The Effect of IFRS Adoption on Foreign Investment in the Japanese Equity Market

Abstract

This study investigates the effects of International Financial Reporting Standards (IFRS) adoption on foreign investment in the Japanese equity market. Previous research suggests that a positive relationship between IFRS adoption and foreign investment typically emerges when a country meets specific conditions, such as a strong regulatory environment or credible improvements in reporting uniformity. We contribute to this literature by re-examining the effects of voluntary IFRS adoption on the foreign shareholding ratio of Japanese companies from 2010 to 2023. Our analysis explicitly controls for the impact of reduced cross-shareholdings and increased share buybacks—structural factors likely to affect the capacity for foreign ownership. Using a difference-in-differences approach combined with propensity score matching to mitigate endogeneity, we compare 168 voluntary IFRS adopters against a control group of non-adopters. Unlike previous studies that reported no significant relationship in Japan—largely attributable to different denominator constructions for foreign shareholding ratios or shorter observation periods—our approach demonstrates that IFRS adoption significantly increases foreign investment over an extended horizon.

1. Introduction

The purpose of this study is to re-examine the effects of the voluntary adoption of International Financial Reporting Standards (IFRS) on the foreign shareholder ratio of Japanese companies. Since its mandatory adoption in the European Union (EU) in 2005, IFRS have been widely used in many countries. IFRS aim to enhance the comparability of firms’ financial statements in different countries and reduce international investors’ information acquisition and processing costs, thereby leading to more significant cross-border investments. In 2010, Japan allowed listed companies to adopt IFRS voluntarily. According to the Japan Exchange Group (2025b), 295 companies have adopted IFRS as of March 2026, and 16 have made plans to do so. Although IFRS adopters account for less than 10% of all firms listed on the Tokyo Stock Exchange, their market capitalization accounts for approximately half that of all firms, suggesting that large companies tend to adopt IFRS.

Because accounting standards are part of the institutional infrastructure, IFRS adoption is expected to promote both foreign direct investment (FDI) and foreign portfolio investment (FPI) by enhancing the comparability of financial statements across nations and reducing information asymmetry between local and foreign investors.1 This argument is supported theoretically by Easley and O’Hara (2004), who present a model demonstrating that detailed accounting information reduces a company’s cost of capital.

However, previous studies indicate that the quality of accounting information is determined not only by reporting standards but also by market forces and institutional factors (Ball et al., 2000, 2003; Ball & Shivakumar, 2005; Leuz et al., 2003). Consistent with this perspective, several studies find that the positive relationship between IFRS adoption and foreign investment either exists or becomes more pronounced when a country meets specific conditions, such as a strong regulatory environment, significant pre-convergence differences in accounting systems, or credible improvements in uniformity following IFRS adoption (Chen et al., 2014; M. DeFond et al., 2011; Shima & Gordon, 2011).

Accordingly, single-country and regional studies yielded mixed results. While a positive relationship between IFRS adoption and foreign investment exists in Sweden and Taiwan (Hamberg et al., 2013; Hsu et al., 2015; Saravanan & Firoz, 2026), a negative association exists in China and several African countries because of their weak institutional infrastructure (M. DeFond et al., 2019; Nnadi & Soobaroyen, 2015).

Japan is classified as a code-law (or civil-law) country (Ball et al., 2000; La Porta et al., 1998). Compared with common-law countries, code-law countries generally offer weaker legal protection to investors and less robust enforcement mechanisms. These characteristics suggest that Japan’s institutional infrastructure may not be sufficiently robust. Consistent with this perspective, J. Kim et al. (2019) report that the voluntary IFRS adoption did not significantly alter the foreign shareholder ratio in Japanese companies between 2010 and 2018.

As is typical in code-law countries, Japanese firms tended to adopt a “stakeholder governance” system, in contrast to the “shareholder governance” system prevalent in common-law countries. The stakeholder model is characterized by a reliance on debt financing, shareholding by domestically affiliated interests, and interconnected networks among firms, their trading partners, and main banks (Hoshi & Kashyap, 2001; Shleifer & Vishny, 1997). However, following the collapse of the economic bubble in the 1990s, Japan made extensive efforts to reform its economy and business environment. These changes are evident in the decline in traditional cross-shareholdings2 and the rise in foreign shareholdings.

Although the decline in cross-shareholdings opened doors for foreign investors by the mid-2000s, foreign ownership levels plateaued at roughly 30% over the subsequent decade.3 The current literature suggests that this reduction in cross-shareholding has been replaced either by an increase in share buybacks or by strategic value-enhancing inter-corporate holdings (Franks et al., 2023; Miyajima & Saito, 2023). This uptick in buybacks is driven, in part, by corporate governance reforms accelerated by the introduction of the Stewardship Code (SWC) in 2014 and the Corporate Governance Code (CGC) in 2015. As these codes called for institutional investors to take a more active role in monitoring their investee companies, opportunities increased for activist shareholders—who often advocated for share buybacks to boost dividends, divestiture of unprofitable businesses, or shifts in corporate strategy (Miyachi & Takeda, 2021). Consequently, the shareholder ratio of corporations has not significantly declined, and the foreign shareholder ratio may not have increased.

These developments provide two distinct motivations for the current study. First, by extending the period of analysis, we expect to generate results that differ from the previous literature, as recent corporate governance reforms may have established the institutional framework required for IFRS adoption to exert a measurable impact. Second, we posit that explicitly controlling for the dual influences of declining cross-shareholdings and rising share buybacks will yield findings distinct from earlier research. This expectation is based on the approach taken by J. Kim et al. (2019), who construct the foreign shareholding ratio by subtracting buybacks from the denominator. Given the upward trend in buybacks, their adjustment may lead to an underestimation of the perceived foreign shareholding ratio.

Thus, the research question of this study is whether voluntary IFRS adoption increases the foreign shareholder ratio of Japanese firms, while specifically controlling for the confounding effects of cross-shareholding reductions and share buyback activity over a longitudinal period. Our sample consists of 168 listed firms that applied IFRS to their financial statements by fiscal year 2022. We employ the difference-in-differences (DID) approach to compare the set of voluntary IFRS adopters to non-adopters from 2010 to 2023. We select non-adopters using propensity score matching (PSM) to avoid potential endogeneity. By employing a foreign shareholder ratio constructed differently from J. Kim et al. (2019), our analysis reveals an increase in foreign shareholdings following IFRS adoption when explicitly controlling for cross-shareholdings and share buybacks over an extended period.4

This study contributes to the literature on the effects of IFRS adoption on foreign equity investment in two ways. First, it provides evidence from Japan, which underwent a governance system transition concurrent with its adoption of IFRS. Second, this study demonstrates the importance of controlling confounding factors in the context of multiple institutional changes. We believe that the findings will be valuable to international investors considering investing in countries such as Japan, a non-Western nation with robust legal enforcement. Our findings are also useful for firms considering IFRS adoption and anticipating changes in their shareholder composition.

The remainder of this article is organized as follows. Section 2 provides a literature review and develops the hypothesis. Section 3 outlines the methodology and data. Section 4 presents the empirical results. Finally, we give our concluding remarks in Section 5.

2. Literature Review and Hypothesis

2.1. The Impact of the Adoption of Non-Local Accounting Standards

IFRS adoption is expected to improve the comparability of firms’ financial statements across countries, reduce costs of information acquisition and processing for international investors, and facilitate cross-border investments. Easley and O’Hara (2004) provide theoretical support for this argument, presenting a model demonstrating that detailed accounting information reduces a company’s cost of capital. However, empirical studies provide mixed results. For instance, while IFRS adoption is found to have reduced the cost of capital in European countries (Ghouma et al., 2023), IFRS convergence in India led to increases in the cost of equity, cost of debt, and information asymmetry, alongside a decrease in market liquidity (Bansal, 2023).5 Accordingly, a positive relationship between IFRS adoption and foreign investment is reported only when a country meets specific institutional conditions.

Prior to the introduction of IFRS, several studies examined the impact of the voluntary adoption of International Accounting Standards (IAS) on foreign investment. Using data on more than 25,000 mutual funds worldwide between 1999 and 2002, Covrig et al. (2007) find that foreign mutual fund ownership is higher among voluntary IAS adopters. Additionally, other studies indicate that voluntary IAS adoption reduces information asymmetry by increasing analyst coverage, forecast accuracy, and share turnover, while decreasing bid–ask spreads (Ashbaugh & Pincus, 2001; Cuijpers & Buijink, 2005; Leuz & Verrecchia, 2000).

Although IAS adoption is voluntary, many countries followed the European Union’s lead in 2005 by mandating IFRS adoption. Similar to previous studies on the impact of IAS adoption, several empirical studies using multi-country data report a positive correlation between IFRS adoption and foreign investment.6 For instance, using data on 73 countries between 2001 and 2006, Amiram (2012) finds that foreign equity portfolio investments increase in countries that adopt IFRS. Mita et al. (2018) also show that foreign investor ownership of firms in 18 countries is positively associated with IFRS adoption for 2003–2012 through the mechanism of higher financial statement comparability.

However, other studies report that reporting standards alone do not determine accounting information quality, but rather market forces and institutional factors also have an effect (Ball et al., 2000, 2003; Ball & Shivakumar, 2005; Leuz et al., 2003). Consistent with this view, several studies find that this positive relationship strengthens when a country meets specific conditions. For instance, based on data from 44 countries between 2003 and 2006, Shima and Gordon (2011) show that US investments in foreign equities are positively associated with IFRS only when the country has a strong regulatory environment. Using data on firms in 14 EU countries for 2003–2007, M. DeFond et al. (2011) find that foreign mutual fund ownership increases in countries with credible increases in uniformity following IFRS adoption.7 Chen et al. (2014) also report that IFRS conformity facilitated FDI flows in 30 OECD countries between 2000 and 2005.

2.2. Single-Country/Region Studies on the Impact of IFRS Adoption

Consistent with the view that a robust institutional infrastructure strengthens the positive relationship between IFRS adoption and foreign investment, existing single-country and regional studies have yielded mixed results. Hamberg et al. (2013) focus on Sweden and report a positive correlation between IFRS adoption and foreign ownership for 2001–2007. Using data on Taiwanese firms, Hsu et al. (2015) report that foreign investors increased their shareholdings following the adoption of IAS-27. Saravanan and Firoz (2026) find a significant increase in foreign institutional investors’ ownership in IFRS-compliant firms in India following implementation during the 2011–2017 period.

However, studies using other datasets have not identified this relationship. Nnadi and Soobaroyen (2015) find that full IFRS adoption is negatively associated with net FDI in 34 African countries. M. DeFond et al. (2019) document that foreign institutional investment did not increase after China adopted the IFRS for 2005–2008. Based on the evidence that foreign institutional investment declines among firms with weaker incentives to credibly implement IFRS or a stronger ability to manipulate IFRS fair value provisions, they conclude that China fails to attract foreign investors, because its weak institutional infrastructure does not improve financial reporting quality. Utilizing data from the Gulf Cooperation Council (GCC) countries between 1980 and 2017, Mameche and Masood (2021) report mixed results, observing a short-term increase in FDI followed by a long-term decrease after IFRS adoption.

Institutionally, Japan is classified as a code law (or civil law) country, similar to China (Ball et al., 2000; La Porta et al., 1998). Compared with common law countries, code law countries generally offer weaker legal protection to investors and less robust enforcement mechanisms. These characteristics suggest that Japan’s institutional infrastructure is not sufficiently robust. Consistent with this view, J. Kim et al. (2019) report that the voluntary adoption of IFRS neither significantly changed analyst coverage nor the foreign shareholders’ ratio of Japanese companies during 2010–2018 but significantly decreased their cross-shareholder ratio. Their results indicate that voluntary IFRS adopters reduce barriers to dialogue with investors but fail to improve the information environment, resulting in unchanged foreign shareholder ratios in Japan.

Other studies have shown that Japan’s institutional infrastructure may not be sufficiently strong. J. H. Kim et al. (2024) and Tohara (2024) find that voluntary IFRS adoption increased discretionary accruals in Japan by 2018, indicating earnings management. In addition, Kashiwazaki et al. (2019) and Amano (2020) discover a change in IFRS adopters’ real actions by taking advantage of the difference in goodwill amortization between IFRS and Japanese Generally Accepted Accounting Principles (JGAAP). Kashiwazaki et al. (2019) show that IFRS adopters experience a larger increase in the number of mergers and acquisitions (M&A), which consist mainly of domestic firms’ purchases of domestic firms.8 Similarly, Amano (2020) reports that voluntary IFRS adopters tend to increase their intangible assets.

2.3. Hypothesis Development

While J. Kim et al. (2019) report no positive relationship between IFRS adoption and the ratio of foreign shareholders, this finding may be attributable to their construction of the ratio, which deducts share buybacks from the denominator. Japan’s corporate governance reform was accelerated by the introduction of the SWC in 2014 and the CGC in 2015. By shifting Japanese corporate governance toward shareholder-oriented models, these recent reforms may have established the institutional framework required for IFRS adoption to exert a significant impact. The SWC and CGC required listed firms to reduce cross-shareholdings, increasing the influence of institutional investors and activists focused on shareholder returns (Miyachi & Takeda, 2021). This reduction has, in turn, led to an increase in share buybacks (Franks et al., 2023; Miyajima & Saito, 2023). Consequently, when the foreign shareholder ratio is constructed by subtracting share buybacks from the denominator, increased buybacks negatively affect the resulting ratio.

To address these issues, this study reexamines the impact of voluntary IFRS adoption on the foreign shareholder ratios of Japanese companies within a setting distinct from prior research. First, we extend the period of analysis, as governance reforms introduced in the mid-2010s likely strengthened shareholder oversight and enforcement mechanisms in the subsequent years. Second, we explicitly control for the influence of declining cross-shareholdings and rising share buybacks, both of which may otherwise offset the growth of foreign shareholding ratios. Our hypothesis is as follows:

Hypothesis 1:

The voluntary adoption of IFRS increases the foreign shareholders’ ratio in Japanese listed companies after controlling for share buybacks and cross-shareholdings.

3. Methodology and Data

3.1. PSM

To investigate the impact of voluntary IFRS adoption on the foreign shareholder ratio of Japanese listed companies, we employ the DID approach (Roberts & Whited, 2013), which uses simple panel data methods applied to sets of voluntary IFRS adopters and non-adopters for 2010–2023, the period after voluntary IFRS adoption was allowed in Japan. Our sample consists of 168 listed firms with available data that report financial statements based on the IFRS by fiscal year 2022. Non-adopters are selected using PSM (Rosenbaum & Rubin, 1983) to avoid potential endogeneity problems. Following previous studies on the factors affecting voluntary IFRS adoption by Japanese companies, we estimate the following probit model (J. Kim et al., 2019; Sato & Takeda, 2017):

$$\begin{array}{l}{P_r\left(IFRS=1\right)}_i=F(\alpha +{\beta }_1{FOREIGNHELD}_{it}+{\beta }_2{GW}_{it}+{\beta }_3{RD}_{it}{+\beta }_4{CROSS}_{it}\\ +{\beta }_5{FOREIGNSALES}_{it}+{\beta }_6{SIZE}_{it}+{{\beta }_{7}ROA}_{it}+{{\beta }_{8}BTM}_{it}\\ {+\beta }_9{CLOSELYHELD}_{it}{+\beta }_{10}{DEBT}_{it}+{\beta }_{11}{AGE}_{it}+{\beta }_{12}{(IRD1/IRD2)}_t\\ +{\beta }_{13}{INDUSTRY}_i)\end{array}$$

(1)

The definitions of the variables and data sources are summarized in Appendix B. The dependent variable, IFRS, is an indicator variable for firms that voluntarily adopted IFRS between 2010 and 2023. We define this binary variable as unity for IFRS adopters—including both new and existing adopters during the sample period—and zero for non-adopters. FOREIGN HELD is constructed as the ratio of shares owned by foreign shareholders to total shares. As explained above, we do not subtract share buybacks from the denominator; instead, we treat share buybacks as an independent variable. Based on our hypothesis, we expect positive coefficients for this variable.

GW is the goodwill ratio of the total assets. Previous research indicates that goodwill accounting is a key distinction between the IFRS and JGAAP. Studies suggest that firms involved in M&A adopt IFRS to leverage these accounting differences, resulting in a positive association between IFRS adoption and the goodwill ratio (Kashiwazaki et al., 2019; J. Kim et al., 2019).

RD is the ratio of research and development (R&D) expenditure to sales. Unlike JGAAP, IFRS permit the capitalization of a portion of development costs. Studies find that companies actively involved in R&D are more likely to adopt IFRS to benefit from these accounting differences, leading to a positive correlation between IFRS adoption and the R&D ratio (Kashiwazaki et al., 2019; J. Kim et al., 2019).

CROSS is the cross-shareholder ratio. Since a reduction in cross-shareholdings expands the capacity for foreign investors—who are more likely to benefit from IFRS adoption—we anticipate a negative coefficient for this variable (J. Kim et al., 2019). FOREIGN SALES is the ratio of foreign sales to total sales. Firms with a high foreign sales ratio are likely to adopt IFRS to reduce the transaction costs arising from different accounting standards. Following J. Kim et al. (2019), we expect positive coefficients for this variable.

The remaining firm characteristics are as follows: SIZE (the natural logarithm of total assets in the previous fiscal year), ROA (return on assets), BTM (book-to-market ratio), CLOSELY HELD (closely-held shareholder ratio), DEBT (net debt-equity ratio), and AGE (natural logarithm of the number of years in business plus one). Previous studies using Japanese data show positive coefficients for SIZE and ROA and negative coefficients for AGE, BTM, CLOSELY HELD, and DEBT (Kashiwazaki et al., 2019; J. Kim et al., 2019; Sato & Takeda, 2017).

Large firms (large SIZE) are generally considered to be more capable of allocating sufficient resources to prepare for changes in accounting standards than small firms; accordingly, they are more likely to adopt IFRS.9 Firms with high ROA tend to operate profitably, whereas those with high BTM often exhibit strong growth potential. Such firms may have stronger incentives to adopt the IFRS to facilitate access to external financing. The higher the concentration of share ownership among major shareholders and executives (large CLOSELY HELD), the lower the need for information disclosure to investors, and the more likely they will not adopt an aggressive disclosure strategy. Firms with large DEBT are unlikely to adopt IFRS because banks can obtain the necessary information from companies through private conversations. Firms with long histories (large AGE) are likely to have more complex operations and, therefore, may be less inclined to adopt IFRS.

IRD1 and IRD2 are macroeconomic variables related to interest rate parity in international finance, which affects cross-border financial transactions.10 IRD1 and IRD2 capture the differences in interest rates between the US and Japan. IRD1 is the difference between the US federal fund rates and Japanese uncollateralized overnight call rates, whereas IRD2 is the difference in yields on three-month treasury bills in both countries. In accordance with the interest rate parity, which indicates capital flows from countries with low interest rates to those with high interest rates, we expect negative coefficients for these variables.

INDUSTRY is a set of dummy variables that equals zero for food companies and one for the other 25 industries: textile products, chemicals, drugs, petroleum, rubber products, stone, clay and glass products, iron and steel, non-ferrous metal and metal products, machinery, electric and electronic equipment, shipbuilding and repair, motor vehicles and auto parts, precision equipment, other manufacturing, fish and marine products, construction, wholesale trade, retail trade, credit and leasing, real estate, land transportation, air transportation, communication service, utilities—electric, and services.

3.2. DID

To perform the DID analysis, we select matched non-adopters based on one-to-one matching using the estimation results of the probit model for years of voluntary IFRS adoption. The estimated probit model is similar to probit model (1) but excludes INDUSTRY and macroeconomic variables (IRD1, IRD2).

We then estimate the following panel data model using ordinary least squares estimation for IFRS adopters and matched non-adopters for 2010–2023.

$$\begin{array}{l}{FOREIGNHELD}_{it}=\alpha +{\beta }_1{IFRS}_i+{\beta }_2{POST}_t+{\beta }_3{\left(IFRS\times POST\right)}_{it}\\ +{\beta }_4{BUYBACK}_{it}+{\beta }_5{CROSS}_{it}+{\beta }_6{BTM}_{it}\\ +{\beta }_7{(IRD1/IRD2/ERD1/ERD2)}_t+{\beta }_8{INDUSTRY}_i+{\beta }_9{YEAR}_t+{\epsilon }_{it}\end{array}$$

(2)

Most variables are the same as those in probit model (1). The variable POST is a dummy variable representing the post-adoption period; it takes a value of 1 for the years following voluntary IFRS adoption and 0 otherwise. We focus on the interaction term, $IFRS\times POST$, defined as 1 for IFRS adopters in the post-adoption period and 0 otherwise, to capture the impact of voluntary adoption. Based on our hypothesis, we expect positive coefficients for the interaction term,$IFRS\times POST$.

BUYBACK is the ratio of the treasury stock to the total number of outstanding shares. We expect negative coefficients for BUYBACK, because an increase in share buybacks reduces room for foreign shareholders. In contrast, we predict a positive coefficient for CROSS, as a reduction in cross-shareholdings creates more capacity for foreign ownership. ERD1 and ERD2 capture exchange rate movements. ERD1 is the natural logarithm of the yen-dollar exchange rate and ERD2 is the year-on-year rate of change in the yen-dollar exchange rate. Based on interest rate parity, which indicates capital flows from countries with weak currencies to those with strong currencies, we expect negative coefficients for IRD1, IRD2, ERD1, and ERD2. YEAR is a set of year dummies.

As a robustness check, we replace $FOREIGNHELD$ with an alternative variable related to the foreign shareholder ratio, $FOREIGNHELDNUM$, which is measured by the number of shareholders.

3.3. Sample Selection

According to the Japan Exchange Group website,11 273 listed firms voluntarily adopted IFRS by fiscal year 2022. After excluding financial institutions such as banks, securities companies, and insurance companies, 261 firms retained a complete set of variables for use in probit model (1). Subsequently, 168 firms remained in our sample following the caliper matching process.

4. Results and Discussion

4.1. Determinants of IFRS Adoption

We first examine the determinants of voluntary IFRS adoption among all listed Japanese firms for the period 2010–2023.

Table 1. Basic statistics and correlation coefficient matrix before matching.

	Mean	Std	Min	25%	50%	75%	Max
IFRS	0.0750	0.2634	0.0000	0.0000	0.0000	0.0000	1.0000
FOREIGN HELD	0.1109	0.1241	0.0000	0.0142	0.0649	0.1725	1.0000
GW	0.0142	0.0442	0.0000	0.0000	0.0000	0.0064	0.8060
RD	0.0438	1.2858	0.0000	0.0000	0.0028	0.0190	122.50
CROSS	0.0706	0.0900	0.0000	0.0000	0.0350	0.1115	0.6660
FOREIGN SALES	0.1553	0.2382	0.0000	0.0000	0.0000	0.2662	1.0000
SIZE	10.669	1.7590	3.8501	9.4966	10.532	11.688	19.512
ROA	0.0312	0.0882	−2.2252	0.0122	0.0308	0.0559	5.8562
CLOSELY HELD	0.5362	0.1561	0.0100	0.4190	0.5300	0.6550	1.0000
DEBT	0.1959	0.1848	0.0000	0.0368	0.1529	0.3099	7.6556
AGE	3.8353	0.6759	0.7308	3.5391	4.0768	4.2819	4.9553
BTM	1.1864	0.7990	0.0027	0.5973	1.0342	1.5907	9.8512
IRD1	0.5311	0.7485	−0.0114	0.0554	0.0950	0.6981	3.1295
IRD2	0.5934	0.8276	−0.0900	0.0220	0.1350	0.9250	3.9110

presents the descriptive statistics used in the probit model (1). The correlation coefficient matrix is provided in Appendix C,12 while

Table 2. Factors associated with IFRS adoption.

Dependent Variable = IFRS
Probit Model	[1]		[2]		[3]		[4]
FOREIGN HELD					0.2448	**	0.2488	**
					(0.035)		(0.032)
GW	6.6831	***	6.6908	***	6.6698	***	6.6774	***
	(0.000)		(0.000)		(0.000)		(0.000)
RD	0.0198	***	0.0200	***	0.0194	***	0.0195	***
	(0.000)		(0.000)		(0.000)		(0.000)
CROSS	−1.6888	***	−1.6812	***	−1.6290	***	−1.6202	***
	(0.000)		(0.000)		(0.000)		(0.000)
FOREIGN SALES	0.6907	***	0.6952	***	0.6678	***	0.6720	***
	(0.000)		(0.000)		(0.000)		(0.000)
SIZE	0.4059	***	0.4063	***	0.3959	***	0.3960	***
	(0.000)		(0.000)		(0.000)		(0.000)
ROA	0.3938	**	0.3943	**	0.3565	*	0.3567	*
	(0.042)		(0.041)		(0.075)		(0.075)
BTM	−0.4006	***	−0.4035	***	−0.3894	***	−0.3922	***
	(0.000)		(0.000)		(0.000)		(0.000)
CLOSELY HELD	0.4289	***	0.4352	***	0.4623	***	0.4692	***
	(0.000)		(0.000)		(0.000)		(0.000)
DEBT	−0.0216		−0.0227		0.0108		0.0104
	(0.774)		(0.764)		(0.888)		(0.893)
AGE	−0.1666	***	−0.1667	***	−0.1657	***	−0.1657	***
	(0.000)		(0.000)		(0.000)		(0.000)
IRD1	−0.0833	***			−0.0845	***
	(0.000)				(0.000)
IRD2			−0.0858	***			−0.0870	***
			(0.000)				(0.000)
Intercept	−5.2068	***	−5.2048	***	−5.1557	***	−5.1529	***
	(0.000)		(0.000)		(0.000)		(0.000)
INDUSTRY	Yes		Yes		Yes		Yes
Observations	35,473		35,435		35,473		35,435
Pseudo R2	0.4039		0.4042		0.4041		0.4043

Notes: The estimation results are based on the probit model and show a heteroscedasticity-robust estimator based on the results of the White and Breusch–Pagan tests. p-values are given in parentheses. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively.

reports the estimation results of the probit model. The results show a heteroscedasticity-robust estimator based on the results of the White and Breusch–Pagan tests. For all models, the signs of the coefficients are the same as expected, except for DEBT that has insignificant coefficients.

Table 3. Basic statistics and correlation coefficient matrix after matching.

Panel A: Covariate balance after matching
Variable			Treatment (a)				Control (b)				Difference (a)–(b) t-Statistics (p-Value)
FOREIGN HELD			0.2435				0.2360				0.0075 0.4721 (0.637)
CROSS			0.0494				0.0496				−0.0002 −0.0309 (0.975)
GW			0.0488				0.0523				−0.0035 −0.4066 (0.685)
RD			0.0353				0.0196				0.0157 *** 2.9407 (0.004)
FOREIGN SALES			0.3374				0.3164				0.0210 0.6367 (0.525)
SIZE			12.4515				12.3552				0.0963 0.4914 (0.623)
ROA			0.0552				0.0471				0.0081 0.7299 (0.466)
BTM			0.7697				0.7751				−0.0054 −0.0881 (0.930)
CLOSELY HELD			0.5001				0.5125				−0.0124 −0.6899 (0.491)
DEBT			0.2158				0.1853				0.0305 1.5918 (0.112)
AGE			3.8186				3.7615				0.0571 0.6934 (0.489)
Panel B: Descriptive statistics of the main variables
Variable			Mean		Std		Min	25%	50%		75%			Max
FOREIGN HELD			0.2323		0.1418		0.0000	0.124	0.2251		0.3258			0.7808
IFRS			0.4940		0.5000		0.0000	0.0000	0.0000		1.0000			1.0000
POST			0.4476		0.4973		0.0000	0.0000	0.0000		1.0000			1.0000
CROSS			0.0512		0.0677		0.0000	0.0000	0.0230		0.0798			0.4520
BUYBACK			0.0321		0.0455		0.0000	0.0021	0.0149		0.0464			0.4386
BTM			0.8528		0.5527		0.0058	0.4476	0.7662		1.1303			4.9520
IRD1			0.5327		0.7442		−0.0114	0.0554	0.0955		0.6981			3.1295
IRD2			0.5974		0.8224		−0.0900	0.0280	0.1390		0.9250			3.9110
ERD1			0.0169		0.0845		−0.1963	−0.0627	0.0078		0.0764			0.3471
ERD2			0.0207		0.0881		−0.1782	−0.0608	0.0078		0.0794			0.4149
Panel C: Correlation coefficient matrix
Variable	FOREIGN HELD		IFRS	POST		CROSS	BUYBACK	BTM	IRD1		IRD2		ERD1	ERD2
FOREIGN HELD			0.021	0.094 ***		0.101 ***	−0.013	−0.169 ***	0.058 ***		0.080 ***		0.031 *	0.031 *
IFRS	0.014			−0.010		0.011	0.028 *	−0.038 **	−0.004		−0.004		−0.008	−0.008
POST	0.102 ***		−0.010			−0.072 ***	0.017	−0.050 ***	0.503 ***		0.550 ***		0.085 ***	0.085 ***
CROSS	−0.013		0.039 **	−0.077 ***			0.170 ***	0.305 ***	0.036 **		0.009		−0.042 ***	−0.042 ***
BUYBACK	−0.116 ***		0.006	−0.000		0.072 ***		0.093 ***	0.003		0.004		0.017	0.017
BTM	−0.193 ***		−0.030 *	−0.003		0.223 ***	0.027 *		−0.106 ***		−0.129 ***		−0.029 *	−0.029 *
IRD1	0.050 ***		−0.008	0.360 ***		0.002	−0.012	−0.085 ***			0.945 ***		−0.237 ***	−0.237 ***
IRD2	0.056 ***		−0.008	0.397 ***		−0.000	−0.012	−0.095 ***	0.989 ***				−0.069 ***	−0.069 ***
ERD1	0.023		−0.006	0.015		−0.029 *	0.004	−0.017	−0.120 ***		−0.100 ***			1.000 ***
ERD2	0.021		−0.007	0.003		−0.028 *	0.004	−0.019	−0.128 ***		−0.109 ***		0.999 ***
Panel D: Distribution of voluntary IFRS adopters
Industry	2010	2011	2012	2013	2014	2015	2016	2017	2018	2019	2020	2021	2022	Total
01_Foods	0	0	0	0	0	0	0	1	2	3	2	1	0	9
03_Textile Products	0	0	0	0	0	0	0	0	0	0	0	1	0	1
07_Chemicals	0	0	0	0	0	2	0	1	5	1	1	1	0	11
09_Drugs	0	0	0	1	3	1	0	1	2	0	1	0	1	10
11_Petroleum	0	0	0	0	0	0	0	1	0	0	0	0	0	1
13_Rubber Products	0	0	0	0	0	0	2	1	0	1	0	0	0	4
15_Stone, Clay and Glass Products	0	0	1	1	0	0	0	0	0	0	1	0	0	3
17_Iron and Steel	0	0	0	0	0	1	0	0	0	2	0	0	1	4
19_Non ferrous Metal and Metal Products	0	0	0	0	0	0	1	1	0	0	0	0	0	2
21_Machinery	0	0	0	0	0	2	2	2	1	4	1	3	0	15
23_Electric and Electronic Equipment	1	0	0	1	1	4	2	2	1	4	1	0	0	17
27_Motor Vehicles and Auto Parts	0	0	0	0	0	5	4	1	2	2	1	0	0	15
31_Precision Equipment	0	1	0	0	0	0	1	2	3	0	0	0	0	7
33_Other Manufacturing	0	0	0	0	0	0	0	0	0	1	2	0	0	3
41_Construction	0	0	0	0	0	0	0	0	0	0	0	1	0	1
43_Wholesale Trade	0	0	0	1	5	1	1	2	0	1	1	1	0	13
45_RetailTrade	0	0	0	0	0	0	0	0	3	0	0	0	0	3
52_Credit and Leasing	0	0	0	1	0	2	0	0	1	1	0	0	0	5
53_Real Estate	0	0	0	1	0	0	0	0	0	0	0	0	2	3
57_Trucking	0	0	0	0	0	1	0	0	0	0	0	0	0	1
65_Communication Services	0	0	0	0	0	0	0	1	0	0	0	0	0	1
67_Utilities—Electric	0	0	0	0	0	0	0	0	0	0	0	1	0	1
71_Services	0	0	0	2	0	6	6	3	6	7	1	5	2	38
Total	1	1	1	8	9	25	19	19	26	27	12	14	6	168

Note: Panel A: This table reports the means of each covariate in the treatment and control groups after propensity score matching, as well as the difference in means and corresponding p-values from Welch’s t-test. p-values are reported in parentheses. *** indicates statistical significance at the 1% level. Panel C: This correlation coefficient matrix provides the Pearson’s product-moment correlation coefficients in the lower left and Spearman’s rank correlation coefficients in the upper right relative to the diagonal components. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively.

provides the basic statistics used for the DID analysis. Panel A presents the results of the balancing tests after matching. Except for RD, the means of all variables do not differ significantly between IFRS adopters and non-adopters. Panel B presents the descriptive statistics for IFRS adopters and their matched non-adopters, based on propensity score matching. Panel C presents the correlation coefficient matrix. For both Pearson’s product-moment correlation coefficient and Spearman’s rank correlation coefficient, FOREIGN HELD is positively associated with POST at the 1% significance level, while it is not significantly associated with IFRS. The signs of the coefficients are mixed for CROSS and BUYBACK between Pearson’s and Spearman’s. Panel D presents the industry distribution of the voluntary IFRS adopters included in the analysis.

Table 4. Univariate DID analysis.

Panel A: DID Analysis: Foreign shareholders’ ratio (FOREIGN HELD)
Sample	Pre-IFRS period (a)	Post-IFRS period (b)	Difference (b)–(a)
Matched non-adopters (i)	0.2195	0.2414	0.0219 ***
	(n = 1121)	(n = 913)	(0.001)
IFRS adopters (ii)	0.2160	0.2556	0.0396 ***
	(n = 1113)	(n = 872)	(0.000)
(ii)–(i)	−0.0035	0.0142 **	0.0177
	(0.528)	(0.033)
Panel B: DID Analysis: Foreign shareholders’ numbers (FOREIGN HELD NUM)
Sample	Pre-IFRS period (a)	Post-IFRS period (b)	Difference (b)–(a)
Matched non-adopters (i)	11.2436	11.7888	0.5452 ***
	(n = 1121)	(n = 913)	(0.000)
IFRS adopters (ii)	11.4973	12.5744	1.0771 ***
	(n = 1113)	(n = 872)	(0.000)
(ii)–(i)	0.2537 **	0.7856 ***	0.5319
	(0.026)	(0.000)

Note: p-values for Welch’s t-test are in parentheses. *** and ** indicate statistical significance at the 1% and 5% levels, respectively.

presents the results of the univariate DID analysis of the two foreign shareholder-related variables of IFRS adopters and their matched non-adopters between the pre-IFRS period and post-IFRS period. The pre-IFRS period indicates the period before IFRS adoption, whereas the post-IFRS period indicates the period after adoption. Panels A and B correspond to the results of $FOREIGNHELD$ and $FOREIGNHELDNUM$.

Before voluntary IFRS adoption, Panel A shows that the foreign shareholder ratio ($FOREIGNHELD$) is not significantly different between IFRS adopters and non-adopters in the pre-IFRS period. Panel B shows that the number of foreign shareholders ($FOREIGNHELDNUM$) is significantly higher among IFRS adopters than among non-adopters. After voluntary IFRS adoption, Panels A and B show that both IFRS adopters and non-adopters significantly increase foreign shareholder ratios ($FOREIGNHELD$) and the number of foreign shareholders ($FOREIGNHELDNUM$) at the 1% level. In the post-IFRS period, while the number of foreign shareholders ($FOREIGNHELDNUM$) is still larger for IFRS adopters, the foreign shareholders’ ratio ($FOREIGNHELD$) also becomes larger for IFRS adopters at the 5% level. Thus, the univariate analysis indicates that IFRS adoption increases foreign investment.

4.2. Factors Affecting the Foreign Shareholder Ratio

Table 5. Panel data analysis: foreign shareholders’ ratio estimates.

Panel A: Factors associated with foreign shareholders’ ratio
Dependent Variable = FOREIGN HELD
OLS Regression	[1]		[2]		[3]		[4]		[5]
IFRS	−0.0102	*	−0.0103	*	−0.0104	*	−0.0104	*	−0.0099	*
	(0.080)		(0.078)		(0.074)		(0.075)		(0.088)
POST	0.0143	**	0.0140	**	0.0134	**	0.0135	**	0.0069
	(0.023)		(0.028)		(0.028)		(0.027)		(0.387)
IFRS × POST	0.0260	***	0.0261	***	0.0262	***	0.0262	***	0.0257	***
	(0.002)		(0.002)		(0.002)		(0.002)		(0.002)
BUYBACK	−0.2851	***	−0.2854	***	−0.2914	***	−0.2913	***	−0.2869	***
	(0.000)		(0.000)		(0.000)		(0.000)		(0.000)
CROSS	0.0094		0.0090		0.0100		0.0099		0.0057
	(0.734)		(0.744)		(0.715)		(0.718)		(0.841)
BTM	−0.0632	***	−0.0632	***	−0.0645	***	−0.0644	***	−0.0613	***
	(0.000)		(0.000)		(0.000)		(0.000)		(0.000)
IRD1	0.0013
	(0.667)
IRD2			0.0014
			(0.615)
ERD1					0.0375
					(0.131)
ERD2							0.0343
							(0.152)
Intercept	0.2408	***	0.2408	***	0.2442	***	0.2441	***	0.2169	***
	(0.000)		(0.000)		(0.000)		(0.000)		(0.000)
INDUSTRY	Yes		Yes		Yes		Yes		Yes
YEAR	No		No		No		No		Yes
Observations	4019		4018		3989		3989		4019
R2	0.177		0.177		0.178		0.178		0.182
Adj R2	0.170		0.170		0.171		0.171		0.173
Panel B: Factors associated with foreign shareholders’ numbers
Dependent Variable = FOREIGN HELD NUM
OLS Regression	[1]		[2]		[3]		[4]		[5]
IFRS	0.1091	**	0.1092		0.1260		0.1259		0.1109
	(0.025)		(0.249)		(0.182)		(0.182)		(0.241)
POST	0.4886	***	0.4823	***	0.4877	***	0.4876	***	0.2327	*
	(0.000)		(0.000)		(0.000)		(0.000)		(0.051)
IFRS × POST	0.5835	***	0.5833	***	0.5682	***	0.5683	***	0.5885	***
	(0.000)		(0.000)		(0.000)		(0.000)		(0.000)
BUYBACK	−6.7152	***	−6.7142	***	−6.9447	***	−6.9440	***	−6.7230	***
	(0.000)		(0.000)		(0.000)		(0.000)		(0.000)
CROSS	0.8600		0.8554		0.9050	*	0.9035	*	0.7665
	(0.105)		(0.107)		(0.081)		(0.081)		(0.161)
BTM	−0.9259	***	−0.9241	***	−0.9455	***	−0.9456	***	−0.9280	***
	(0.000)		(0.000)		(0.000)		(0.000)		(0.000)
IRD1	0.0396
	(0.387)
IRD2			0.0420
			(0.320)
ERD1					−0.0413
					(0.918)
ERD2							−0.0915
							(0.813)
Intercept	11.8680	***	11.8647	***	11.9404	***	11.9418	***	11.6453	***
	(0.000)		(0.000)		(0.000)		(0.000)		(0.000)
INDUSTRY	Yes		Yes		Yes		Yes		Yes
YEAR	No		No		No		No		Yes
Observations	4019		4018		3989		3989		4019
R2	0.215		0.215		0.214		0.214		0.218
Adj R2	0.208		0.208		0.207		0.207		0.209

Notes: Panel A: The regression results are based on an ordinary least squares estimation and show a heteroscedasticity-robust estimator based on the White and Breusch–Pagan test results. p-values are given in parentheses. ***, **, and * indicate statistical significance at 1%, 5%, and 10% levels, respectively. Panel B: The regression results are based on ordinary least squares estimation and show a heteroscedasticity-robust estimator based on the White and Breusch–Pagan test results. p-values are given in parentheses. ***, **, and * indicate statistical significance at 1%, 5%, and 10% levels, respectively.

presents the estimation results of multivariate regression model (2). As a robustness check, we replace the dependent variable $FOREIGNHELD$ in Panel A with an alternative variable related to the foreign shareholder ratio, $FOREIGNHELDNUM$ in Panel B. The results show a heteroscedasticity-robust estimator based on the results of the White and Breusch–Pagan tests.

In Panel A, the IFRS dummy variable has significantly negative coefficients in all models, whereas POST has significantly positive coefficients at the 5% level in all models except Model [5]. Our target interaction term, $IFRS\times POST$, has significantly positive coefficients at the 1% level for all models. These results are consistent with our hypothesis. In other words, the increase in foreign shareholder ratios is more pronounced in listed Japanese companies that voluntarily adopt IFRS than in their non-adopting counterparts. Across all model iterations, the coefficients for BUYBACK are negative and significant at the 1% level; conversely, CROSS remains positive without reaching statistical significance. These results imply that the room for foreign shareholders is narrowed by buyback activities but is not meaningfully increased by declining cross-shareholdings. These results are consistent with Miyajima and Saito’s (2023) claim that firms conduct share buybacks using the money created by reducing cross-shareholdings. Panel B provides similar results to Panel A, showing positive coefficients for POST and the interaction ($IFRS\times POST$) and negative coefficients for BUYBACK.

4.3. Robustness Checks

As a robustness check, we estimate the multivariate regression model (2) using winsorized variables. The results are presented in

Table 6. Panel data analysis to estimate foreign shareholders’ ratio using winsorized data.

Dependent Variable = FOREIGN HELD
OLS Regression	[1]		[2]		[3]		[4]		[5]
IFRS	−0.0103	*	−0.0104	*	−0.0104	*	−0.0104	*	−0.0100	*
	(0.074)		(0.072)		(0.070)		(0.070)		(0.081)
POST	0.0135	**	0.0132	**	0.0125	**	0.0126	**	0.0061
	(0.029)		(0.034)		(0.038)		(0.037)		(0.428)
IFRS × POST	0.0259	***	0.0260	***	0.0261	***	0.0260	***	0.0256	***
	(0.002)		(0.002)		(0.002)		(0.002)		(0.002)
BUYBACK	−0.2746	***	−0.2750	***	−0.2813	***	−0.2812	***	−0.2761	***
	(0.000)		(0.000)		(0.000)		(0.000)		(0.000)
CROSS	0.0275		0.0271		0.0278		0.0277		0.0249
	(0.353)		(0.360)		(0.346)		(0.348)		(0.415)
BTM	−0.0671	***	−0.0671	***	−0.0682	***	−0.0682	***	−0.0651	***
	(0.000)		(0.000)		(0.000)		(0.000)		(0.000)
IRD1	0.0009
	(0.745)
IRD2			0.0011
			(0.693)
ERD1					0.0408	*
					(0.097)
ERD2							0.0381
							(0.109)
Intercept	0.2422	***	0.2422	***	0.2455	***	0.2454	***	0.2181	***
	(0.000)		(0.000)		(0.000)		(0.000)		(0.000)
INDUSTRY	Yes		Yes		Yes		Yes		Yes
YEAR	No		No		No		No		Yes
Observations	4019		4018		3989		3989		4019
R2	0.177		0.177		0.178		0.178		0.182
Adj R2	0.170		0.170		0.171		0.171		0.173

Notes: The regression results are based on an ordinary least squares estimation and show a heteroscedasticity-robust estimator based on the White and Breusch–Pagan test results. p-values are given in parentheses. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively. All continuous variables are winsorized at the 1st/99th percentiles to mitigate the influence of outliers.

, which is based on ordinary least squares estimation and shows a heteroscedasticity-robust estimator based on the results of the White and Breusch–Pagan tests. All continuous variables are winsorized at the 1st/99th percentiles to mitigate the influence of outliers. These results are similar to those shown in Panel A of Table 5. For all models, the coefficients on IFRS are significantly negative at the 10% level; POST has significantly positive coefficients at the 5% level, except for Model [5]; the interaction term, $IFRS\times POST$, has significantly positive coefficients at the 1% level; BUYBACK has significantly negative coefficients at the 1% level; and the coefficients on CROSS are positive but statistically insignificant.

Table 7. Panel data analysis including industry-specific time trends.

Dependent Variable = FOREIGN HELD
OLS Regression	[1] Panel A of Table 5		[1]’ Industry Time Trend		[5] Panel A of Table 5		[5]’ Industry Time Trend
IFRS	−0.0102	*	−0.0117	**	−0.0099	*	−0.0117	**
	(0.080)		(0.049)		(0.088)		(0.049)
POST	0.0143	**	0.0003		0.0069		0.0049
	(0.023)		(0.974)		(0.387)		(0.547)
IFRS × POST	0.0260	***	0.0296	***	0.0257	***	0.0293	***
	(0.002)		(0.001)		(0.002)		(0.001)
BUYBACK	−0.2851	***	−0.2934	***	−0.2869	***	−0.2948	***
	(0.000)		(0.000)		(0.000)		(0.000)
CROSS	0.0094		0.0096		0.0057		0.0123
	(0.734)		(0.728)		(0.841)		(0.666)
BTM	−0.0632	***	−0.0687	***	−0.0613	***	−0.0673	***
	(0.000)		(0.000)		(0.000)		(0.000)
IRD1	0.0013		−0.0012
	(0.667)		(0.685)
Intercept			0.2514	***			0.2581	***
			(0.000)				(0.000)
INDUSTRY	Yes		Yes		Yes		Yes
YEAR	No		No		Yes		Yes
INDUSTRY time trends	No		Yes		No		Yes
Observations	4019		4019		4019		4019
R2	0.177		0.193		0.182		0.196
Adj R2	0.170		0.180		0.173		0.181

Notes: The regression results are based on an ordinary least squares estimation and show a heteroscedasticity-robust estimator based on the White and Breusch–Pagan test results. p-values are given in parentheses. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively.

presents additional robustness checks. Models [1] and [5] represent the baseline results from Table 5, Panel A, while Models [1]’ and [5]’ incorporate 28 industry-specific time trends to account for evolving governance and reporting patterns in each industry.13 The results indicate that the coefficients on the interaction term, $IFRS\times POST$, remain positive and significant at the 1% level across both specifications. Furthermore, F-tests for Models [1]’ and [5]’ reject the null hypothesis that the industry trends are jointly zero at the 1% significance level. In other words, IFRS adoption leads to deviations from preexisting industry-specific trends. Consequently, these findings underscore the robustness of the main results.

4.4. Validity of Parallel Trend Assumption

Next, we test the validity of the parallel trend assumption, a crucial requirement in the DID approach, which posits that IFRS adopters and non-adopters would follow a similar pattern in the pre-adoption period. This implies that, absent the IFRS adoption, both groups would have evolved similarly. A violation of this assumption would suggest divergent trends regardless of IFRS adoption. To evaluate the parallel trend assumption, this study employs two statistical approaches.

The first approach is the inclusion of linear IFRS adopter-specific time trend in regression model (2). The results are presented in

Table 8. Panel data analysis including IFRS adopter-specific time trends.

Dependent Variable = FOREIGN HELD
OLS Regression	[1] Panel A of Table 5		[1]” IFRS Time Trend		[5] Panel A of Table 5		[5]” IFRS Time Trend
IFRS	−0.0102	*	0.0161	*	−0.0099	*	0.0165	*
	(0.080)		(0.089)		(0.088)		(0.080)
POST	0.0143	**	−0.0118		0.0069		−0.0082
	(0.023)		(0.193)		(0.387)		(0.375)
IFRS × POST	0.0260	***	0.0565	***	0.0257	***	0.0563	***
	(0.002)		(0.000)		(0.002)		(0.000)
TREND			0.0050	***
			(0.000)
IFRS × TREND			−0.0056	***			−0.0057	***
			(0.001)				(0.001)
BUYBACK	−0.2851	***	−0.2814	***	−0.2869	***	−0.2827	***
	(0.000)		(0.000)		(0.000)		(0.000)
CROSS	0.0094		0.0067		0.0057		0.0088
	(0.734)		(0.809)		(0.841)		(0.758)
BTM	−0.0632	***	−0.0633	***	−0.0613	***	−0.0616	***
	(0.000)		(0.000)		(0.000)		(0.000)
IRD1	0.0013		−0.0009
	(0.667)		(0.762)
Intercept			0.2172	***			0.2072	***
			(0.000)				(0.000)
INDUSTRY	Yes		Yes		Yes		Yes
YEAR	No		No		Yes		Yes
Observations	4019		4019		4019		4019
R2	0.177		0.181		0.182		0.185
Adj R2	0.170		0.173		0.173		0.175

Notes: The regression results are based on an ordinary least squares estimation and show a heteroscedasticity-robust estimator based on the White and Breusch–Pagan test results. p-values are given in parentheses. ***, **, and * indicate statistical significance at 1%, 5%, and 10% levels, respectively.

. Models [1] and [5] reprint the baseline results from Table 5, Panel A, while Models [1]” and [5]” incorporate IFRS adopter-specific time trends, expressed by the interaction term $IFRS\times TREND$, where TREND is a time trend taking the value of 1 for 2010, 2 for 2011, and so on. The results show that the p-value for the interaction term, $IFRS\times TREND$, rejects the null hypothesis that the IFRS-specific time trends are zero at the 1% significance level. The coefficients on the main interaction term of interest, $IFRS\times POST$, remain positive and significant at the 1% level, with their magnitudes increasing by approximately 2% in the trend-adjusted specifications. This confirms that IFRS adoption causes a distinct deviation from preexisting adopter-specific trends.

The second approach utilizes an event study design.14 Following Saravanan and Firoz (2026), we incorporate year-specific indicator variables (Period), each interacting with an IFRS dummy variable (IFRS × Period), to identify firms voluntarily adopting IFRS. Specifically, we include indicator variables for the two to three years prior to adoption (t − 2 and t − 3), and zero to three years post-adoption (t0, t + 1, t + 2 and t + 3). Consequently, the year immediately preceding adoption (t − 1) serves as the benchmark period against which all other interaction coefficients are compared. By examining the statistical significance of the pre-adoption interaction terms (IFRS × Period_t−k, k = 2, 3), this study assesses whether there is any systematic divergence in trends between IFRS adopters and non-adopters prior to implementation.

Table 9. Panel data analysis augmented with leads and lags.

Dependent Variable = FOREIGN HELD
OLS Regression	[6] Leads and Lags		[7] Leads and Lags
IFRS	0.0004		0.0006
	(0.970)		(0.958)
IFRS × Periodt−3	−0.0141		−0.0163
	(0.324)		(0.261)
IFRS × Periodt−2	−0.0049		−0.0050
	(0.733)		(0.726)
IFRS × Periodt0	0.0026		0.0024
	(0.852)		(0.863)
IFRS × Periodt+1	0.0072		0.0069
	(0.605)		(0.626)
IFRS × Periodt+2	0.0073		0.0086
	(0.612)		(0.561)
IFRS × Periodt+3	0.0132		0.0147
	(0.376)		(0.343)
BUYBACK	−0.4517	***	−0.4469	***
	(0.000)		(0.000)
CROSS	−0.0029		−0.0066
	(0.935)		(0.859)
BTM	−0.0537	***	−0.0534	***
	(0.000)		(0.000)
IRD1	−0.0047
	(0.190)
Intercept	0.2410	***	0.2589	***
	(0.000)		(0.000)
INDUSTRY	Yes		Yes
YEAR	No		Yes
Observations	2147		2147
R2	0.161		0.163
Adj R2	0.146		0.144

Notes: The regression results are based on an ordinary least squares estimation and show a heteroscedasticity-robust estimator based on the White and Breusch–Pagan test results. p-values are given in parentheses. *** indicates statistical significance at 1% level.

provides estimates of a subset of Models [1] and [5] in Table 5, augmented with leads and lags of the IFRS adoption. For both models, the coefficients on the adoption leads (IFRS × Period_t−k, k = 2, 3) are negative and statistically insignificant, suggesting little evidence of an anticipatory response among IFRS adopters. In the year of adoption, the coefficient on the interaction term (IFRS × Period_t0) becomes positive, after which this increment fluctuates between 0.7% and 1.5% over the subsequent three years (for IFRS × Period_t+k where k$\in \left\{1,2,3\right\}$). This pattern of coefficients provides evidence that IFRS adoption increases the foreign shareholding ratio, rather than vice versa.

Overall, our analysis is consistent with the hypothesis that the foreign shareholder ratio of listed Japanese companies increases significantly after voluntary IFRS adoption when explicitly controlling for cross-shareholdings and share buybacks over an extended period. These results are not observed in the study by J. Kim et al. (2019). This study contributes to the literature in two ways. First, we integrate recent findings on Japanese corporate governance with the literature on the effects of IFRS adoption by explicitly incorporating cross-shareholdings and share buybacks over an extended sample period. Second, we provide evidence that diverges from the findings of a previous study by employing a different variable and model, showing a positive correlation between IFRS adoption and the foreign shareholder ratio in Japan from 2010 to 2023.

4.5. Heterogeneity Analysis

This subsection presents an additional heterogeneity analysis based on firm size. Traditionally, larger firms exhibit more robust corporate governance and diminished information asymmetry. We divide IFRS adopters and non-adopters into two groups based on total assets and estimate the multivariate regression model (2). The results are presented in

Table 10. Heterogeneity analysis to estimate foreign shareholders’ ratio by firm size.

Dependent Variable = FOREIGN HELD
OLS Regression	[1] for Larger Firms		[5] for Larger Firms		[1] for Smaller Firms		[5] for Smaller Firms
IFRS	−0.0060		−0.0059		−0.0344	***	−0.0338	***
	(0.462)		(0.472)		(0.000)		(0.000)
POST	0.0052		0.0026		−0.0110		−0.0233	**
	(0.539)		(0.800)		(0.170)		(0.020)
IFRS × POST	0.0281	***	0.0271	***	0.0101		0.0094
	(0.007)		(0.010)		(0.386)		(0.419)
BUYBACK	−0.0375		−0.0378		−0.1541	***	−0.1591	***
	(0.633)		(0.626)		(0.001)		(0.000)
CROSS	−0.2098	***	−0.2191	***	0.0180		0.0106
	(0.000)		(0.000)		(0.659)		(0.798)
BTM	−0.0542	***	−0.0509	***	−0.0666	***	−0.0643	***
	(0.000)		(0.000)		(0.000)		(0.000)
IRD1	0.0042				0.0028
	(0.239)				(0.486)
Intercept	0.2656	***	0.2446	***	0.2327	***	0.2046	***
	(0.000)		(0.000)		(0.000)		(0.000)
INDUSTRY	Yes		Yes		Yes		Yes
YEAR			Yes				Yes
Observations	2000		2000		1988		1988
R2	0.231		0.239		0.168		0.175
Adj R2	0.218		0.221		0.155		0.158

Notes: The regression results are based on an ordinary least squares estimation and show a heteroscedasticity-robust estimator based on the White and Breusch–Pagan test results. p-values are given in parentheses. *** and ** indicate statistical significance at 1% and 5% levels, respectively.

. The interaction term, $IFRS\times POST,$ carries significantly positive coefficients at the 1% level in both Models [1] and [5] for larger firms but remains insignificant for smaller firms. These results indicate that the post-IFRS adoption increase in foreign shareholding is significantly more pronounced among larger firms compared to their smaller counterparts.

5. Concluding Remarks

This study examines the effect of voluntary IFRS adoption on the foreign shareholder ratio of listed Japanese companies for the period 2010–2023. Unlike the previous study, which employed a different variable and model, our DID analysis demonstrates an increase in the foreign shareholder ratio following IFRS adoption, after explicitly controlling for cross-shareholdings and share buybacks that tend to affect opportunities for foreign shareholders. This study contributes to the literature on the effects of IFRS adoption on foreign equity investment in two ways. First, it provides evidence that Japan has undergone a governance system transition concurrent with its adoption. Second, this study demonstrates the importance of controlling confounding factors when multiple institutional changes are ongoing. We believe that the findings of this study are valuable not only to international investors considering investments in countries such as Japan but also to firms contemplating IFRS adoption and anticipating changes in shareholder composition.

Although this study contributes to the literature on the effects of IFRS adoption by providing evidence from the Japanese context, we acknowledge its limitations. Following J. Kim et al. (2019), we employ a panel data model to conduct a DID analysis. A panel data model is useful for increasing the sample size. However, the periods before and after IFRS adoption differ across sample firms. Instead, M. DeFond et al. (2011) and M. L. DeFond et al. (2012) compare changes in the variables three years before and after IFRS adoption. Unlike Europe, the number of IFRS adopters in Japan is not large. As the number of IFRS adopters increases in the future, however, a research design similar to that of M. DeFond et al. (2011) and M. L. DeFond et al. (2012) could be utilized. Another promising direction for future research is the integration of behavioral finance factors—such as trust and political risk aversion—which may influence the dynamics between corporate governance and foreign investment.