Expected Credit Spreads and Market Choice: Evidence from Japanese Bond Issuers

Abstract

This study explores the impact of credit spreads—defined as the difference between corporate bond yields and matched government bond yields—and macro-financial conditions on Japanese firms’ decision-making regarding whether to issue corporate bonds in domestic or international markets. Using firm-level panel data from 2010 to 2019, we employ fixed-effects regressions to identify the determinants of credit spreads and assess their influence on issuance location. The results suggest that firms strategically opt for foreign markets when anticipating narrower spreads, despite the typically higher borrowing costs associated with overseas issuance. Sensitivity to credit spreads systematically varies with issuer characteristics—such as leverage and credit ratings—and market elements—including the United States volatility and stock performance. Interaction models further demonstrate that market selection dynamically responds to pricing signals and uncertainty. By connecting credit spread formation to venue choice, this study provides a new perspective on cross-border financing in segmented capital markets. These findings offer theoretical insights and practical implications for understanding how firms adapt their debt strategies in response to global financial conditions.

1. Introduction

In recent decades, Japanese corporations have increasingly diversified their funding strategies by accessing international bond markets, driven by shifts in global capital flows, monetary policy divergence, and evolving investor demand. Although domestic bond issuance remains a cornerstone of corporate finance in Japan—supported by established relationships with local financial institutions and a stable investor base—foreign issuance offers opportunities for cost reduction, currency matching, and broader investor outreach.

However, the decision to issue bonds in a foreign market extends beyond firm size or globalization trends. It reflects a complex interplay of firm-specific characteristics (such as credit ratings and financial leverage), macroeconomic indicators (including interest rate differentials and global volatility), and market-level factors (e.g., liquidity conditions and regulatory friction).

A key consideration is the credit spread, where premium firms must have a benchmark risk-free rate to raise capital. Credit spreads—hereafter referred to as the credit spread index (CS)—serve as forward-looking proxies for the cost of debt and are influenced by both global risk appetite and issuer fundamentals. However, their impact on the choice to issue credit domestically or abroad remains underexplored.

Most empirical studies have either examined the determinants of credit spreads in isolation or treated the choice of issuance market as a binary outcome shaped by macroeconomic shocks or firm-level constraints. This segmented approach overlooks how firms jointly assess pricing conditions across markets when developing debt strategies. In Japan, where monetary easing and yield curve controls have suppressed domestic spreads, the relative pricing advantage of foreign markets may have become more salient, highlighting the need to consider firm decision-making as a dynamic response to evolving cost landscapes. Building on this observation, we hypothesize that Japanese firms internalize expected credit spreads when selecting issuance markets, favoring foreign venues when anticipated pricing advantages outweigh domestic familiarity. This hypothesis guides our empirical investigation and frames the strategic nature of market choice under segmented conditions. Recent work by Anderson and Cesa-Bianchi (2024) further emphasizes the importance of firm-level heterogeneity in shaping credit spread sensitivity, showing that financial leverage and credit quality significantly influence how firms respond to market conditions. Their findings support the notion that issuance decisions are not uniform across firms but rather reflect differentiated expectations and constraints. This aligns with our hypothesis that Japanese firms evaluate issuance markets based on firm-specific spread expectations. This perspective reinforces the need to examine how Japanese firms internalize pricing signals when navigating segmented bond markets.

The novelty of this study lies in its dual focus on (1) the integration of spread expectations into firms’ market selection behavior and (2) the contextual emphasis on Japanese corporations, whose financing decisions are shaped by a distinct institutional and monetary environment. Unlike prior research centered on U.S. or emerging markets, our framework captures the dynamic optimization process of firms navigating segmented pricing landscapes.

For instance, Feyen et al. (2015) show that global liquidity conditions drive aggregate bond issuance in emerging markets, emphasizing macro-level push factors. In contrast, our study focuses on firm-level responses to expected credit spreads, highlighting strategic behavior under segmented pricing conditions.

This study addresses this gap by asking: To what extent does the issuance market environment influence credit spreads? How do these spread expectations affect firms’ decisions to raise capital in domestic and foreign bond markets? Using a panel dataset of Japanese bond issuers from 2010 to 2019, we employ firm-level fixed effects regressions to disentangle the effects of issuer characteristics, macro factors, and market-specific conditions on credit spreads. This study also investigates whether firms internalize these expected pricing outcomes in their market access decisions.

This study contributes to the literature on corporate financing under capital market frictions by reframing international bond issuances as a strategic response to credit pricing across segmented markets. First, it integrates the pricing dimension—specifically expected credit spreads—into the analysis of issuance market choice, which previous studies have primarily treated as separate domains. Second, while global research focuses on the United States (U.S.) or emerging market firms, this study examines the behavior of Japanese corporations operating within a low-yield, bank-oriented financial system. Third, it enhances the understanding of corporate financial strategy by interpreting bond issuance location as an endogenous outcome of spread-based optimization rather than merely a reflection of fixed constraints.

Taken together, these contributions offer a new perspective on how firms navigate global capital markets and provide policy-relevant insights into the conditions that promote or inhibit international bond issuances. It also informs policymakers about the channels through which market segmentation and interest rate policies may influence cross-border corporate behavior.

2. The Related Literature

2.1. Theoretical Underpinning

The choice between domestic and foreign bond markets is not arbitrary; rather, it reflects a complex interplay between firm-level characteristics and the broader macro-institutional landscape in which firms operate. Theoretical frameworks such as signaling theory suggest that firms may issue in foreign markets to convey credibility and reputational strength (Miller & Puthenpurackal, 2005). Similarly, market segmentation models posit that institutional and regulatory barriers create pricing differentials across markets, prompting firms to engage in strategic issuance behavior (Bhattacharyay, 2013).

Credit spreads serve as a fundamental pricing metric in bond markets, reflecting both issuer-specific credit risk and broader market liquidity conditions. Collin-Dufresn et al. (2001) theoretically highlight the roles of default probability, market illiquidity, and macroeconomic volatility in determining spreads. These insights provide the conceptual foundation for understanding how firms may anticipate pricing outcomes when selecting issuance venues.

Additionally, financial market dynamics—such as equity volatility and global liquidity shocks—are theorized to influence bond pricing and issuance behavior. Indicators like the Chicago Board Options Exchange Volatility Index (VIX), a widely used measure of market volatility and investor risk aversion, capture investor sentiment and risk aversion, which may alter the relative attractiveness of foreign markets. This study builds on these theoretical perspectives to explore how firms incorporate expected credit spread differentials and risk perceptions into their market choice decisions. These theoretical insights collectively suggest that firms do not select issuance markets randomly, but rather in anticipation of pricing outcomes and perceived risks. Building on this foundation, the following section develops testable hypotheses that link expected credit spread conditions to market choice behavior, particularly in the context of Japanese firms.

In the Japanese bond market, several structural constraints limit firms’ flexibility in accessing domestic capital. Regulatory requirements for public issuance—including mandatory credit ratings, minimum issuance sizes, and disclosure standards—can pose significant hurdles, especially for mid-sized firms. Moreover, the investor base remains relatively narrow, dominated by domestic institutional investors with conservative risk preferences and a strong home bias. These features contribute to limited market depth and pricing inefficiencies, particularly during periods of macroeconomic uncertainty.

Recent analysis by the Bank of Japan, Japan’s central monetary authority (Ochi & Osada, 2024), highlights persistent challenges in market functioning, especially in the secondary market, where low liquidity and limited trading volumes hinder efficient price discovery and capital allocation.

Against this backdrop, foreign bond markets offer Japanese firms alternative avenues for capital raising, often with greater pricing transparency, broader investor reach, and more flexible issuance conditions. The decision to issue abroad may reflect strategic responses to domestic constraints, as well as signaling motives aimed at enhancing credibility in global financial circles. In this context, expected credit spread differentials serve not only as cost indicators but also as proxies for market accessibility and reputational positioning.

2.2. The Empirical Literature

The extensive empirical literature has examined the factors that influence corporate bond issuance decisions across different markets. Previous studies have highlighted various determinants of market access and preferences, including capital account openness, regulatory conditions, and firm-specific considerations such as asset tangibility and reputational concerns. For example, X. Wang et al. (2023) demonstrate that capital account liberalization substantially increases international bond issuance among Chinese firms, particularly those with high tangible asset ratios. Similarly, Mizen and Tsoukas (2014) explore firm-level issuance behavior across nine Asian economies, emphasizing the roles of market depth, liquidity, and supranational policy initiatives.

Zhu (2013) challenges conventional interpretations of capital structure determinants by addressing methodological issues in ratio-based models. Using a probit framework, this study finds that more profitable firms are more likely to issue debt, consistent with behavior observed in many Asian economies. These findings underscore the significance of both structural and policy-driven enablers of market access in emerging countries.

In the context of U.S. issuers, Miller and Puthenpurackal (2005) and Tawatnuntachai and Yaman (2007) identify diverse firm-specific motivations for foreign issuance, including reputational signaling, pursuit of scale efficiencies, and unfavorable domestic conditions. Ahwireng-Obeng and Ahwireng-Obeng (2022) reinforce the role of firm-level fundamentals—such as size, leverage, and firm age—in shaping bond issuance behavior in African emerging markets.

Recent empirical studies have also examined the determinants of credit spreads. Q. Wang et al. (2019) identify credit ratings, issuance size, and currency denomination as key variables, particularly in emerging and green bond markets. Badoer and James (2016) introduce the concept of “gap-filling behavior,” showing how firms adjust issuance maturity and volume in response to supply shortages in government securities. These behaviors reflect forward-looking pricing strategies. Bhattacharyay (2013) emphasizes the importance of institutional depth and exchange rate stability in enabling efficient pricing and issuance across Asian markets.

Geopolitical risks and domestic institutional quality have also emerged as influential factors. Mertzanis and Tebourbi (2025) find a nonlinear relationship between geopolitical uncertainty and green bond issuance, suggesting that political risk perceptions affect both the timing and location of issuance.

However, relatively few empirical studies have examined Japanese firms, particularly regarding their response to cross-border pricing incentives. While limited, recent studies have begun to explore credit spread dynamics in the Japanese context. Okimoto and Takaoka (2024) examine the role of ESG performance in shaping credit spreads among Japanese firms, highlighting the pricing relevance of non-financial attributes. Huang et al. (2025) investigate the structural determinants of credit spreads across advanced economies, including Japan, and identify persistent gaps between theoretical predictions and observed market behavior. These findings underscore the need for further firm-level analysis of issuance decisions under varying pricing conditions. Japan’s unique financial system—characterized by low yields, a dominant banking sector, and persistent investor home bias—has not been systematically analyzed in relation to market choice. Although Japan remains a key player in global capital markets, limited evidence exists on how Japanese firms navigate the tradeoff between domestic stability and foreign pricing opportunities.

This study addresses this empirical gap by linking expected credit spread conditions with issuance market decisions. It incorporates a comprehensive set of firm-level (e.g., credit ratings and financial leverage) and macro-level (e.g., interest rate differentials and global volatility) variables, and employs firm-level fixed effects panel regressions to control for unobserved heterogeneity.

2.3. Hypothesis Development

This study develops hypotheses linking expected credit spread conditions to firms’ issuance market choices. Specifically, it incorporates firm-level and macro-level variables and applies a fixed effects panel regression framework to examine how Japanese firms respond to pricing incentives across domestic and foreign bond markets. Japan’s unique financial system—marked by low yields, a dominant banking sector, and persistent investor home bias—creates distinct constraints and incentives that may shape firms’ market selection behavior, yet remains underexplored in the existing literature. Although Japan remains a key player in global capital markets, relatively few studies have explored how Japanese firms navigate the tradeoff between domestic stability and foreign pricing opportunities. This study builds on the existing literature by linking expected credit spread conditions with issuance market decisions; incorporating a comprehensive set of firm-level (e.g., credit ratings and financial leverage) and macro-level (e.g., interest rate differentials and global volatility) variables; and employing firm-level fixed effects panel regressions to control for unobserved heterogeneity.

By addressing the intersection of pricing expectations and market choice, this study contributes to both the theoretical understanding of capital market segmentation and the ongoing policy discussions on corporate funding behavior under yield suppression and international arbitrage.

Although prior studies have examined various aspects of corporate bond issuances—including market choice, credit pricing, and institutional influences—these strands of the literature have often evolved separately. Studies on issuance market selection typically emphasize firm characteristics or macro-level constraints, whereas studies on credit spreads focus on pricing determinants in isolation. Consequently, little attention has been paid to how expected pricing conditions may drive strategic issuance behavior across markets. To address this gap, this study aims to investigate how firms incorporate expected credit spread differences into their bond issuance decisions across domestic and foreign markets.

Drawing on these theoretical and contextual foundations, the following hypotheses are proposed to capture how Japanese firms respond to pricing incentives and risk perceptions in their issuance decisions. Building on the prior literature and the institutional context of Japanese bond markets, this study proposes the following hypotheses:

H1. 

Firms are more likely to issue foreign bonds when expected credit spreads are narrower than those for domestic bonds.

H2. 

Higher firm leverage is associated with a greater likelihood of domestic issuance owing to higher perceived credit risk abroad.

H3. 

Heightened global financial volatility (e.g., higher U.S. VIX [USVIX]) reduces the likelihood of foreign issuance as investors demand higher risk premiums.

These hypotheses suggest that issuance location is chosen in anticipation of pricing outcomes and risk perceptions, not in isolation. Figure 1 presents a conceptual framework that summarizes the hypothesized relationships developed in this section. It illustrates how issuer characteristics (e.g., leverage and rating) and market conditions (e.g., USVIX and TOPIX) jointly influence expected credit spreads, which, in turn, affect firms’ decisions to issue bonds in domestic or foreign markets.

3. Methodology

This section presents the empirical framework for testing key hypotheses regarding Japanese firms’ issuance behaviors in the domestic and foreign bond markets. The analysis proceeds in two stages. First, it examines pricing conditions associated with each bond issuance via credit spread regressions. Then, it analyzes how these pricing conditions, along with firm characteristics and market variables, systematically affect the firm’s foreign bond issuance decisions.

3.1. Econometric Framework

To test the study hypotheses, we estimate the following models using firm-level panel data with fixed effects:

• Step 1: Credit Spread Determination

Following Collin-Dufresn et al. (2001), we define the credit spread of corporate bonds using yield, leverage, market, and VIX. The first model examines the determinants of credit spreads at the time of issuance:

CS_it = β₁(ForeignDummy)_it + β₂X_it + α_i + δ_t + ε_it

where CS_it represents the credit spread of firm i in year t, calculated as the difference between the yield on the firm’s corporate bond and that on a matched government bond with comparable maturity and currency denomination. Specifically, we use Japanese Government Bonds, which serve as benchmark risk-free instruments in the domestic market, with similar maturities as benchmarks for isolating firm-specific credit risk components. This matching ensures that the spread reflects differences in creditworthiness, rather than term structure effects.

The indicator variable (ForeignDummy)_it equals 1 if a bond is issued in a foreign market during year t and 0 otherwise. This distinction enables the estimation of systematic pricing differences between domestic and international bond markets.

The control vector X_it includes firm-level financial characteristics and market-level conditions. The firm-level controls include credit ratings and leverage ratios. We convert Moody’s credit ratings into a 1–10 numeric scale, in which higher values reflect lower credit quality, and all firms in the sample remain within the investment-grade range. Market-level variables include the Japan Volatility Index (JPVIX), which reflects domestic market uncertainty, and the U.S. Volatility Index (USVIX), which captures global financial risk sentiment in each Japanese and foreign market, with the Tokyo Stock Price Index (TOPIX), a benchmark for Japanese equity market performance and the Standard & Poor’s 500 Composite Stock Price Index (S&P 500), a key indicator of U.S. equity market conditions, being represented similarly. These controls capture the time-varying risks and macro-financial conditions that may jointly influence credit spreads and issuance behavior.

To account for unobserved heterogeneity, the model incorporates firm fixed effects α_i to absorb time-invariant firm characteristics and year fixed effects δ_t to control for macroeconomic shocks and global financial conditions common across all firms.

Interaction terms, such as ForeignDummy × Leverage or ForeignDummy × USVIX, are included to examine heterogeneity in spread responses.

These interaction terms allow us to assess whether the effect of foreign bond issuance on credit spreads varies depending on firm-specific financial risk and market volatility. Specifically, the coefficient on ForeignDummy × Leverage captures whether highly leveraged firms face higher spreads when issuing foreign bonds. Similarly, ForeignDummy × USVIX reflects whether foreign issuers are more sensitive to U.S. market uncertainty.

By incorporating these terms, we move beyond average treatment effects and uncover conditional dynamics that are critical for understanding the cost of foreign debt issuance under varying corporate and macroeconomic conditions. These terms collectively test how firm-specific traits and market-level signals jointly shape credit spread outcomes, thereby capturing the dynamic nature of market selection under segmented conditions.

• Step 2: Market Choice Model

Following the probabilistic frameworks used by Zhu (2013) and Ahwireng-Obeng and Ahwireng-Obeng (2022), we model the likelihood of foreign bond issuances as a function of predicted credit spreads, firm characteristics, and fixed effects. Subsequently, using a fixed-effects logit model, we test how pricing conditions and firm characteristics influence decisions to issue in foreign markets.

Prob(ForeignIssue_it = 1) = f (^CS_it, X_it, α_i, δ_t)

In this analysis, the dependent variable is a binary indicator that takes the value of 1 if a firm issues a bond in a foreign market in year t and 0 otherwise. The key explanatory variable, ^CS_it, represents the firm-specific expected credit spread, which can be either directly estimated or proxied using observable financial and market conditions. This estimation utilizes either a fixed effects logit model or linear probability model, both incorporating firm and year fixed effects (α_i and δ_t, respectively) to control for unobserved heterogeneity. To ensure the robustness of the results, we employ additional specifications, such as alternative constructions of key variables and interaction terms.

3.2. Data and Variable Construction

This study utilizes a firm-level dataset spanning 1 January 2010 to 31 December 2019. The sample period encompasses key macro-financial phases, including the U.S. Federal Reserve’s third round of quantitative easing (QE3), a monetary policy tool aimed at stimulating the economy through large-scale asset purchases.

To avoid confounding effects associated with the onset of the COVID-19 pandemic, the dataset was deliberately restricted to pre-pandemic years. Recent research has documented substantial shifts in economic behavior and financial market dynamics following the emergence of the pandemic (Chetty et al., 2024). To ensure analytical consistency and isolate structural patterns unaffected by these disruptions, this study excludes data from 2020 onward.

Although these macroeconomic episodes primarily impact foreign bond issuance, the number of foreign bond observations during these sub-periods is limited and, thus, unlikely to bias the main results. To ensure comparability under pricing conditions, we focus exclusively on fixed-rate corporate bonds, excluding floating-rate notes. To calculate credit spreads, each corporate bond is matched with a government bond of the same currency and maturity. The final sample comprises 13 Japanese firms that issued both domestic and foreign bonds during the study period. These firms represent a diverse range of major Japanese institutions across the financial, infrastructure, and industrial sectors. These firms include Daiwa Securities Co., Ltd., Tokyo, Japan; Development Bank of Japan Inc., Tokyo, Japan; Japan Bank for International Cooperation, Tokyo, Japan; Central Japan Railway Company, Nagoya, Japan; Mitsubishi Corporation, Tokyo, Japan; Mitsui Fudosan Co., Ltd., Tokyo, Japan; Mizuho Bank, Ltd., Tokyo, Japan. MUFG Bank, Ltd., Tokyo, Japan; Central Nippon Expressway Company, Ltd., Nagoya, Japan.; Nippon Telegraph and Telephone Corporation, Tokyo, Japan; Sumitomo Corporation, Tokyo, Japan; Sumitomo Mitsui Trust Bank, Ltd., Tokyo, Japan; and Takeda Pharmaceutical Company, Ltd., Tokyo, Japan. All issuers maintained investment-grade credit ratings (Moody’s Baa or higher) throughout the sample period.

This dataset combines multiple sources to capture firm-, bond-, and market-level information. Data on domestic bond issuances are obtained from the Japan Securities Dealers Association, Tokyo, Japan, specifically its Public and Corporate Bond Issue Lists. Information on foreign bond issuances is collected from Moody’s official website, while government bond yields used for spread calculations are sourced from the Ministry of Finance, Tokyo, Japan.

Firm-level financial indicators, including leverage ratios and balance sheet data, are retrieved from the EOL database provided by Pronexus Inc., Tokyo, Japan. For equity market conditions, data on the TOPIX are obtained from Yahoo! Finance, Japan, Tokyo, Japan. Additionally, the Nikkei 225 Volatility Index, used as a proxy for domestic risk sentiment, is sourced from Nikkei Indexes. To measure foreign market conditions, we use the S&P 500 and VIX. Both datasets are obtained from Yahoo! Finance, New York, United States. Credit spreads are calculated by subtracting the matched government bond yield from the corporate bond yield at issuance. To ensure precise matching, each corporate bond is paired with a government bond of the same currency and maturity, using yield data from the bond issuance date. For example, a five-year corporate bond issued on a given date is matched with the yield of a five-year government bond on that same date. If a government bond with the exact same maturity was not available on the issuance date, the observation was excluded from the sample to avoid interpolation and maintain consistency. As a result, all matched pairs in the final sample have identical maturities. Financial ratios and macro-financial indicators are merged by firm and year to construct a balanced panel for fixed-effects regression analysis. The Japanese bond market exhibits distinct institutional characteristics that may influence issuance behavior and pricing dynamics. According to the Bank of Japan (2025), recent developments—such as the expansion of the Securities Lending Facility and efforts to enhance JGB market liquidity—reflect ongoing structural adjustments aimed at improving market functioning. These features provide important context for interpreting firm-level issuance decisions and observed credit spread patterns. All statistical analyses were performed using Stata/BE version 19 (StataCorp LLC, College Station, TX, USA). To enhance clarity regarding the operational definitions and measurements of variables used in our econometric framework, we provide a summary table below (

Table 1. Variable summary.

Variable Name	Description	Type	Source/Notes
CS	Credit spread at issuance (corporate bond yield minus matched government bond yield)	Dependent (Step 1)	Calculated from bond issuance data and government bond yields
Foreign Dummy	Indicator for foreign bond issuance (1 = foreign, 0 = domestic)	Independent	Categorical variable identifies issuance location
Credit rating	Firm’s credit rating (converted to numeric scale: 1 = highest, 10 = lowest)	Independent	Moody’s ratings, investment-grade only
Leverage	Firm’s leverage ratio (total liabilities/total assets)	Independent	Firm-level financial data from EOL database
JPVIX	Japan Volatility Index	Market control	Proxy for domestic market uncertainty
USVIX	U.S. Volatility Index	Market control	Proxy for global market uncertainty
TOPIX	Tokyo Stock Price Index	Market control	Proxy for domestic equity market conditions
S&P 500	Standard & Poor’s 500 Index	Market control	Proxy for foreign equity market conditions
Prob(Foreign Issue)	Probability of foreign bond issuance	Dependent (Step 2)	Modeled using fixed-effects logit regression
^CS	Expected credit spread (predicted from Step 1 regression)	Independent (Step 2)	Used in Step 2 to model market choice
JGB	Yield on matched Japanese Government Bond	Market control	Used to calculate domestic credit spread
UST	Yield on matched U.S. Treasury Bond	Market control	Used to calculate foreign credit spread
Maturity year	Bond maturity in years	Control variable	Bond-level data
Interest rate	Coupon rate of the bond	Control variable	Bond-level data

). This includes all firm-level financial characteristics and market-level conditions employed in Steps 1 and 2.

Table 2. Summary statistics.

	Obs	Mean	Std. Dev	Min	Max	Unit/Notes
CS	465	0.325	0.358	−0.001	2.410	Credit spread index
Foreign Dummy	471	0.312	0.464	0.000	1.000	Dummy Variable
Leverage rate	471	0.825	0.120	0.233	0.967	Ratio
Credit rating	471	4.548	1.104	2.000	10.000	Rating scale (1–10)
Issuance (JPY)	324	26,795.68	19,622.70	5000	120,000	Millions of Yen
Issuance (USD)	147	895.995	493.509	53	2500	Millions of USD
Interest rate	471	1.140	1.106	0.001	5.000	%
Maturity year	471	7.531	5.901	2.000	40.000	Year
JGB	471	0.301	0.467	−0.376	2.128	Yield spread (bps)
UST	442	1.778	0.829	0.240	4.520	Yield spread (bps)
TOPIX	471	1237.612	337.209	716.840	1911.070	Index level
SP500	471	1907.950	556.850	1064.880	3096.630	Index level
JPVIX	471	21.797	4.897	13.880	36.600	VIX
USVIX	471	16.317	4.917	9.510	37.810	VIX

summarizes the main variables used in this analysis. In total, our initial sample comprises 471 firm–year observations. Of these, six observations are excluded from the credit spread regressions due to missing UST for the matched maturity, leaving 465 observations for the CS analysis. An additional 23 observations lack UST on the domestic issuance date, resulting in a final sample of 442 observations for the fixed-effects logit models. The credit spread index (CS) has an average of 0.33 and a standard deviation of 0.369, indicating moderate variability among firms. The sample comprises domestic and international issuers, with approximately 31% classified as foreign (ForeignDummy). The average leverage ratio is 0.83, suggesting a balanced capital structure. The mean credit rating is 4.55 on a 1–10 scale. The average bond issuance amounts are 26,796 million yen for yen-denominated bonds and USD 896 million for U.S. dollar-denominated bonds, indicating significant variation among firms. The interest rates range from 0.001% to 5%, averaging 1.14%, while bond maturities average 7.53 years, ranging from 2 to 40 years. Yield spreads for Japanese Government Bonds (JGBs) and U.S. Treasuries (USTs), which represent benchmark sovereign yields in their respective markets, average 0.30 and 1.77 basis points, respectively. The stock market indices (TOPIX and S&P 500) average 1238 points and 1908 points, respectively, while the volatility indices (JPVIX and USVIX) average 21.8 points and 16.3 points, respectively, indicating general market stability during the sample period.

Table 3. Correlation matrix.

	Foreign Dummy	CS	Leverage Rate	Credit Rating	JGB	UST	TOPIX	SP500	JPVIX	USVIX
Foreign Dummy	1
CS	0.798	1
Leverage rate	0.022	−0.051	1
Credit rating	0.200	0.317	−0.186	1
JGB	−0.250	−0.183	−0.135	−0.329	1
UST	−0.055	−0.120	−0.278	0.094	0.536	1
TOPIX	0.216	0.154	−0.181	0.520	−0.460	0.312	1
SP500	0.213	0.164	−0.209	0.577	−0.548	0.234	0.924	1
JPVIX	−0.080	−0.019	0.107	−0.319	0.308	−0.089	−0.407	−0.524	1
USVIX	−0.131	−0.067	−0.033	−0.225	0.275	−0.050	−0.471	−0.443	0.437	1

details the pairwise correlations among the key independent variables. The CS exhibits a strong positive correlation with ForeignDummy (r = 0.798), implying that foreign issuances generally have a wider spread. CS shows weak correlations with other firm-level and market variables. Credit ratings positively correlate with CS (r = 0.317), suggesting that higher-rated firms may be associated with wider spreads, potentially reflecting market segmentation or issuer characteristics. Credit ratings negatively correlate with JGB yields (r = −0.329) and equity indices (TOPIX: r = −0.460), while positively correlating with SP500 (r = 0.577), indicating nuanced relationships between firms’ creditworthiness and market conditions. The leverage rate exhibits minimal correlations with all other variables (e.g., r = −0.051 with CS), suggesting that multicollinearity from capital structure indicators is unlikely. Among market variables, JGB and UST yields are moderately correlated (r = 0.536). Equity indices and volatility measures display strong inverse relationships (e.g., SP500 and USVIX: r = −0.443), consistent with conventional market dynamics. Overall, the correlations remain within acceptable limits, with no variable pairs surpassing conventional multicollinearity thresholds. Furthermore, the variance inflation factors for all independent variables are below the standard threshold of 10, with an average of 3.85, indicating no significant concerns about multicollinearity.

4. Results and Analysis

4.1. Results

This section details the empirical findings from a two-step estimation approach, exploring the link between foreign bond issuances and corporate credit spreads. To empirically test our hypotheses, we first estimate credit spreads using fixed-effects panel regressions. Then, we assess foreign issuance probabilities via fixed-effects logit models. Both approaches control for unobserved heterogeneity through firm and year fixed effects.

To empirically evaluate the hypotheses presented in Section 3.1, we employ a two-stage regression approach. First, a linear fixed-effects model is used to estimate credit spreads at issuance, accounting for both firm- and market-level variables. Subsequently, a fixed-effects logit model is used to assess the likelihood of foreign issuances based on credit spreads and financial characteristics. The results, summarized in Table 3 and

Table 4. Regression results (Step 1).

	Model
	1			2
Variable	Coef.	SE	Sig	Coef.	SE	Sig
Credit Spread (CS)
Foreign Dummy	0.621	0.100	***	0.617	0.103	***
Leverage	0.718	0.475		−0.166	0.232
Credit rating	0.092	0.080		0.087	0.037	**
TOPIX				0.000	0.000	*
S&P 500				0.000	0.000
JPVIX				0.007	0.002	**
USVIX				−0.005	0.003
JGB				−0.052	0.038
UST				0.037	0.033
Constant	−0.872	0.590		0.402	0.356
Observations	465			442
R-squared	0.772			0.673
Year FE	Yes			Yes

Note: All regressions include year fixed effects. Standard errors are clustered at the issuer level. *** p < 0.01, ** p < 0.05, * p < 0.1.

, strongly support H1 and H3, while partially confirming H2.

Table 4 (Model 1) reveals that the coefficient for ForeignDummy is positive and significant (β = 0.621, p < 0.01), indicating that foreign bond issuance is generally associated with wider credit spreads. This finding supports H1, suggesting that firms may face higher borrowing costs when issuing abroad. Model 2 includes additional market-level controls, and the significance of ForeignDummy remains (β = 0.617, p < 0.01), reinforcing the robustness of this relationship. Volatility measures show consistent effects across both models: JPVIX positively correlates with spreads (β = 0.007, p < 0.01), suggesting that domestic market uncertainty increases borrowing costs, while USVIX has a negative coefficient (β = −0.005), does not reach conventional significance thresholds, indicating a weaker link between global volatility and spread compression. Credit ratings are positively associated with credit spreads in Model 2 (β = 0.087, p < 0.05), which partially supports H2 and may reflect issuer segmentation or rating inflation. Other market variables, including equity indices and sovereign yields, show minimal effects. The R-squared values (0.772 for Model 1 and 0.673 for Model 2) indicate strong explanatory power, and year fixed effects are included in all specifications.

Table 5. Fixed effect logit model (Step 2).

Dependent Variable: Foreign Bond Issuance Dummy
	Model
	1			2-1			2-2
Variable	Coef.	SE	Sig	Coef.	SE	Sig	Coef.	SE	Sig
CS	36.005	9.143	***	49.106	16.488	***	49.106	16.488	***
Leverage	62.002	54.784		128.72	62.713	**	128.72	62.713	**
Credit rating	−2.35	1.24	*	−4.739	1.902	**	−4.739	1.902	**
TOPIX				0.008	0.003	**	0.008	0.003	**
S&P 500
JPVIX
USVIX				−0.207	0.215		−0.207	0.215
JGB
UST
Observations	465			465			465
LR chi2(3)	365.08		***	374.16		***	374.16		***
Number of Issuers	13			13			13
Fixed Effects	Issuer-level (conditional logit)			Issuer-level (conditional logit)			Issuer-level (conditional logit)
	Model
	2-3			2-4			2-5
Variable	Coef.	SE	Sig	Coef.	SE	Sig	Coef.	SE	Sig
CS	35.051	9.019	***	45.217	14.181	***	48.878	15.552	***
Leverage	68.909	55.363		115.401	50.357	**	122.548	57.203	**
Credit rating	−2.705	1.411	*	−4.152	1.609	***	−3.23	1.608	**
TOPIX				0.009	0.007
S&P 500				−0.001	0.003
JPVIX							−0.31	0.15	**
USVIX							−0.275	0.217
JGB	−2.229	3.712
UST	1.445	1.197
Observations	442			465			465
LR chi2(3)	352.77		***	373.29		***	372.19		***
Number of Issuers	13			13			13
Fixed Effects	Issuer-level (conditional logit)			Issuer-level (conditional logit)			Issuer-level (conditional logit)

Note: All regressions are estimated using conditional fixed-effects logistic models. R-squared is not reported because it is not defined for fixed-effects logit models. Standard errors are clustered at the issuer level. *** p < 0.01, ** p < 0.05, * p < 0.1.

presents the results from six conditional fixed-effects logit models examining the determinants of foreign bond issuance. All models use issuer-level fixed effects and cluster standard errors at the firm level. Model 1 includes core firm-level variables, while Models 2-1 through 2-5 incrementally incorporate market-level indicators to test the robustness of H1–H3.

Across all specifications, credit spreads (CS) consistently exhibit a strong positive association with the likelihood of foreign issuance (e.g., Model 1: β = 36.005, p < 0.01; Model 2-5: β = 48.878, p < 0.01), reaffirming H1. This suggests that firms facing wider domestic spreads are more likely to seek international financing. The economic magnitude is substantial, indicating that even moderate increases in CS can significantly shift firms’ market choice behavior.

Leverage shows a positive and significant effect in several models (e.g., Model 2-1: β = 128.72, p < 0.05; Model 2-5: β = 122.548, p < 0.05), which contrasts with the initial expectation under H2. This may reflect strategic behavior among highly leveraged firms that pursue foreign markets to diversify funding sources or mitigate domestic constraints. Credit ratings consistently show a negative relationship with foreign issuance (e.g., Model 2-1: β = −4.739, p < 0.05), suggesting that lower-rated firms are more likely to issue abroad, possibly due to limited domestic access or rating arbitrage.

Market-level variables yield mixed results. Domestic equity performance (TOPIX) is positively associated with foreign issuance in Model 2-1 (β = 0.008, p < 0.05), indicating that favorable domestic conditions may encourage international financing. Volatility measures show limited significance overall, though JPVIX is negatively associated with foreign issuance in Model 2-5 (β = −0.310, p < 0.05), partially supporting H3. This suggests that firms may be deterred from foreign markets during periods of domestic uncertainty, while global volatility (USVIX) does not exhibit consistent effects.

Overall, the results strongly support H1, offer partial support for H3, and provide a nuanced view of H2. The expanded model set enhances robustness and highlights the complex interplay between firm fundamentals and market conditions in shaping international financing decisions. For instance, based on Model 2-1, which employs a fixed-effects logit specification, when the credit spread index (CS) increases from 0.30 to 0.60—within the observed range—the predicted probability of foreign issuance rises from approximately 32.6% to 95.4%. This substantial marginal effect underscores the economic significance of spread sensitivity in firms’ financing decisions, suggesting that even moderate changes in credit conditions can meaningfully influence market choice.

Taken together, these results confirm the theoretical expectations, demonstrating that firms actively respond to credit-pricing signals and risk conditions when selecting issuance venues. The consistent impact of spreads and volatility across the models highlights the strategic nature of market selection within a segmented financial environment.

4.2. Robustness

To evaluate the robustness of our primary conclusions, we use a range of alternative models, including interaction terms and subsample analyses. These expanded models enable us to verify whether the relationships identified in Section 4.1 remain consistent under different assumptions and estimation methods. Specifically, we investigate how firm characteristics and market conditions interact with issuance location to affect credit spread outcomes, further supporting H1–H3.

Model 1 (

Table 6. Regression results across interaction models (Models 1–4).

	Model
Variable	1	2		3-1		3-2		3-3		3-4		4
Credit Spread (CS)
Leverage rate	1.542	0.785		0.985	**	0.581		0.961	*	1.008	*	0.856
Credit rating	0.064	0.097		0.084		0.094		0.083		0.09		0.085
Foreign Dummy (Foreign)		0.577	*	0.555	***	0.534	***	1.026	***	1.087	***	0.160
Foreign × Credit rating		−0.017
Foreign × Leverage rate		0.147
Foreign × JGB				0.405	***
Foreign × UST						0.051
Foreign × TOPIX								−0.003	**
Foreign × S&P 500										0.000	***
Foreign × JPVIX												0.019	***
Foreign × USVIX												0.003
Constant	−1.2389	−0.951		−1.05	*	−0.772		−1.026		−1.094	*	−0.932
Observations	465	465		465		442		465		465		465
R-squared	0.329	0.77		0.717		0.774		0.787		0.794		0.792
F-statistic	1.46	12.01	***	20.12	***	18.17	***	16.83	***	18.99	***	13.25	***

Note: All regressions include year fixed effects. Standard errors are clustered at the issuer level. *** p < 0.01, ** p < 0.05, * p < 0.1.

) acts as a baseline, confirming that crucial firm-level factors such as credit rating, leverage ratio, and bond maturity significantly influence credit spreads at issuance. These findings generally align with the predictions of H1 and H2, reinforcing that cost expectations and financial structure influence pricing behavior.

To enhance our understanding of how firm characteristics impact spread formation in issuance markets, we include interaction terms between issuer traits and the ForeignDummy variable. This method enables us to examine whether the influence of leverage, credit ratings, and other features differs systematically based on issuance location, thereby providing a more detailed examination of H2 and strengthening the reliability of our previous results. To further address potential endogeneity concerns—particularly the possibility of reverse causality between issuance location and credit spreads—we conduct a series of robustness checks. These include alternative model specifications and subsample analyses designed to test the stability of our findings. The results remain consistent with our primary estimates, reinforcing the robustness of H1–H3 and suggesting that the observed relationships are not driven by endogenous selection into foreign markets. Model 2 introduces the interaction terms between ForeignDummy and the issuer characteristics. A positive and significant interaction between ForeignDummy and credit ratings indicates that highly rated firms issuing in foreign markets tend to encounter wider spreads. This may reflect global investors’ reputational risk pricing or transparency gaps. However, the negative interaction with leverage suggests that firms with higher debt experience a narrower spread abroad, potentially supporting H2, which suggests that leverage limits foreign market access because of increased credit concerns.

Models 3-1 to 3-4 incorporate interactions between issuance location and market conditions, such as government bond yields and equity performance. The negative and statistically significant interaction between ForeignDummy and UST yields aligns with H1, indicating that favorable pricing conditions abroad (i.e., lower benchmark yields) result in narrower spreads for foreign issuers. Similarly, the negative coefficients of the ForeignDummy × SP500 and TOPIX interactions imply that improved equity performance compresses spreads in a foreign context, further highlighting the pricing sensitivity predicted by H1.

Finally, Model 4 examines interactions with forward-looking risk indicators. The positive and highly significant coefficient of ForeignDummy × USVIX confirms that global volatility increases spread levels for foreign bonds, consistent with H3, which anticipates that rising uncertainty prompts investors to demand higher premiums. Conversely, the interaction with the JPVIX is insignificant, suggesting asymmetric effects of domestic and global uncertainty on credit pricing. Across all specifications, the R-squared values improve from 0.329 in the baseline model to above 0.794 in extended models, indicating that accounting for issuer–market interactions enhances explanatory power. These results reinforce the strategic nature of market selection and confirm that spread formation varies systematically according to firm type and market environment, which is consistent with our core hypotheses.

5. Discussion

Our regression analysis (Table 4, Model 1) reveals a notable positive correlation between credit spread and the probability of bond issuance. This finding aligns with those of Badoer and James (2016), who contend that market-based costs such as credit risk are crucial for firms’ capital structure decisions. While they primarily concentrated on long-term debt issuances among U.S. firms, we expand the scope by including market volatility indices and international exposure.

The impact of the macroeconomic variables is also apparent. As illustrated in Table 4, the Coefficient for JPVIX is positive and statistically significant, suggesting that heightened domestic market volatility increases credit spreads at issuance. In contrast, USVIX exhibits a negative but statistically insignificant coefficient, indicating that global volatility may exert a more muted or complex influence on bond issuance behavior. These findings partially support the perspectives of Lo Duca et al. (2016), who highlight the sensitivity of global bond issuances to changes in monetary policy during quantitative easing.

Building on this macro-financial lens, a complementary perspective is offered by Fatmawatie et al. (2024), who examine the determinants of Indonesian government bond yields using GARCH-type models. Their findings reveal that macroeconomic variables such as domestic interest rates, foreign exchange reserves, and exchange rates significantly influence sovereign bond yields, while inflation shows no consistent effect. Although their study focuses on sovereign debt in a developing market context, the results underscore the broader relevance of macroeconomic conditions in shaping bond pricing and issuance behavior.

In contrast, our study emphasizes firm-level issuance decisions within a segmented market, highlighting how volatility indices and credit spreads interact with corporate characteristics. This juxtaposition suggests that while macro-level indicators are crucial for sovereign debt pricing, firm-level strategies and market segmentation play a decisive role in corporate bond issuance.

Table 5 (Model 1 and extended specifications) highlights the consistent and statistically significant impact of credit spread on the likelihood of foreign bond issuance, affirming its central role in pricing decisions. This robustness across models underscores the importance of market-based risk indicators in segmented markets such as Japan.

In addition, firm-level characteristics such as leverage and credit rating also show significant effects. Higher leverage is positively associated with issuance probability, while stronger credit ratings are negatively correlated, suggesting that firms with lower perceived risk may face less urgency to issue foreign bonds. These findings complement Zhu (2013), who emphasizes managerial discretion and strategic motives, but our results indicate that systemic financial indicators may exert a more consistent influence under Japan’s market conditions.

Notably, the effect of volatility indices diverges from earlier models. JPVIX shows a negative and statistically significant coefficient in Model 2-5, contrasting with its positive effect in Table 4. These results suggest that domestic volatility may deter foreign bond issuance when firm-level fixed effects are accounted for, highlighting the nuanced role of macro-financial conditions depending on model specification. Moreover, the importance of the ForeignDummy (Table 6, Model 1 to 3-3) reinforces that firms with international ties are more inclined to issue bonds, aligning with the findings of Ahwireng-Obeng and Ahwireng-Obeng (2022) and Mizen and Tsoukas (2014), who observe similar patterns in emerging and Asian markets. This finding indicates that global integration eases access to capital markets. Our interaction models further reveal that sensitivity to market conditions varies significantly based on firm characteristics. Specifically, the positive and statistically significant coefficient for the interaction between Foreign Dummy and Leverage Rate (Model 3-2) suggests that highly leveraged firms with international exposure are more responsive to credit spread dynamics. Additionally, foreign interactions with market indices such as UST, TOPIX, and S&P 500 (Models 3-3 and 3-4) show significant effects, indicating that firms’ issuance behavior is shaped not only by domestic volatility but also by foreign asset market conditions.

These findings have important practical implications for corporate financing. The consistent and significant impact of credit spreads on the likelihood of bond issuance (Table 4, Table 5 and Table 6) highlights the need for firms to monitor market conditions closely when contemplating debt issuance. Firms operating in international markets—evidenced by the positive coefficient of ForeignDummy—might gain from developing adaptable funding strategies that consider foreign market dynamics. Furthermore, because credit ratings consistently show predictive power across the models (e.g., Table 4 and Table 6), enhancing transparency and maintaining strong credit profiles may improve access to bond markets, particularly during times of financial uncertainty. In short, our findings help clarify how pricing expectations shape firms’ strategic financing decisions, particularly in segmented bond markets. While the findings illuminate how credit spreads influence issuance location decisions, this study has several limitations. First, the sample focuses solely on investment-grade Japanese firms, which may limit its generalizability to smaller or lower-rated entities. Broadening the analysis to include non-investment-grade issuers and small and medium enterprises could provide deeper insights into the role of perceived credit risk. Second, the relatively small number of firms issuing in both domestic and foreign markets limits the statistical power and may introduce selection bias.

Future research could benefit from utilizing multi-country datasets or higher-frequency observations to validate spread sensitivity patterns in diverse institutional contexts. Finally, this study does not capture post-issuance outcomes, such as secondary market performance or refinancing behavior, which could further reflect pricing and market choice dynamics. Investigating these post-issuance behaviors would complement our understanding of long-term financial strategies and enhance the implications of spread-based decision-making. Building on the limitations noted above, Collin-Dufresn et al. (2001) examine credit spread changes in the secondary market and find that a large portion of spread variation is driven by a common factor unrelated to firm-specific credit risk or liquidity proxies. Their findings underscore the complexity of post-issuance dynamics and suggest that local supply–demand shocks may play a dominant role. Incorporating such secondary market behavior into future research could enrich our understanding of how initial pricing expectations evolve over time and influence refinancing or trading decisions.

Beyond academic implications, our findings also offer practical insights for policymakers and regulators. In a persistently low-yield domestic environment, Japanese firms appear to engage in strategic market selection, responding to pricing signals and volatility across borders. This behavior reflects a form of capital market arbitrage that may influence outbound capital flows. From a regulatory perspective, facilitating access to international bond markets—through harmonized disclosure standards, enhanced credit rating transparency, or reduced issuance frictions—could support firms’ financing flexibility and improve the efficiency of cross-border capital allocation.

Furthermore, the findings offer actionable insights for monetary authorities and financial regulators. The consistent influence of credit spreads on issuance behavior suggests that yield curve control policies should account for firm-level heterogeneity, as uniform rate targeting may produce uneven incentives across sectors. Additionally, the contrasting effects of JPVIX and USVIX highlight the importance of managing currency risk and global volatility exposure, particularly for internationally active firms. As Mizen and Tsoukas (2014) emphasize, firms operating across borders face distinct financing constraints and risk profiles, which require tailored regulatory responses. In this context, reforms that support flexible hedging strategies, cross-border issuance, and market segmentation awareness could enhance the effectiveness of monetary policy transmission and improve overall market functioning. These considerations are especially relevant in the Japanese bond market, where firms increasingly respond to global pricing signals and volatility dynamics.

In summary, our findings carry clear, actionable recommendations for key stakeholders. First, regulators may consider adjusting yield-curve control frameworks and reducing issuance frictions to foster smoother cross-border capital flows. Second, investors can enhance portfolio resilience by monitoring anticipated credit spread differentials and global volatility indices when allocating to domestic versus foreign bonds. Finally, corporate issuers are advised to develop dynamic funding strategies that incorporate real-time spread forecasts and flexibility in currency denomination to optimize borrowing costs under segmented market conditions.

6. Conclusions

This study explores how Japanese firms balance domestic and international bond issuance by investigating the factors that determine credit spreads and their interactions with market selection. Using firm-level panel data from 2010 to 2019, we find that credit spreads, market volatility, and firm-specific characteristics significantly affect borrowing costs and issuance location choices.

The fixed-effects regression results indicate that while foreign bond issuances are generally associated with higher credit spreads, firms strategically select foreign markets when they expect more favorable pricing conditions. Interaction models further demonstrate that issuer-specific traits (such as leverage and credit ratings) and market conditions (such as U.S. volatility and stock performance) influence the spread sensitivity of foreign bonds. These results provide empirical confirmation and offer theoretical refinement. By showing that firms respond to pricing signals in segmented markets, our findings support models of strategic market selection under informational frictions. The observed sensitivity to volatility and spread expectations suggests that issuance behavior reflects forward-looking optimization, consistent with dynamic capital structure theories.

These findings highlight that bond issuance decisions are not fixed but are dynamic choices influenced by expected pricing outcomes and global financial signals. This study provides new insights into how Japanese firms engage in capital market arbitrage by adjusting their strategies based on expectations and uncertainties.

By linking credit spread formation with market selection, this study contributes to the ongoing discourse on international corporate finance, highlighting firms’ strategic behavior within segmented financial systems. These insights may also inform policy discussions on how pricing incentives and market structures shape cross-border financing decisions, particularly amid persistently low domestic yields and heightened global volatility.

Nevertheless, this study has some limitations. The sample is restricted to large Japanese firms with access to both domestic and foreign bond markets and publicly available Moody’s credit ratings. Consequently, all issuers in the sample are investment-grade. While this reflects the characteristics of firms active in cross-border issuance, it may limit the generalizability of the findings. Additionally, we acknowledge that unobserved firm-level heterogeneity—such as managerial preferences or internal capital constraints—may influence issuance decisions in ways not fully captured by our model.

Future research should examine the issuance behavior of small and medium-sized enterprises, which typically face greater constraints in accessing public bond markets, and non-investment-grade firms. Although data availability remains a challenge—particularly due to the prevalence of private placements among small and medium-sized enterprises in Japan—such investigations could offer broader insights into corporate financing behavior and market access across different firm segments. Future crises and economic downturns may be partly anticipated through shifts in green bond spreads. Looking ahead, future research could extend our framework to green bonds by examining how “greenium” evolves in response to investor behavior and interactions with other financial instruments, thereby shedding light on market anticipation mechanisms for systemic stress.

Moreover, the two-stage structure of the empirical model raises potential concerns about endogeneity. While the analysis assumes that expected credit spreads influence issuance market choice, the choice of issuance location may have affected observed spreads, leading to reverse causality. This bidirectional relationship may introduce bias if credit spreads are not entirely exogenous. Additionally, textual disclosure in green bond prospectuses—such as mandated sections on use of proceeds and climate-related risk factors—could provide exogenous variation to help address endogeneity between credit spreads and unobserved issuer characteristics. Addressing this issue in future work—through instrumental variable methods or control function approaches—could help isolate causal effects and enhance the robustness of the empirical strategy. These insights not only deepen our understanding of bond pricing dynamics under segmented market conditions but also open avenues for more robust causal inference in sustainable finance.