have enhanced monitoring skills, they can scrutinize management more thoroughly than individual shareholders and do so at a relatively low cost (Shleifer and Vishny 1986). Thus, numerous foreign institutional investors may have a substantial influence on a firm’s governance structure and a positive effect on firm performance. Nevertheless, we are aware that the causation effect starts with high-performing firms that attract well-connected financial advisors (FAs), whose clientele includes institutional investors and recommends offshore ownership for transactional purposes. Additionally, we analyze the performance of firms with hidden and non-hidden ownerships before an IPO. The results show that firms with hidden ownership outperform firms without hidden ownership. However, this may not be exclusively due to the presence of many institutional investors. These firms have already achieved high returns before their IPO. The firm’s outstanding performance attracts prominent FAs with solid ties to institutional investors. Strong connections between reputable FAs and numerous institutional investors can further influence the continuously high firm performances, suggesting that firms strategically implement offshore hidden ownership for transaction purposes.

The findings of this study may provide new insights into hidden ownership puzzles and contribute significantly to the current corporate governance literature. The unique data sources used in this study can uncover hidden ownership structures and challenge conventional understanding of ownership structures and firm performance. Previous empirical research has not included hidden ownership in the analysis. Therefore, this pioneering study provides new insights into hidden ownership, improving the understanding of ownership control patterns. Additionally, the results of this study suggest that policymakers and regulators should carefully review the existing rules and regulations regarding disclosure and transparency requirements. This review will help mitigate the issue of information asymmetry between the ultimate shareholders and minority shareholders.

The remainder of this paper is organized as follows. Section 2 reviews the relevant literature and develops the hypotheses. Section 3 discusses the study’s research methods. Section 4 discusses the results, and Section 5 concludes the paper.

2. Literature Review and Hypothesis Development

2.1. Theoretical Background

The concept that a firm’s ownership structure directly influences its performance is a central and ongoing topic of debate in the corporate finance literature. Demsetz and Lehn (1985) emphasized that different ownership structures have various objectives and levels of managerial oversight. In particular, the impact of managerial ownership on firm value has long been questioned in corporate governance. Managerial ownership may improve firm performance by aligning the interests of managers and owners (Jensen and Meckling 1976). Simultaneously, an excessive increase in management ownership may relate to an entrenchment effect, which harms the firm’s value (Morck et al. 1988; Stulz 1988).

The widely owned firm with dispersed and passive shareholders outlined by Berle and Means (1932) is representative of only a few U.S. and UK companies, according to a large body of recent empirical evidence. Recently, cross-country research has shown that pyramidal ownership and cross-holding structures are being used by most companies worldwide. A study to identify the ultimate ownership by La Porta et al. (1999) utilizes ownership structure data of large corporations from 27 developed and developing countries. The results indicate that most firms are controlled by families with a high level of ownership concentration. As mentioned, these ultimate shareholders use a pyramid ownership structure and their own participation in management to gain excess control rights over their cash flow rights (Claessens et al. 2002; Bebchuk and Roe 1999). Family owners and managers with significant control over a company may prioritize their family interests over those of minority shareholders. This can result in principal–principal agency issues, such as the expropriation of wealth from minority shareholders and the entrenchment of controlling family shareholders (Morck and Yeung 2003; Friedman et al. 2003; Anderson and Reeb 2004; Andres 2008; Johnson et al. 2000; La Porta et al. 1999). Therefore, a family firm with high ownership concentration within a company can lead to agency issues that ultimately harm the company’s performance (Hart 1995; Pérez-González 2006; Schulze et al. 2001). Consistent with the research conducted by Claessens et al. (2002), it has been observed that the dominant shareholder possesses the ability to establish a robust position and exploit resources, which can result in negative consequences for both overall business performance and the economy at large (Chrisman et al. 2012; Morck et al. 2000).

2.2. Ownership Structure and Firm Performance

A classic study by Berle and Means (1932) outlines dispersed ownership in a typical corporation and argues for its negative relationship with firm performance. Several empirical studies have investigated the relationship between ownership structures and firm performance. Morck et al. (1988) investigate how different types of ownership affect a firm’s value. Their research shows that the entrenchment effect predominates when the ownership range is between 5% and 25%. The alignment effect influences a low or high fraction of ownership when the range of ownership is between 0% and 25% or more than 25%, respectively. Demsetz and Villalonga (2001) examine the relationship between firm performance and ownership structure in the US economy. They have found no statistically significant relationship between firm performance and ownership structure, similarly to Al Farooque et al. (2020). This result fits the idea that even though diffuse ownership might worsen some agency problems, it can also lead to benefits that significantly compensate for the problems.

In the context of letter research, there are two contrasting viewpoints on the relationship between family-owned businesses characterized by concentrated ownership and their overall performance. Academic researchers have discovered that family-owned firms perform better than non-family-owned firms. For example, using panel data, Andres (2008) examines how founding family ownership affected the performance of 275 listed German companies. The study shows that family businesses are more profitable than companies with other block holders or widely held firms. Nevertheless, as a family business, only companies in which the original family members are still active on the executive or supervisory board perform better. Anderson and Reeb (2003) conducted a study that explored the relationship between the ownership of firms by founding families and their performance. Their findings revealed that family-owned firms outperformed non-family-owned firms, particularly when a family member assumed the role of CEO. Similarly, Maury (2006) conducted a study to evaluate the performance of family firms in Western European countries. The results of the study indicate that firms actively managed by family members tend to have better overall performance. In the context of developing economies such as Thailand, Wiwattanakantang (2001) research indicates that controlling owners’ involvement leads to improved performance when measured by accounting metrics such as ROA and sales-to-asset ratio. This observation aligns with the conclusions drawn by Yammeesri et al. (2006), which suggest a positive correlation between concentrated ownership and firm performance.

In contrast, numerous scholars have discovered a negative correlation between family-owned businesses and the performance of these firms. Family-owned businesses may encounter many challenges that have the potential to impact their financial performance negatively (Villalonga and Amit 2006). As evidenced by Xu et al. (2015) and Beuren et al. (2016), certain firms may exhibit a propensity to appoint family members to managerial positions, even in cases where such individuals may not possess the required managerial skills. In addition, Fan and Leung (2020) investigate the functions of ownership structures within a controlling family. They focus on the family trust structure, which is used by many families worldwide to preserve ownership. Share transfer restrictions help maintain power in the family; however, they might lead to problems with family shirking. A study on Hong Kong’s publicly traded family businesses shows that trust-controlled companies that are more aware of these issues tend to increase dividends at the expense of long-term investments, and thus, they underperform. In accordance with Benjamin et al. (2016), it has been found that when the ownership level of a family shareholder is higher, there tends to be a higher probability for a firm to pay out larger dividends. However, this may have a negative impact on the firm’s overall performance. Limpaphayom and Ngamwutikul (2004) examined the post-issue operating results of businesses that conduct seasoned equity offerings (SEO) on the Stock Exchange of Thailand (SET). They provide evidence of SEO firms’ poor operational performance after their IPO. Additionally, a link exists between the declining operational performance after an issue and insider ownership concentration. Kim et al. (2004) examined the relationship between managerial ownership and the evolution of Thai IPO firms’ performance in the context of developing markets. Their findings revealed a negative correlation between management ownership and subsequent changes in firm performance, particularly at the medium-to-high ownership level. Similarly, the study conducted by Sciascia and Mazzola (2008) revealed that firms characterized by familial participation in managerial roles have diminished levels of performance. In line with the findings of Harymawan et al. (2019) and Jara Bertin and Iturriaga (2014), it is evident that firms without family ownership exhibit superior performance compared to firms with family ownership.

Numerous studies on ultimate ownership focus on institutional ownership. As knowledgeable and large owners, institutional investors are incentivized to scrutinize management at a lower cost than individual shareholders (Shleifer and Vishny 1986). This can significantly impact the governance structure and a firm’s course of action, which is one of the main benefits of institutional ownership. Pedersen and Thomsen (2003) demonstrate a positive influence on a firm regardless of whether its major shareholder is a financial institution or an enterprise. Foreign investors are distinct types of ultimate owners that share specific similarities with institutional investors. Cornett et al. (2007) examined the connection between institutional investor involvement and the operating performance of large corporations. They reported a strong link between a company’s operating cash flow returns and the number and percentage of institutional investors. Additionally, Elyasiani and Jia (2010) examined the relationship between corporate firm performance and the extent and stability of institutional ownership. Considering shareholding percentage, they reported a positive correlation between firm performance and institutional ownership stability. This relationship is strong and supports the idea that stable institutional investors play an important role in monitoring. Several studies, particularly those on developing and emerging economies, demonstrate that institutional investors favorably impact firm performance, which may be attributed to their enhanced monitoring abilities.

Overall, an examination of the existing empirical research reveals considerable cross-country variance in the relationship between ownership structure and firm performance. There are two divergent perspectives on the correlation between concentrated ownership in family-owned firms and their overall firm performance. These differences in findings might be because the impact of ultimate ownership on a firm’s performance is highly context-dependent and depends on factors such as laws, institutions, and the economic environment (Chernykh 2008). Moreover, hidden ownership has been overlooked in research on ownership structure. Consequently, previous estimations of ownership may yield inaccurate results. In Thailand, ownership structures are highly concentrated (Wiwattanakantang 2001; Khanthavit et al. 2003) but not visible because owners are motivated to divide their shares and conceal their identities through the use of foreign offshore companies. The outcomes of this study may assist in revealing shareholder structure and revolutionizing existing beliefs about ownership structures and corporate performance.

This study provides further evidence on the relationship between hidden ownership and corporate performance. As agency problems are common in emerging markets such as Thailand, the private benefit of control is likely to be removed from minority shareholders, mainly when the ultimate owners hide via offshore holding companies. Thus, the motivation of the ultimate owners is questionable. The following hypothesis is proposed:

H1. 

Wealth expropriation from minority shareholders to the ultimate owner causes a negative correlation between hidden ownership and firm performance.

3. Research Design

3.1. Data and Sample

Thailand’s emerging market is our main focus because it has several advantages. (1) Most family-owned firms have conflicts of interest between majority and minority shareholders. (2) An inadequate market for corporate control in emerging markets and inefficient governance mechanisms may worsen the agency problem. (3) Complete information is accessible. These benefits boost our confidence in the accuracy of our data and findings.

All firms that went public on SET between 2011 and 2015 were included in the sample used for this research. We examined data from 2011 and followed the firm’s performance over three years. The regulations of the Bank of Thailand restricting who may own a bank or financial institution led to the exclusion of companies in the financial services sector from the sample, including banks, insurance companies, finance and securities firms, listed mutual fund firms, and property investment funds. This process yielded a sample of 49 firms. The SET provides additional financial information through SETSMART, and financial data were gathered from Datastream, a Thomson Financial service. More information is gathered manually from sources such as business filing reports, the annual company reports, and the minutes of general meetings with the board members, firm affiliates, and family ties between board members. This study investigates the chain of ownership of a company’s shares using special data sources to uncover hidden ownership patterns.

Table 1. Sample selection.

Panel A: Sample selection
Total number of firm-year observations from 2011 to 2015	180
Less: Firm-year observations in the financial industry	33
Less: Firm-year observations dropping due to insufficient data	11
Final sample size from 2011 to 2015	136
Panel B: Distribution of IPO firms by industry
Industry	N	%
Agribusiness and Food Industry	26	19.12
Consumer Products	6	4.41
Industrials	12	8.82
Property and Construction	34	25.00
Resources	18	13.24
Services	40	29.41
Total	136	100.00

provides a summary of the sample determination process in Panel A, while Panel B displays the distribution of firms across six industries. Our data reveal that the services industry has the highest number of firm-year observations, comprising 29.41% of the sample. The property and construction industry follows closely with 25.00%, while the agribusiness and food industry accounts for 19.12%. Conversely, the consumer products industry has the lowest number of firm-year observations, accounting for only 4.41% of the sample.

3.2. Variable Definition

• Dependent variables: As revealed by Brown and Caylor (2009), all performance indicators have specific flaws. As the measurement errors of performance indicators are only partially associated, researchers should consider many performance measures rather than relying on a single one. Tobin’s Q, ROA, and ROE were used as dependent variables in this research on firm performance. ROA and ROE are both based on accounting; however, Tobin’s Q is a measure of performance based on the market. These measurements vary from one another in two crucial ways (Demsetz and Villalonga 2001). The first is in terms of time, with accounting-based performance measures looking backward and market-based performance measures looking forward. The second distinction concerns those who measure performance. Each has its benefits and drawbacks.
• Independent variables: We investigated the direct and indirect ownership chains of a company’s shares using unique data sources. For direct shareholdings, we considered not only people with the same last name as a family but also people with known familial ties regardless of last name. Several sampled firms had indirect shareholdings, and the ultimate owner of the sampled firm held shares via another public or private company. For each firm in the sample, the Business Online (BOL) database was used to track any indirect ownership of shares via the networks of public and private firms. To determine the extent of hidden ownership, we identified the offshore holdings attributed to the ultimate owner using unique data sources. The variables for hidden ownership are the dummy for hidden ownership, the degree of hidden ownership, and adjusted ultimate ownership.
• Control variables: We consider several variables found to affect firm performance in the literature. We employed a control set that included the firm’s age (year of registration), size (total assets), leverage (total debt over total equity), board size (number of directors), percentage of independent directors on the management board, and an industry dummy to account for variances across sectors.

3.3. Research Model

To examine the relationship between hidden ownership and firm performance, we adopted a method similar to that of Wiwattanakantang (2001) and Glaewketgarn (2013) using random-effects regressions. Furthermore, we added control variable references from Chen and Guay (2020) and Ashraf et al. (2020). The general form of the model used in this study is as follows:

$$Y_{it}=\alpha +\beta {(hidden\_ownership)}_{it}+\gamma {\mathrm{X}}_{it}+{\epsilon }_{it}$$

where $Y_{it}$ is the dependent variable which represents the firm’s ROA, ROE, and Tobin’s Q performance. For ${\alpha }_0$ = coefficient, i = firm i, $t$ = year, and $\epsilon$ = error term. For the independent variables, ${(hidden\_ownership)}_{it}$ represents three types of hidden ownership depending on the model: the dummy of hidden ownership, the degree of hidden ownership, and the adjusted ultimate ownership. For the control variables, ${\mathrm{X}}_{it}$ is a vector of all firm-level characteristics, including industry dummies. The control variables include firm age, which represents the number of years since the company was founded, converted into a log value. Firm size is the log transformation of a firm’s assets. We measure financial leverage as total debt divided by total equity. The number of directors on the board indicates the size of the board. Independent directors are an indicator of the overall independence of the management board. If a company is involved in a relevant industry, the industry dummy variable equals one.

Table 2. Variables definitions.

Variables	Definition
Dependent variables
Tobin’s Q	The ratio between the market value and replacement value of the same physical asset, calculated by dividing the sum of equity market value and liabilities book value by the sum of equity book value and liabilities book value.
Return on equity (ROE)	The ratio of net income before extraordinary items scaled by book equity, defined as common equity plus deferred tax.
Return on asset (ROA)	The ratio of net income before extraordinary items scaled by total assets.
Independent variables
Dummy of hidden ownership	Equals 1 if the firm has hidden ownership of offshore holdings. Equals 0 otherwise.
Degree of hidden ownership	The number of shares in offshore holdings for hidden ownership over the entire number of firm shares.
Adjusted ultimate ownership	Proportion of ultimate owners plus the fraction of hidden ownership over the entire number of company shares.
Control variables
Firm age	The number of years of the incorporation of the firm.
Firm size	Log transformation of the firm’s assets in Thai baht (THB).
Leverage	Financial leverage proxied by total debt scaled by total equity.
Board size	Total number of directors on board.
Independent directors	Percentage of independent directors on management board.
Industry dummy	Equals 1 if the firm is in the relative industry. Equals 0 otherwise.

provides details on the definition of the variables.

4. Empirical Result

4.1. Descriptive Statistics

A table showing a description of the sample is presented in

Table 3. Descriptive statistics.

Variable	Obs.	Mean	S.D.	Min.	P.5	Median	P.95	Max.
Firms’ performance variable
ROA	136	8.41	5.30	−10.86	−2.20	8.46	17.99	46.81
ROE	136	13.12	10.27	−121.90	−7.28	12.43	35.23	86.16
Tobin’s Q	136	2.40	1.46	0.73	0.91	1.98	6.60	12.02
Ownership variable
Ultimate ownership	136	55.76	16.51	18.22	25.86	57.35	77.29	84.03
Hidden ownership	136	1.81	3.70	0.00	0.00	0.00	12.07	18.95
Firms’ control variable
Firm age	136	21.89	11.97	2.00	6.00	20.00	48.00	53.00
Size	136	6.72	0.45	5.97	6.04	6.69	7.70	8.40
Leverage	136	0.90	0.72	0.03	0.15	0.73	2.58	17.51
Board size	136	9.54	1.61	7.00	7.00	9.00	12.00	15.00
Independent directors	136	3.87	0.76	2.00	3.00	4.00	5.00	7.00

Note: The number of observations is reported in “firm-year” format. This study includes 136 firm-years. Firms with hidden ownership account for 52 firm-years, whereas those without hidden ownership account for 84 firm-years. The firm’s age is shown in years for the firms’ control variables. Firm size is the log of a firm’s assets in the THB. Leverage is the ratio of total debt to total equity. Board size is the total number of board members. Independent directors is the number of independent directors on the management board.

. The table includes information on the average ROA, ROE, and Tobin’s Q, which are 8.41, 13.12, and 2.40, respectively, with standard deviations of 5.30, 10.27, and 1.46, respectively. Additionally, the table provides firm characteristics, such as ownership fraction, firm age, asset size, degree of leverage, number of board members, and independent directors. The table shows that the ultimate ownership proportion is 55.76% with a standard deviation of 16.51%. The proportion of hidden ownership proportion is 1.81% with a standard deviation of 3.70%. The firm’s age is 23.44 years with a standard deviation of 11.25 years. The firm size is represented by the logarithm of its assets, which is 6.81 with a standard deviation of 0.33. The average leverage of the firm is 0.78 with a standard deviation of 0.61. The average size of the board of directors is 9.69 members with a standard deviation of 1.78 members. The number of independent directors is 4.10 members with a standard deviation of 0.82.

The sample description, as presented in Table 3, suggests that the study encompasses 136 firm-years. The identification of offshore holdings associated with ultimate owners was accomplished through the utilization of unique data sources. Firms with hidden ownership account for 52 firm-years, while those without hidden ownership account for 84 firm-years. After considering hidden ownership and adjusting it to the final ultimate ownership share, the overall ownership percentages of both groups were found to be comparable. Furthermore, the study revealed that the characteristics of firms, including variables such as ownership, firm age, asset size, degree of leverage, number of board members, and independent directors, were comparable between the two groups.

4.2. Regression Results and Hypothesis Testing

This section discusses the regression findings regarding the impact of hidden ownership on firm performance. We expand the main regression models for the regression analysis into several models that incorporate accounting and market performance metrics. Several types of hidden ownership that may influence a company’s performance have been investigated. Each hidden ownership pattern was regressed independently using three distinct equation models (models 1–9); the findings were compared, and the results of each model were explained. Following the test, we anticipated a significant result with a negative correlation due to the agency problem, which is wealth expropriation from minority shareholders to the ultimate owner, thereby generating a negative association between hidden ownership and firm performance.

Table 4. The effect of hidden ownership and firm performance (ROA).

Variables	Model 1	Model 2	Model 3
Hidden ownership dummy	1.995 * (1.868)
Degree of hidden ownership		0.268 ** (2.012)
Adjusted ultimate ownership			−0.016 (−0.539)
Firm age	−0.020 (−0.484)	−0.037 (−0.901)	−0.017 (−0.410)
Firm size	0.180 (0.137)	0.255 (0.195)	0.534 (0.404)
Leverage	−2.548 *** (−3.442)	−2.489 *** (−3.354)	−2.765 *** (−3.711)
Board size	−0.317 (−0.914)	−0.431 (−1.249)	−0.392 (−1.123)
Independent directors	−0.390 (−0.507)	−0.245 (−0.326)	−0.019 (−0.025)
Industry dummy (Agro and Food)	−0.597 (−0.440)	0.267 (0.200)	0.135 (0.096)
Industry dummy (Consumer Products)	0.003 (0.001)	−0.097 (−0.042)	−0.908 (−0.391)
Industry dummy (Industrials)	−1.392 (−0.768)	−1.496 (−0.834)	−1.891 (−1.030)
Industry dummy (Property and Construction)	0.652 (0.494)	0.849 (0.637)	0.515 (0.367)
Industry dummy (Resources)	−2.077 (−1.314)	−1.903 (−1.194)	−2.625 (−1.639)
Constant	14.050 (1.628)	14.449 * (1.675)	12.832 (1.465)
Observations	136	136	136
Adjusted R2	0.124	0.128	0.102

Notes: The dependent variable is ROA. ***, **, and * indicate significances at the 1%, 5%, and 10% levels, respectively.

displays the results of the regression analysis for the three tested models. The predictors for return on assets (ROA) include hidden ownership dummy, degree of hidden ownership, adjusted ultimate ownership, and other control variables such as firm age, firm size, leverage, board size, number of independent directors, and industry dummy. The first model indicates that the hidden ownership dummy has a positive and significant relationship with ROA at the 90% confidence level, with a coefficient of 1.995 and a p-value of 0.064. The second model shows that the degree of hidden ownership is positively associated with ROA at a 95% confidence level, with a coefficient of 0.268 and a p-value of 0.046. In contrast, the third model reveals that adjusted ultimate ownership has no significant relationship with ROA. Additionally, the study finds that the leverage variable is strongly and negatively associated with ROA across all three models, with a 99% confidence level. The other control variables were found to be insignificant.

After examining

Table 5. The effect of hidden ownership and firm performance (ROE).

Variables	Model 4	Model 5	Model 6
Hidden ownership dummy	3.550 (1.611)
Degree of hidden ownership		0.469 * (1.708)
Adjusted ultimate ownership			−0.036 (−0.582)
Firm age	−0.020 (−0.235)	−0.050 (−0.593)	−0.013 (−0.156)
Firm size	−0.529 (−0.194)	−0.388 (−0.143)	0.099 (0.036)
Leverage	0.095 (0.062)	0.193 (0.126)	−0.306 (−0.200)
Board size	−0.783 (−1.093)	−0.983 (−1.382)	−0.915 (−1.275)
Independent directors	−1.363 (−0.858)	−1.100 (−0.708)	−0.692 (−0.444)
Industry dummy (Agro and Food)	−2.812 (−1.004)	−1.284 (−0.465)	−1.417 (−0.492)
Industry dummy (Consumer Products)	1.268 (0.263)	1.068 (0.223)	−0.364 (−0.076)
Industry dummy (Industrials)	−3.418 (−0.914)	−3.619 (−0.977)	−4.213 (−1.116)
Industry dummy (Property and Construction)	0.599 (0.220)	0.933 (0.339)	0.469 (0.163)
Industry dummy (Resources)	−3.243 (−0.994)	−2.959 (−0.899)	−4.118 (−1.251)
Constant	29.468 (1.655)	30.117 * (1.691)	27.582 (1.532)
Observations	136	136	136
Adjusted R2	0.007	0.009	0.006

Notes: The dependent variable is ROE. * indicate significances at the 10% levels.

, Model 5 demonstrates a positive and significant correlation between the degree of hidden ownership and ROE, with a coefficient of 0.469 and a p-value of 0.090 at a 90% confidence level. However, this association is not observed in Models 4 and 6, as neither the hidden ownership dummy nor the adjusted ultimate ownership variable appears to impact ROE. Additionally, none of the three models indicate that the other control variables have a significant effect on ROE. Overall, the regression results suggest that the degree of hidden ownership is a crucial predictor of ROE, while other variables may not play a significant role.

Table 6. The effect of hidden ownership and firm performance (Tobin’s Q).

Variables	Model 7	Model 8	Model 9
Hidden ownership dummy	0.580 ** (2.208)
Degree of hidden ownership		0.046 (1.390)
Adjusted ultimate ownership			−0.007 (−0.893)
Firm age	−0.024 ** (−2.424)	−0.027 *** (−2.664)	−0.023 ** (−2.229)
Firm size	−0.748 ** (−2.305)	−0.693 ** (−2.121)	−0.646 * (1.979)
Leverage	−0.615 *** (−3.377)	−0.626 *** (−3.387)	−0.681 *** (−3.707)
Board size	0.021 (0.242)	−0.008 (−0.090)	−0.001 (−0.010)
Independent directors	−0.279 (−1.472)	−0.213 (−1.133)	−0.168 (−0.897)
Industry dummy (Agro and Food)	0.222 (0.664)	0.432 (1.297)	0.459 (1.329)
Industry dummy (Consumer Products)	0.297 (0.516)	0.174 (0.302)	0.029 (0.051)
Industry dummy (Industrials)	−1.695 *** (−3.806)	−1.797 *** (−4.021)	−1.815 *** (−4.007)
Industry dummy (Property and Construction)	−0.586 * (−1.807)	−0.599 * (−1.803)	−0.596 * (−1.721)
Industry dummy (Resources)	−0.462 (−1.189)	−0.524 (−1.319)	−0.594 (−1.504)
Constant	9.472 *** (4.465)	9.321 *** (4.335)	9.193 *** (4.254)
Observations	136	136	136
Adjusted R2	0.298	0.282	0.275

Notes: The dependent variable is Tobin’s Q. ***, **, and * indicate significances at the 1%, 5%, and 10% levels, respectively.

presents the regression results for the three tested models. In Model 7, the hidden ownership dummy is positively and significantly associated with Tobin’s Q (at a 95% confidence level), with a coefficient of 0.580 and a p-value of 0.029. However, neither the degree of hidden ownership nor the adjusted ultimate ownership variable is related to Tobin’s Q. Additionally, Tobin’s Q is substantially and adversely correlated with firm age, firm size, leverage, and several industry dummy factors, as observed in the control variables.

The results of Table 4, Table 5 and Table 6 demonstrate a strong positive correlation between hidden ownership and overall firm performance. At the 0.05% and 0.10% significance levels, the hidden ownership dummy was significantly and positively correlated with ROA and Tobin’s Q, which are two measures of firm performance. Additionally, ROA and ROE, two accounting measures of firm performance, were strongly and positively correlated with the degree of hidden ownership at the same significance levels. These results are consistent with the direction and sign of the coefficient, as every significance test indicated a positive correlation. However, no significant relationship was found between ROA, ROE, Tobin’s Q, and adjusted ultimate ownership, including hidden ownership. In summary, this study’s findings contradict the hypothesis that hidden ownership may result in the transfer of wealth from minority shareholders to the ultimate owner. Instead, the research supports the idea that companies with hidden ownership exhibit better performance compared to those without. These findings challenge conventional beliefs about hidden ownership structures and their impact on firm performance, contributing to a deeper understanding of ownership control patterns.

4.3. Robustness Analysis

To establish the validity of our findings, we employed a two-stage least squares regression model while accounting for the potential influence of endogeneity on our data. Given that both hidden ownership and firm performance were determined to be endogenous, we utilized the average ownership proportion over time, the change in ownership proportion, and the proportion of foreign institutional ownership as instrument variables in the first stage of regression. In the second stage, we re-estimated the regression using the fitted values from the first stage instead of the hidden ownership variables, resulting in accurate and dependable results.

Table 7. Endogeneity controls using 2SLS estimations.

Variables	Hidden Ownership	ROA	ROA	ROE	ROE	Tobin’s Q	Tobin’s Q
	1st Stage	2nd Stage	2nd Stage	2nd Stage	2nd Stage	2nd Stage	2nd Stage
Hidden ownership dummy		2.387 ** (2.188)		4.430 * (1.968)		0.728 *** (2.727)
Degree of hidden ownership			0.272 * (1.848)		0.433 (1.427)		0.038 (1.040)
Firm age	−0.046 (−1.105)	−0.021 (−0.520)	−0.037 (−0.901)	−0.022 (−0.266)	−0.048 (−0.563)	−0.024 ** (−2.488)	−0.027 ** (−2.598)
Firm size	−2.216 (−1.161)	0.075 (0.057)	0.251 (0.191)	−0.751 (−0.276)	−0.350 (−0.129)	−0.785 ** (−2.438)	−0.685 ** (−2.085)
Leverage	−1.320 (−1.387)	−2.405 *** (−3.227)	−2.485 *** (−3.331)	0.374 (0.243)	0.160 (0.104)	−0.568 *** (−3.117)	−0.634 *** (−3.405)
Board size	0.557 ** (2.125)	−0.316 (−0.916)	−0.431 (−1.248)	−0.775 (−1.089)	−0.978 (−1.369)	0.022 (0.263)	−0.007 (−0.077)
Independent directors	−0.639 (−1.154)	−0.429 (−0.561)	−0.248 (−0.329)	−1.461 (−0.927)	−1.072 (−0.687)	−0.295 (−1.580)	−0.207 (−1.096)
Industry dummy (Agro and Food)	−1.117 (−1.080)	−0.838 (−0.613)	0.272 (0.202)	−3.298 (−1.169)	−1.330 (−0.479)	0.141 (0.421)	0.422 (1.260)
Industry dummy (Consumer Products)	−0.807 (−0.218)	−0.009 (−0.004)	−0.0.85 (−0.037)	1.312 (0.275)	0.962 (0.200)	0.306 (0.540)	0.151 (0.259)
Industry dummy (Industrials)	−1.596 (−0.646)	−1.371 (−0.764)	−1.487 (−0.825)	−3.325 (−0.898)	−3.700 (−0.993)	−1.678 *** (−3.824)	−1.815 *** (−4.036)
Industry dummy (Property and Construction)	−0.069 (−0.053)	−0.599 (0.458)	0.857 (0.638)	0.530 (0.197)	0.854 (0.308)	−0.597 * (−1.868)	−0.617 * (−1.840)
Industry dummy (Resources)	2.549 (1.447)	−2.071 (−1.328)	−1.889 (−1.171)	−3.172 (−0.986)	−3.086 (−0.926)	−0.449 (1.177)	−0.552 (−1.373)
Average ownership proportion across time	0.071 *** (2.817)
change in ownership proportion	4.473 ** (2.583)
foreign institutional ownership proportion	0.663 *** (8.618)
Constant	6.833 (0.606)	14.753 * (1.714)	14.482 * (1.672)	30.913 * (1.741)	29.814 * (1.665)	9.713 *** (4.612)	9.254 *** (4.282)
Observations	136	136	136	136	136	136	136
Adjusted R2	0.723	0.133	0.124	0.017	0.002	0.312	0.277

Notes: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively. The degree of hidden ownership and the hidden ownership dummy are the dependent variables in the first stage of regression.

and

Table 8. Endogeneity controls using 2SLS estimations.

Variables	Adjusted Ultimate Ownership	ROA	ROE	Tobin’s Q
	1st Stage	2nd Stage	2nd Stage	2nd Stage
Adjusted ultimate ownership		−0.026 (−0.862)	−0.058 (−0.919)	−0.011 (−1.402)
Firm age	0.007 (0.249)	−0.014 (−0.348)	−0.008 (−0.092)	−0.022 ** (−2.140)
Firm size	−1.679 * (−1.865)	0.532 (0.403)	0.094 (0.035)	−0.647 ** (−1.991)
Leverage	−0.030 (−0.061)	−2.784 *** (−3.743)	−0.347 (−0.227)	−0.689 *** (−3.765)
Board size	0.142 (0.624)	−0.392 (−1.123)	−0.914 (−1.276)	−0.001 (−0.008)
Independent directors	−0.249 (−0.482)	−0.004 (−0.005)	−0.659 (−0.423)	−0.162 (−0.870)
Industry dummy (Agro and Food)	0.349 (0.378)	0.268 (0.191)	−1.131 (−0.393)	0.511 (1.482)
Industry dummy (Consumer Products)	2.064 (1.337)	−0.923 (−0.398)	−0.396 (−0.083)	0.023 (0.041)
Industry dummy (Industrials)	0.723 (0.599)	−1.757 (−0.957)	−3.925 (−1.041)	−1.763 *** (−3.906)
Industry dummy (Property and Construction)	1.465 (1.543)	0.678 (0.483)	0.818 (0.284)	−0.533 (−1.542)
Industry dummy (Resources)	1.492 (1.333)	−2.480 (−1.548)	−3.807 (−1.157)	−0.538 (−1.367)
Average ownership proportion across time	0.963 *** (49.024)
change in ownership proportion	45.065 *** (32.037)
foreign institutional ownership proportion	0.635 *** (8.828)
Constant	−33.527 *** (−5.575)	13.237 (1.513)	28.452 (1.582)	9.350 *** (4.344)
Observations	136	136	136	136
Adjusted R2	0.860	0.105	0.005	0.282

Notes: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively. The adjusted ultimate ownership is the dependent variables in the first stage of regression.

show the results of the two-stage least squares (2SLS) regression estimates. The analysis reveals that all instrumental variables, including the average ownership percentage over time, the change in ownership proportion, and the foreign institutional ownership percentage, have statistically significant coefficients. The study also examines the instrumental variable of hidden ownership and finds that the estimated values of hidden ownership in the subsequent regression analysis yield a positive and statistically significant coefficient, indicating consistency with the main findings in Table 4, Table 5 and Table 6. Therefore, there is confidence that endogeneity does not influence the main regression results in Table 4, Table 5 and Table 6.

4.4. Additional Analysis of Ownership Structure

The results of the univariate analysis are presented in

Table 9. The proportion of foreign investors’ holdings after excluding hidden ownership and ultimate ownership portion.

	Firm with Hidden Ownership		Firm without Hidden Ownership
Investors Type	Mean	S.D.	Mean	S.D.
Individual	0.16	0.45	0.33	0.94
Institutional	2.35	3.58	0.87	1.88
Total	2.52	3.52	1.19	2.43

Note: The number of observations is reported in “firm-year” format. This study includes 136 firm-years. Firms with hidden ownership account for 52 firm-years, whereas those without hidden ownership account for 84 firm-years.

. The percentage of foreign shareholders in firms with hidden ownership is relatively high, and they are likely institutional investors. Institutional investors may promote improved governance, proactive monitoring, and improved corporate performance (Shleifer and Vishny 1986; Pedersen and Thomsen 2003; Cornett et al. 2007; Elyasiani and Jia 2010).

The percentages of foreign investors are listed in Table 9. After excluding hidden ownership and ultimate ownership, the average proportion of foreign investors in firms with hidden ownership was determined to be 2.52% with a standard deviation of 3.52%. Foreign institutional investors account for the largest percentage (2.35%, with a standard deviation of 3.58%). In contrast, the average proportion of foreign investors in firms without hidden ownership is 1.19% with a standard deviation of 2.43%. Foreign institutional investors account for 0.87% with a standard deviation of 1.88%. The proportion of institutional investors in firms without hidden ownership is considerably lower than that in firms with hidden ownership. The variations in the number and type of foreign investors may substantially influence a firm’s governance structure. According to several studies (Pedersen and Thomsen 2003; Cornett et al. 2007; Elyasiani and Jia 2010), institutional investors positively affect business performance. Thus, institutional investors with enhanced monitoring skills might explain this effect since they can scrutinize management more thoroughly than individual shareholders and do so at a relatively low cost (Shleifer and Vishny 1986).

Nevertheless, we are aware that the causality effect begins with good-performance firms that attract well-connected FAs, whose clients include institutional investors, and who suggest offshore ownership for transaction purposes. As shown in

Table 10. Firm’s performance one year before its first public offering.

	Firm with Hidden Ownership		Firm without Hidden Ownership
Firm Performance	Mean	S.D.	Mean	S.D.
ROA	14.47	10.76	10.70	9.01
ROE	34.89	21.90	24.47	15.78

Note: The number of observations is reported in “firm-year” format. This table shows a firm’s performance one year before its IPO, which includes 38 firm-year observations. Firms with hidden ownership account for 15 firm-years, whereas firms without hidden ownership account for 23 firm-years.

, we compare the performance of hidden and non-hidden firms to examine a firm’s performance before its IPO.

Table 10 shows a firm’s performance one year before its IPO. The firms with hidden ownership have average ROA and ROE values of 14.47 and 34.89, respectively, with standard deviations of 10.76 and 21.90, respectively. Conversely, firms without hidden ownership have average ROA and ROE values of 10.70 and 24.47, respectively, with standard deviations of 9.01 and 15.78, respectively. Table 10 shows that firms with hidden ownership perform better than firms without hidden ownership.

5. Conclusions

This study investigates the chain of ownership of a company’s shares using special data sources to uncover hidden ownership patterns. As hidden ownership may negatively affect a firm’s performance, studying the link between hidden ownership and business performance is essential. The results of this study dispel the traditional hypothesis of wealth expropriation from minority shareholders to the ultimate owner. Furthermore, the analysis shows that firms with hidden ownership outperform those without hidden ownership. The existence of hidden ownership is not detrimental to firm value; in contrast, the presence of hidden ownership might enhance firm value. Furthermore, the findings exhibit robustness following tests that account for potential endogeneity among the variables, hence affirming their alignment with the main regression result.

Additionally, the percentage of foreign shareholders in firms with hidden ownership is relatively high, and foreign shareholders are likely institutional investors. Institutional investors may improve governance, proactive monitoring, and corporate performance (Shleifer and Vishny 1986; Pedersen and Thomsen 2003; Cornett et al. 2007; Elyasiani and Jia 2010). Furthermore, firms with hidden ownership outperform those without hidden ownership, although this may not be solely due to the high number of institutional investors. These firms had already achieved high returns before their IPO. Strong firm performance attracts reputable FAs with strong connections to institutional investors. Strong connections between reputable FAs and a large number of institutional investors can further influence high and continuously high firm performance, suggesting that firms strategically implement offshore hidden ownership for transaction purposes.

This study seeks to make a significant academic contribution by utilizing comprehensive ownership structure data to shed light on the mysterious ownership puzzles. Through this, it will contribute to the existing body of knowledge in the field of corporate governance and establish itself as a groundbreaking study in this area. The findings of this study contradict the hypothesis that hidden ownership results in the transfer of wealth from minority shareholders to the ultimate owner. In fact, the analysis shows that firms with hidden ownership outperform firms without hidden ownership since the ultimate owner strategically employs offshore hidden ownership for transactional objectives. This study challenges traditional assumptions regarding hidden ownership arrangements and corporate performance, thereby improving our understanding of ownership control patterns.

Given the findings of this research, it is crucial for policymakers and regulators to thoroughly review current regulations and standards related to disclosure and transparency mandates. This review is necessary to address the issue of information asymmetry between majority and minority shareholders. The study also provides new avenues for future investigations into the motivations behind hidden ownership, such as tax optimization, adaptability in buying and selling, and manipulation of stock prices. The potential benefit of tax management is that it serves as an additional incentive for hidden ownership. The difference in tax structures motivates ultimate owners to choose hidden ownership through offshore entities rather than onshore entities. A future investigation of the capital tax system could explain the reasons behind ultimate owners’ preference for masking ownership through foreign offshore entities. Additionally, the ultimate owner intentionally uses hidden offshore ownership to achieve transactional goals. Flexibility in stock transactions and the ability to manipulate prices may be potential motivations for hiding ownership. The ultimate owner may use this flexibility to bypass regulatory restrictions and disclosure requirements or to prevent the transmission of inaccurate market signals when trading in equities.

This study, like others, has some limitations. First, it focused on IPO firms in Thailand with a limited scope. Although the results may be valid specifically for Thailand, they may not apply to other countries with different ownership structures. Therefore, caution should be exercised when generalizing the findings of this study. Second, the study was restricted to data collected between 2011 and 2015 because the unique database only provided access to this period. Extending the study to include more recent data would help improve the consistency of results across time. Finally, the focus of this research is on IPO firms. Private firms that go public often face high degrees of information asymmetry. However, extending this study to include currently listed firms would be interesting and may provide a better understanding of ownership control patterns.