Capital Structure Theories in US Corporate Divestitures: A Study on Spin-Off Firms

Abstract

Some giant US conglomerates are now undergoing corporate spin-offs or are considering such spin-offs in the near future. Corporate spin-offs offer a unique opportunity to assess corporate capital structure decisions. The leverage ratio of the spin-off firms represents their initial capital structure. We investigate the capital structure of corporate spin-offs and find evidence that they adhere to the trade-off theory. This study provides evidence that the subsidiary firms tend to aim for a target capital ratio during the sample period. The results indicate that the partial adjustment model with firm fixed effects is a good fit for the data sample. The parent companies in corporate spin-offs exhibit a similar pattern but with a slower adjustment speed. The tendency to target capital ratios is observable in both market value and book value leverage measures for the parent and subsidiary firms. Indicators of the pecking order assumption do not possess statistically significant coefficients. Changes in share price affect market debt ratios in the short term. With alternative definitions of leverage, the estimated adjustment speeds vary. In the case of longer horizons, the results align with a continuous rate of adjustment.

1. Introduction

Corporate capital structure is of immense importance to the risk nature of the firm. There is considerable interest in examining how firms make decisions about their capital structure in divestitures. Divestitures have several different forms, including selloffs, spin-offs, carve-outs, split-ups, and liquidations, each serving different strategic purposes for companies. When a corporate spin-off occurs, the subsidiary does not make its initial capital structure decision; its capital structure and leverage ratio are decided by the parent firm. Therefore, there is an opportunity to observe the initial corporate leverage ratio.

Studying the leverage or capital structure of spin-offs is important because it provides insight into how newly independent firms make financial decisions and optimize their operations. Spin-offs offer a unique setting where the parent company deliberately determines the new entity’s capital structure, allowing researchers to observe how factors like growth opportunities, asset structure, and industry characteristics influence leverage choices (Dittmar, 2004; Mehrotra et al., 2003). Understanding these decisions helps clarify which capital structure theories—such as the trade-off theory, which suggests firms balance the costs and benefits of debt, or pecking order theory which suggests how companies prioritize their sources of financing, they prefer internal funds followed by debt and using equity as a last resort—best explain real-world behavior, as spin-offs often deviate from the patterns seen in established firms (Dittmar, 2004; Simu, 2019; Mehrotra et al., 2003). Analyzing spin-offs also reveals how firms adjust toward target leverage and how these adjustments impact market reactions and future performance (Heinze & Finke, 2022). Additionally, research shows that spin-offs improve the efficiency of capital allocation and investment decisions, as managers focus more effectively on the core business after separation (Feldman, 2016; Gertner et al., 2002). Overall, examining the capital structure of spin-offs deepens understanding of financial policy, firm value, and the mechanisms that drive efficient resource allocation in corporate restructuring (Bock et al., 2018; Fu et al., 2022).

Our research aims to explore how companies establish their capital structure after a spin-off. This objective aligns with the work of Alderson and Betker (1995) and Gilson (1997), who investigated leverage decisions in firms emerging from financial distress. Gilson (1997) found that firms tended to maintain higher debt levels due to significant transaction costs. Dittmar (2004) analyzed the capital structures of subsidiaries following spin-offs and found evidence supporting the tradeoff theory, though her study does not address the pecking order theory. Lemmon and Zender (2010) tested many theories of capital structure and incorporated a measure of debt capacity. Using a large and long-term dataset, they summarized that the pecking order capital structure theory effectively explained the financing behavior of firms.

Building on these studies, we empirically compared the tradeoff theory and the pecking order theory in the context of corporate spin-offs relative to their parent companies. Our goal is to identify the theory that better explains the capital structure decisions of spin-offs and their underlying reasons. Additionally, we investigate whether these firms make changes to their leverage to converge toward a targeted ratio in the years following the divestiture.

Our evidence suggests that subsidiary firms tend to pursue a policy of targeting leverage ratios during the sample period. The parent companies in corporate spin-offs follow the same pattern, but the adjustment speed is slower. This targeting behavior is evident both in the market-value and the book-value leverage measures for both parent and subsidiary firms. The indicator for pecking order assumptions does not provide significant results. We find that changes in share price affect market debt ratios in the short term. Our additional analysis confirmed the robustness of the results using fixed effects, GMM, and Lasso methods, highlighting consistent lag effects, valid instruments, and key predictive variables.

The rest of this paper proceeds as follows. In Section 2, we discuss the related literature and develop the hypotheses for this study. In Section 3, we review the data sample and explain the construction of the variables. In Section 4, we present the methodology, and in Section 5, we present and discuss the results. We conclude in Section 6.

2. Literature Review and Hypothesis Development

2.1. Capital Structure Theories

There is a lot of prior research pertaining to the study of capital structure decisions by firms. Shyam-Sunder and Myers (1999) investigated whether historical decisions had any impact on firms’ capital structure decisions. They used various simulations in their study and showed that the trade-off theory provided the best explanation for capital structure choices. They showed that because of the cyclical nature of earnings and capital investments, firms seemed to revert back to certain debt ratios (leverage). This reversion might erroneously convey the idea that firms are trying to acquire a target debt ratio.

Frank and Goyal (2003) refuted the empirical evidence presented by Shyam-Sunder and Myers (1999) in support of the pecking order theory because it does not hold when tested using a broader sample of firms over a longer time frame. Similarly, Chirinko and Singha (2000) criticized the empirical approach used by Shyam-Sunder and Myers, arguing that it lacks the statistical power to effectively differentiate between competing financing theories. Lemmon and Zender (2010) developed previous studies by incorporating variation in firms’ debt capacity. Their findings reveal that, when external financing is required, firms with greater debt capacity tend to rely on debt to cover their funding needs, whereas firms with limited capacity are more dependent on external equity. Additionally, Bolton and Freixas (2000) developed a model closely aligned with the pecking order theory, offering a practical framework for understanding debt capacity. Their model emphasizes the role of asymmetric information between firms and the market in shaping financing decisions and proposes using firm risk and age to measure debt capacity.

2.2. Adjustment Speed Literature

Traditional partial adjustment models (e.g., A. Hovakimian et al., 2001; Flannery & Rangan, 2006) typically assume that leverage converges to its target at a constant pace. A number of earlier studies have explored the broader role of transaction costs in determining the speed of adjustment (e.g., Korajczyk & Levy, 2003; Strebulaev, 2007; Shivdasani & Stefanescu, 2010; Faulkender et al., 2012). Other research focuses on specific opportunity costs that influence capital structure decisions, including financing costs and market conditions (e.g., Altınkılıç & Hansen, 2000; Frank & Goyal, 2004; Flannery & Rangan, 2006; Liu, 2009; Warr et al., 2012), governance and investment opportunities (e.g., Chang et al., 2014; Elsas et al., 2014; Zhou & Xie, 2016), and strategic or liquidity considerations such as mergers and acquisitions or credit line access (e.g., Uysal, 2011; Lockhart & Flannery, 2010).

Despite this body of work, empirical evidence (e.g., Fama & French, 2002) consistently finds that the speed at which capital structure adjusts is generally low. As a result, understanding how and why firms modify their leverage towards a targeted ratio has attracted significant academic interest. According to the trade-off theory, firms maintain an optimal leverage and gradually move toward this target over time, primarily due to adjustment costs. Fischer et al. (1989) introduced a foundational dynamic model of capital structure in the presence of such costs. Several studies (e.g., Korajczyk & Levy, 2003; Strebulaev, 2007; Shivdasani & Stefanescu, 2010) have since built on this framework to examine how adjustment costs influence leverage behavior. More recent work shifts focus to the heterogeneity of adjustment speeds across firms, depending on the specific opportunity costs (e.g., Chang et al., 2014; Elsas et al., 2014; Zhou & Xie, 2016).

Closely related, another line of research emphasizes the role of equity misvaluation, suggesting that market conditions at the specific time of issuance can alter the cost and pace of adjusting capital structure. When a firm’s equity is overvalued, firms tend to favor equity issuance over debt, which can delay progress toward the target leverage ratio. For instance, past studies find market timing, proxied by market-to-book ratio, can explain leverage models and influence leverage decisions even outside active market-timing periods (Flannery & Rangan, 2006; Liu, 2009). Warr et al. (2012) further show that market timing does impact adjustment costs, thereby influencing the rate at which firms adjust their leverage toward optimal levels.

Some researchers suggest that firms choose target debt-to-equity ratios by balancing the costs and benefits associated with leverage. For instance, Graham and Harvey (2001) found that 81% of firms take a target debt ratio or range into account when making financing decisions. In contrast, different views exist, such as pecking order theory (e.g., Donaldson, 1961; Myers, 1984). According to pecking order theory, managers believe the market tends to undervalue their shares due to information asymmetry. Consequently, firms prioritize financing investments with internal funds first, resort to issuing safer debt if internal resources are not sufficient, and only issue new equity as the last option. This theory suggests that corporate leverage is primarily shaped by its historical profitability and investment requirements. Rather than targeting specific leverage ratios, firms typically do not exhibit a strong inclination to adjust leverage in response to changes caused by financing demands or earnings fluctuations.

Other capital structure theories also question the notion that firms deliberately align with a preferred leverage ratio. For example, Baker and Wurgler (2002) contend that corporate leverage is the cumulative result of its past ability to issue overvalued equity. Since share prices tend to fluctuate around their intrinsic values, managers are more likely to raise equity capital during periods of high market valuation. This market timing hypothesis suggests that managers aim to capitalize on information asymmetries to benefit existing shareholders. As a result, if market timing plays a dominant role in leverage decisions, there would be no tendency for firms to go back to a specific target leverage ratio. In addition, Welch (2004) proposes that due to managerial inertia, fluctuations in stock prices significantly influence market-value-based debt ratios. Over time, these price effects become more influential in explaining leverage than traditional explanatory factors.

In summary, the three primary theories of capital structure, the pecking order, the market timing, and the inertia theory, suggest that leverage has little to no impact on a firm’s value. As a result, managers do not actively attempt to offset shifts in leverage. The tradeoff theory, however, suggests that because of market imperfections, leverage results can have a direct impact on firm value. Accordingly, managers take deliberate actions to realign their leverage when it deviates from the target. The pace of a firm’s return to its target leverage is determined by the cost of adjustment. If adjustment costs were zero, the trade-off theory implies that firms would always maintain their target leverage. On the other hand, if these costs were prohibitively high, firms would not adjust their leverage at all. Therefore, estimating the effect of capital structure adjustment costs is a crucial first step in evaluating competing theories of capital structure.

A firm’s capital structure evolves in response to its historical operating performance and financial decisions. When a firm experiences a streak of unexpectedly high profits, its leverage tends to fall below the target debt ratio. Conversely, a firm facing a series of unexpectedly low profits is likely to exhibit leverage above its target. Firms typically act to correct such deviations by adjusting their capital structure. However, adverse market conditions or high transaction costs can delay these adjustments, leading to temporary divergences from the target debt ratio.

These deviations can create a negative relationship between leverage and the firm’s profitability pattern, documented by many prior studies. Dittmar (2004), for instance, found that firms make deliberate capital structure choices influenced by factors such as the size of the firm, its growth prospects, and collateral value. Unlike much of the existing research, Dittmar’s study concludes that profitability does not significantly affect debt decisions. It suggests that the factors driving the negative relationship between leverage and profitability do not apply to spun-off subsidiaries. The study notes that if profitability correlates with unobserved growth opportunities, the negative relationship can be explained by the trade-off theory. If growth is not the underlying factor, however, the theory cannot explain the inverse link between profitability and leverage. Dittmar’s (2004) study found results that support the predictions of the trade-off theory of capital structure for subsidiaries after spin-off. This paper is an extension of Dittmar’s (2004) article as we examine both the trade-off and pecking order theories of capital structure for spin-off firms.

2.3. Spin-Offs

Corporate spin-offs are distinct from other divestiture strategies like carve-outs of equity or the sale of assets. In a spin-off, the spun-off unit becomes an independent company with its own legal and operational structure, and shares of the new entity are distributed to existing shareholders of the parent firm on a pro-rata basis. These shareholders retain financial claims on both the original company and the spun-off firm. The primary structural change is the establishment of a new management team and board of directors for the spun-off entity. If the two companies operate more efficiently post-spin-off than the combined firm did previously, this separation is expected to improve performance and enhance shareholder wealth.

Academic research (e.g., Aron, 1991; Parrino, 1997; Nanda & Narayanan, 1999; Chemmanur & Yan, 2004) has proposed several theoretical reasons why firms pursue spin-offs, generally concluding that long-term performance tends to improve. Empirical studies (e.g., Cusatis et al., 1993; Daley et al., 1997; Desai & Jain, 1999; Burch & Nanda, 2003; Chemmanur et al., 2014; Klein & Rosenfeld, 2010) support this view; they show that spin-offs do lead to better results and increased shareholder value over time.

The capital structure of spin-off subsidiaries is primarily determined by growth opportunities, with firms exhibiting lower leverage ratios compared to their parent companies, but similar to those that have not been spun off, of comparable size and industry. Research indicates that profitability does not significantly influence leverage decisions in spin-offs, which aligns with the capital structure theory of trade-off, suggesting that firms examine the costs and benefits of debt when making these decisions. Unlike non-spin-off firms, the determinants of leverage in spin-offs are distinct, as the parent company actively chooses the subsidiary’s capital structure at the time of separation. This process allows for a more deliberate alignment with the subsidiary’s specific growth prospects and risk profile, rather than simply mirroring the parent’s financial policies. Empirical evidence also suggests that using year-end debt ratios may obscure the true factors influencing leverage, as spin-offs often start with capital structures intentionally set to meet target ratios from the outset. Overall, the findings emphasize the importance of growth opportunities and the deliberate, theory-driven approach to capital structure in spin-off entities, rather than reliance on historical profitability or passive inheritance of the parent’s financial characteristics (Dittmar, 2004; Simu, 2019).

Empirical research on the capital structure of spin-offs reveals that these newly independent firms typically have lower leverage ratios than their parent companies, but their leverage is similar to that of comparable firms that have not been spun off. Growth opportunities are consistently found to be the main determinant of leverage in spin-offs, while profitability has little to no impact, supporting the capital structure theory of trade-off over the pecking order theory in this context (Dittmar, 2004; Simu, 2019). Research also shows that the leverage of spin-offs is deliberately set by the parent at the time of separation, allowing for a tailored approach that reflects the subsidiary’s specific risk and growth profile rather than simply inheriting the parent’s financial policies (Dittmar, 2004; Simu, 2019; Mehrotra et al., 2003). Some studies highlight that spin-off subsidiaries adjust their leverage over time to align with target capital structures, especially when they start out overleveraged or underleveraged, and that these adjustments are influenced by firm-specific factors such as their opportunities for growth and profitability. Additionally, the reliability of traditional tests for target leverage behavior in spin-offs has been questioned, as mean reversion and active financing decisions can obscure true target-following behavior (Heinze & Finke, 2022). Overall, the empirical evidence underscores that spin-off capital structure decisions are distinct from those of both parent and non-spin-off firms, with a strong emphasis on growth prospects and a deliberate, theory-driven approach to leverage.

Research comparing capital structure in spin-offs and mergers highlights key differences in how firms approach leverage and financial policy. In spin-offs, the newly independent subsidiary typically adopts a leverage ratio that is lower than its parent but similar to comparable standalone firms, with growth opportunities being the main determinant of leverage, supporting the trade-off theory of capital structure rather than the pecking order theory (Dittmar, 2004; Simu, 2019; Heinze & Finke, 2022). Profitability does not significantly influence leverage in spin-offs, and managerial incentives or governance factors appear to have little effect (Dittmar, 2004; Mehrotra et al., 2003) In contrast, mergers can create financial synergies by pooling cash flows, reducing risk, and potentially increasing optimal leverage, but these benefits depend on the similarity of the merging firms’ risk profiles and default costs; if these differ greatly, the inability to tailor capital structures to each activity may outweigh diversification gain (Leland & Skarabot, 2003). Thus, while spin-offs allow for more customized capital structures aligned with the specific risk and growth profile of the new entity, mergers often seek to exploit risk reduction and tax benefits, though not always successfully. Overall, spin-offs tend to result in capital structures that reflect the unique characteristics of the spun-off business, whereas mergers may impose a restraining one-size-fits-all approach that is not always optimal for all combined activities.

We suppose firms choose an optimal leverage ratio, but their choice of leverage is determined and varies by their firm characteristics. The spin-off sample offers a clearer view of the relationship between firm characteristics and leverage, as subsidiary leverage ratios are set deliberately rather than inherited from prior operating histories. Low- and high-levered firms adjust toward total and long-term leverage targets faster than mid-levered firms, though short-term SOA is slower—especially post-liberalization (T. Nguyen et al., 2021; Qureshi et al., 2017). The negative leverage–profitability link reflects partial adjustment, with changes reversing within five years (Kayhan & Titman, 2007).

2.4. Hypothesis Development

• Null hypothesis

Many firms appear not to possess the optimal capital structures because they do not fully utilize the debt tax shield (Miller, 1977; Graham, 2000). Kopecky et al. (2018) discuss Pareto theory and stated that today’s market environment might not warrant the issuance of debt with the sole purpose of capturing the value generated by the interest tax credit because the benefit of the interest debt shield is already reflected in the equity component of the firm’s value. Therefore, it could be possible that spin-off firms do not try to adjust their leverage after inception. So, our null hypothesis becomes “spin-off firms do not change their capital structure or try to adjust to a target debt ratio.”

2.

Pecking order theory and capital structure of spin-off

To develop a hypothesis about the pecking order theory and capital structure of spin-offs, it is important to consider how the pecking order theory predicts firms’ financing choices. The pecking order theory suggests that firms prefer to finance new investments first with their internal funds, then with debt, and use equity only as a last resort, due to information asymmetry and the costs of external financing (H. Nguyen et al., 2020; Jarallah et al., 2018; Budiarso & Pontoh, 2021; Serrasqueiro & Caetano, 2014; Chen & Chen, 2011). Empirical studies show that more profitable firms tend to use less debt, which supports the pecking order theory’s prediction that internal funds are preferred (H. Nguyen et al., 2020; Jarallah et al., 2018; Budiarso & Pontoh, 2021; Serrasqueiro & Caetano, 2014; Chen & Chen, 2011). In the context of spin-offs, this would imply that newly independent firms with strong internal cash flows will choose lower leverage, while those with fewer internal resources may rely more on debt. Alternatively, it could be possible that spin-off firms do not try to alter their debt ratio after inception. So, the null hypothesis becomes:

• Spin-off firms do not change their capital structure or try to adjust to an optimal debt ratio.

Therefore, a plausible alternative hypothesis is: “Spin-off firms with higher profitability and greater internal resources will have lower leverage, consistent with the pecking order theory, while those with lower profitability will exhibit higher leverage due to greater reliance on external financing.” This hypothesis aligns with findings that profitability negatively affects leverage and that growth opportunities may increase the need for external financing, especially in newly independent spin-offs.

3.

Trade-off theory and capital structure of spin-off

To develop a hypothesis about the trade-off theory and spin-off capital structure, consider that the trade-off theory posits firms decide on their leverage by balancing the tax benefits of debt against the costs of financial distress and bankruptcy (Ai et al., 2020). In the case of spin-offs, empirical research shows that the new, stand-alone entities tend to have leverage ratios lower than their parent companies but similar to comparable independent firms, and that growth opportunities are the primary determinant of leverage, while profitability has little impact (Dittmar, 2004; Simu, 2019). This supports the trade-off theory’s view that firms weigh the costs and benefits of debt to reach an optimal target, rather than relying solely on internal resources or profitability.

Therefore, a suitable hypothesis is “Spin-off firms set their capital structure by balancing the tax advantages of debt with the potential costs of financial distress, resulting in leverage ratios that reflect their specific growth opportunities and risk profiles, consistent with the trade-off theory.” This hypothesis is backed by findings that spin-offs deliberately tailor their leverage to optimize these trade-offs, rather than mirroring their parent’s capital structure or relying solely on internal funds.

In our study we employ an empirical model that captures the potential dynamics of corporate capital structure. We use this model to examine whether firms maintain a target leverage ratio and, if so, how quickly they adjust toward that target. The findings show that firms do, in fact, aim for a specific capital structure. Moreover, it suggests that this targeting behavior significantly influences the capital structures firms adopt in practice.

3. Sample Construction and Variable Measurement

When a company initiates a spin-off, its current shareholders are allocated shares in the newly formed subsidiary proportionate to their existing ownership. The subsidiary which used to be part of the parent company becomes an independent company after the spin-off. Under IRS tax regulations, a newly spun-off company is expected to operate independently, with the parent firm retaining minimal influence—typically less than 20% of the voting power in the new entity. Therefore, spin-offs result in complete divestiture of the subsidiary resulting in a totally new, independent company.

The empirical sample comprises completed spin-off transactions announced between 1992 and 2016, as recorded in the SDC Platinum database. Over 900 spin-offs are listed in SDC in this sample period. Financial data for both post–spin-off parent companies and their subsidiaries were obtained from the Compustat database, which provided coverage for 238 completed spin-off pairs.

We estimate the target debt ratio based on a set of firm-specific variables (X_i,t) commonly cited in prior literature, including studies by Rajan and Zingales (1995), A. Hovakimian et al. (2001), Fama and French (2002), and G. Hovakimian and Titman (2003). We anticipate that these firm characteristics will influence the target leverage ratio in the following ways:

DEP_AT: Depreciation scaled by total assets. Depreciation relative to total assets serves as an indicator of a firm’s reliance on non-debt tax shields. Firms with higher depreciation typically exhibit a diminished need for interest-related tax benefits, often resulting in lower levels of debt financing.

EBIT_AT: EBIT scaled by total assets. A firm exhibiting higher operating profits relative to its total assets may adopt either a conservative or aggressive leverage strategy. Lower leverage could result from ample retained earnings, reducing reliance on external financing, or from a deliberate effort to preserve financial stability by minimizing debt exposure. Alternatively, elevated leverage may signal the firm’s confidence in its capacity to meet debt obligations through strong and consistent cash flows.

LnAT: Natural logarithm of total assets. Larger firms often exhibit higher leverage, potentially due to greater transparency, more stable asset returns, or lower cost of debt.

M/B: Market-to-book ratio. A high market-to-book ratio reflects investor expectations of substantial future growth. To safeguard this perceived valuation, firms often choose to maintain conservative leverage levels.

PPE_AT: Property, Plant, and Equipment scaled by total assets. Firms with a higher share of tangible assets generally possess a stronger capacity to sustain elevated levels of debt financing.

R&D_AT: Research and development expenses scaled by total assets. Firms with substantial investments in research and development often prefer equity financing, as it offers greater flexibility and avoids the constraints associated with debt obligations.

FF12 Factor/INDUSTRY CLASSIFICATION: To capture industry-specific effects not addressed by other control variables, we incorporate the Fama-French 12-industry classification framework.

4. Methodology

The methodological approach adopted in this study draws upon the framework established by Flannery and Rangan (2006).

The firm’s market debt ratio serves as the principal metric for leverage, defined as follows:

MDR_i,t = D_i,t/(D_i,t + S_i,tP_i,t)

(1)

in this case D_i,t represents the book value of firm i’s interest-bearing debt at time t, S_i,t represents the number of common shares outstanding at time t, and P_i,t represents the share price at time t.

The target leverage ratio can differ across firms and over time, modeled as follows:

MDR_i,t+1 = βX_i,t,

(2)

Here, MDR_i,t+1 denotes firm i’s target debt ratio at t + 1, while X_i,t, represents a vector of firm-specific characteristics that shape the trade-offs involved in determining optimal leverage levels. The vector β contains the corresponding coefficients. According to trade-off theory, the assumption that β ≠ 0 implies that firm-specific characteristics play a meaningful role in determining the target leverage ratio.

4.1. Adjustment to Target Leverage

The standard partial adjustment model that we use is given by

MDR_i,t+1 − MDR_i,t = ʎ (MDR_i,t+1 − MDR_i,t) + ẟ_i,t+1

(3)

Each year, the firm adjusts a fraction ʎ of the discrepancy between its actual and target leverage levels. Here ʎ represents the speed of adjustment toward the optimal capital structure, while ẟ denotes the error term capturing unexplained variation.

Substituting the expression for MDR* in Equation (3)

The estimable model becomes

MDR_i,t+1 = (ʎβ)X_i,t+1 + (1 − ʎ) MDR_i,t + ẟ_i,t+1

(4)

Managers seek to minimize the gap between the firm’s actual leverage (MDR_i,t) and its intended target leverage (βX_i,t+1), aligning capital structure decisions with strategic financial goals. This model indicates that

• Over time, the actual corporate leverage will converge to the target level;
• The long-term effect of X_i,t on the leverage ratio is calculated by dividing its estimated coefficient by ʎ;
• The speed of adjustment is uniform across all firms.

4.2. Pecking Order Tests

According to the Pecking Order Theory, fluctuations in a firm’s book-debt ratio can reflect its financing deficit. This relationship can be examined by evaluating a book-value analogue of the primary specification outlined in Equation (4).

ΔBDR_i,t+1 = (ʎβ)X_i,t + ʎBDR_i,t + ɣ₂FINDEF_i,t+1 + ɛ_i,t+1

where BDR denotes the book-debt ratio, calculated as total debt—comprising both long-term and short-term obligations—divided by total assets. FINDEF represents the firm’s financing deficit, measured as the sum of dividends and the change in working capital, normalized by total assets.

4.3. Stock Price Impact

Welch (2004) finds no evidence that managers actively target a specific leverage ratio. Instead, the study suggests that managers tend to passively accept nearly all changes in the market debt ratio (MDR) resulting from fluctuations in share prices. This conclusion is drawn from the following regression model:

MDR_i,t+1 = α₀ + α₁MDR_i,t + α₂IDR_i,t+1 + μ_i,t+1

where IDR_i,t+1, which represents the implied debt ratio equal to D_i,t/(D_i,t + S_i,tP_i,t(1 + R_i,t+1)).

The variable R_i,t+1 captures the realized appreciation in firm i’s share price over the interval from t to t + 1. The parameters α₀, α₁, α₂ are to be estimated within the model. The impact of share price movements is reflected through changes in the firm’s market debt ratio (MDR).

SPE_t+1 = D_t/(D_t + S_tP_t(1 + R_t,t+1)) − MDR_t

IDR_i,t+1 = SPE_i,t+1 + MDR_i,

5. Results and Discussion

5.1. Descriptives

The frequency of corporate spin-offs during the sample period shows notable surges in certain years. As illustrated in Figure 1, the peak in spin-off announcements occurred in 2000, while the greatest number of spin-offs reaching completion was recorded in 1998.

As shown in Figure 2, the debt ratios indicate that parents have much higher debt ratios compared to subsidiaries.

Figure 3 reveals that, according to the Fama-French 12 industry classification, the majority of corporate spin-offs are concentrated within the Computers, Software, and Electronic Equipment sector, as well as the Money and Finance industry. Excluding the industry with the highest number of spin-offs yielded no significantly different results, so the main sample analysis is included in this study.

The key debt ratios for companies in pre-spin-off and spin-off years are shown in

Table 1. Ratios for parent and subsidiary.

	Parent Pre Spin-off	Parent Spin-Off Year	Subsidiary Spin-Off Year
Long-term Debt/Book Value of Equity	0.8303	0.8244	0.559
Total Debt/Equity	1.688	1.7067	2.256
Total Debt/Total Assets	0.5882	0.58189	0.5531
Total Debt/capital	0.4867	0.4811	0.4263
Long-term Debt/invested capital	0.3383	0.3292	0.2486
Book/Market	0.80002	0.6694	0.8805
Cash Flow/Total Debt	0.0720	0.0968	0.1495
Long-term Debt/Total Liabilities	0.3123	0.3078	0.2424
Debt Senior-Convertible	3.0047	4.124837	0
Debt Subordinated Convertible	14.07564	13.8468	3.9055
Debt Convertible	16.62729	17.392	3.59992
Debt Debentures	104.5014	96.5423	11.8303
Long-term Debt maturing in 1 year	58.0385	51.2747	40.7467
Long-term Debt maturing in 2 years	57.7389	70.2826	7.1283
Long-term Debt maturing in 3 years	55.1286	62.9996	3.6481
Long-term Debt maturing in 4 years	58.1591	72.2290	10.8827
Long-term Debt maturing in 5 years	68.6574	76.9210	4.2361
Debt in current liabilities	608.2634	693.1448	668.5947
Current debt changes	−62.16911	−2.902008	10.6138

.

The subsidiaries have less long-term debt and more cash flows than the parents, overall.

Table 2. Correlation for firm characteristics variables for subsidiaries.

	EBIT_AT	M/B	DEP_AT	LnAT	PPE_AT	R&D_AT	FF12 Factor
EBIT_AT	1.000
M/B	−0.9101	1.000
DEP_AT	−0.1594	0.1104	1.0000
LnAT	0.2569	−0.1390	−0.2005	1.0000
PPE_AT	0.0399	−0.0297	0.4499	−0.1065	1.0000
R&D_AT	−0.0038	0.0012	0.0323	0.0110	0.0673	1.0000
FF12 Factor	−0.0125	−0.0080	−0.1177	−0.0222	0.0345	0.0110	1.0000

presents the correlation matrix of firm-specific characteristics for subsidiary entities.

Table 3. Summary statistics. The data sample covers all industrial firms in Compustat which are spun-off with data from 1992 to 2016. Total: 238 firms; 1486 firm years and all Industrial Compustat firms which spin-off their subsidiaries with data during 1992 to 2016. Total: 238 firms; 2629 firm years.

Panel A: Subsidiary
Variable	Observation	Mean	Standard Deviation	Minimum	Maximum
MDR	1486	0.2263	0.2376	0	0.9999
SPE	1486	0.1402	3.8786	−0.6191	10.8889
BDR	1486	0.2932	1.0127	0	3.6667
EBIT_AT	1486	0.0355	0.7378	−2.6833	0.1326
M/B	1486	5.2121	1.4842	0.0712	67.117
DEP_AT	1486	0.0425	0.042	0	1.0741
lnAT	1486	6.4076	2.2309	−5.116	13.4459
PPE_AT	1486	0.2722	0.2487	0	1
R&D_AT	1486	−0.0006	0.0061	−0.1519	0
L3MDR	1486	0.6441	0.6398	0	2.9571
FINDEF	1486	0.4125	0.1486	−0.12	0.3048
MDR1	1486	0.1638	0.1725	0	0.9335
MDR2	1486	0.3819	0.2523	0.0008	0.9999
MDR3	1486	0.1391	0.155	0	0.9335
Panel B: Parent
Variable	Observation	Mean	Standard Deviation	Minimum	Maximum
MDR	2629	0.2525	0.2229	0	1
SPE	2629	0.2378	0.2265	0	0.9993
BDR	2629	1.1655	0.2713	0	0.1115
EBIT_AT	2629	−0.1053	5.9716	−354.5	4.5384
M/B	2629	3.3193	39.3629	0.0185	1445.69
DEP_AT	2629	0.0852	3.3002	0	0.2303
lnAT	2629	7.5833	2.4686	−6.2146	14.7606
PPE_AT	2629	0.2355	0.2316	0	1
R&D_AT	2629	0.0644	0.3935	0	0.1955
L3MDR	2629	0.6861	0.582	0	2.9193
FINDEF	2629	0.2236	0.197	0	3.4979
MDR1	2629	0.1738	0.1566	0	0.9247
MDR2	2629	0.3815	0.2167	0.0008	1
MDR3	2629	0.4561	0.6597	0	0.4076

presents Summary statistics.

5.2. Different Estimation Methods (Equation (4))

The results from different estimation methods for Equation (4) are presented in

Table 4. This table presents alternate estimation methods for specification (4). Regression results for the model MDR_i,t+1 = (ʎβ)X_i,t+1 + (1 − ʎ) MDR_i,t +ẟ_i,t+1, where MDR is the market debt ratio. The “X” variables are the determinants for a firm’s long-term target debt ratio; p-values are shown in parentheses. MDR is the book value of total debt over the market value of assets. EBIT_AT is the earnings before interest and taxes divided by total assets. M/B is the market-to-book ratio. DEP_AT is Depreciation scaled by total assets. LnAT is the natural logarithm of total assets. PPE_AT is Property, Plant & Equipment per total assets. R&D_AT is Research and development expenses per total assets. The FF industry factor, or Fama-French 12 industry classification, is used to account for industry characteristics not captured by other explanatory variables. * indicates significance at the 10% level, ** indicates significance at the 5% level, and *** indicates significance at the 1% level.

Panel A: Subsidiary	(1)	(2)	(3)	(4)	(5)
	Baseline	Fixed Effects Panel	Fixed Effects Panel (Year Dummy)	Instrumental Variable Panel	Random Effects Panel
lagMDRit	0.7857 ***	0.5080 ***	0.5353 ***	1.0972 ***	0.7246 ***
	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)
EBIT_AT	−0.1459 ***	−0.1791 ***	−0.1596 ***	−0.0497	−0.1639 ***
	(0.000)	(0.000)	(0.000)	(0.114)	(0.000)
M/B	−0.0003 **	−0.0002	0.0000	−0.0001 *	−0.0003 **
	(0.043)	(0.132)	(0.969)	(0.443)	(0.038)
DEP_AT	0.1008	0.2951	0.0987	0.3206 **	0.1110
	(0.408)	(0.126)	(0.597)	(0.019)	(0.423)
LnAT	0.0053 **	0.0281 ***	0.0282 ***	0.0006	0.0088 ***
	(0.011)	(0.000)	(0.000)	(0.809)	(0.001)
PPE_AT	0.0356 **	0.0272	0.0271	−0.0267	0.0462 **
	(0.017)	(0.569)	(0.593)	(0.120)	(0.012)
R&D_AT	−0.1330 **	−0.0550	−0.1165	0.0026	−0.1595 **
	(0.010)	(0.644)	(0.304)	(0.965)	(0.011)
F-F industry factors	−0.0008
	(0.329)
Fixed Effects?	No	Yes	Yes	No	No
No. of Obs	1002	1002	1002	1002	1002
R2 Adj.	0.7149	0.6392	0.6855	0.6407	0.7146
Panel B: Parent	(1)	(2)	(3)	(4)	(5)
	Baseline	Fixed Effects Panel	Fixed Effects Panel (Year Dummy)	Instrumental Variable Panel	Random Effects Panel
lagMDRit	0.7944 ***	0.5888 ***	0.6122 ***	1.0961 ***	0.8029 ***
	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)
EBIT_AT	−0.1688 ***	−0.1453 ***	−0.1270 ***	−0.0310	−0.1623 ***
	(0.000)	(0.000)	(0.000)	(0.265)	(0.000)
M/B	0.0000	0.0000	0.0000	0.0000	0.0000
	(0.707)	(0.497)	(0.838)	(0.690)	(0.694)
DEP_AT	0.2636 **	0.3204 **	0.2185	0.1670	0.2933 ***
	(0.014)	(0.031)	(0.131)	(0.166)	(0.006)
LnAT	0.0057 ***	0.0111 ***	0.0104 **	0.0026 **	0.0055 ***
	(0.000)	(0.000)	(0.001)	(0.043)	(0.000)
PPE_AT	0.0001	−0.0015	0.0065	−0.0277 *	−0.0013
	(0.993)	(0.965)	(0.861)	(0.072)	(0.925)
R&D_AT	−0.1232 **	−0.2057 *	−0.1905 **	0.1085 **	−0.1363 ***
	(0.003)	(0.008)	(0.013)	(0.028)	(0.001)
F-F industry factors	−0.0018 *
	(0.010)
Fixed Effects?	No	Yes	Yes	No	No
No. of Obs	1646	1646	1646	1646	1646
R2 Adj.	0.7441	0.7294	0.7522	0.6727	0.7443

.

Column (2) in Panel A presents panel regression with fixed effects for subsidiaries, showing coefficient estimates for the factors influencing target leverage, except for LnAT. The inclusion of fixed effects enhances the statistical significance of most variables, thereby validating their use in the target MDR specification. The estimated coefficient on the lagged MDR implies a much faster adjustment speed (49.2%) in the panel model, meaning a typical firm closes its leverage gap within about one year. The increased speed may be attributed to the inclusion of firm fixed effects in the target specification or to the panel regression’s assumption of time-invariant slope coefficients. Furthermore, firm-specific unobserved effects likely have a strong impact on the estimated adjustment speeds, as they considerably improve the precision of the target debt ratio estimates.

Column (3) shows the estimations of a revised panel model with year dummies. The adjusted R² increases modestly compared with column (2), and the estimated coefficients follow a similar pattern.

Column (4) addresses the endogeneity concern stemming from the correlation between the lagged dependent variable and the error term in panel data, which may bias the estimated speed of adjustment. In this specification, the coefficient on MDR_t rises substantially (from 0.5080 to 1.0972), while the other coefficients remain broadly similar to those previously reported. An implied adjustment speed of 21.2% suggests that the average firm achieves over half of its necessary leverage correction within a five-year period. This relatively swift convergence toward a firm-specific capital structure implies that pecking-order and market-timing theories are unlikely to be the primary determinants of leverage decisions for most firms.

In Table 4, Panel B, we present the corresponding results for the parent firms. The coefficient on lagged MDR suggests an adjustment speed of 41.12% in the panel regression, indicating that the typical firm reduces its leverage gap rapidly—largely within two years. Column (4) mitigates concerns about endogeneity arising from the correlation between the lagged dependent variable and the error term, which can distort the estimated speed of adjustment. After accounting for this, the coefficient on MDR_t rises from 0.5888 to 1.0961.

5.3. Regression Results for the Capital Structure Models

The capital structure models estimated for both parent firms and their subsidiaries using the same panel dataset are reported in

Table 5. This table presents regression results for Capital Structure Models: (1) MDR_i,t+1 = βX_i,t + ẟ_i,t+1. (2) MDR_i,t = (ʎβ)X_i,t + (1 − ʎ) MDR_i,t + ẟ_i,t+1. (3) MDR_i,t+1 = (ʎβ)X_i,t + (1 − ʎ) MDR_i,t+1 + ẟ_i,t+1. (4) MDR_i,t+1 = ʎ(TDROLS) + (1 − ʎ) MDR_i,t + ẟ_i,t+1. (5) MDR_i,t+1 = βL3MDR_i,t + (1 − ʎ) MDR_i,t + ẟ_i,t+1. MDR is the book value of total (short-term plus long-term) debt over market value of assets. EBIT_AT is the earnings before interest and taxes divided by total assets. M/B is the market-to-book ratio. DEP_AT is Depreciation per total assets. LnAT is the natural logarithm of total assets. PPE_AT is Property, Plant & Equipment per total assets. R&D_AT is Research and development expenses per total assets. TDROLS is the target debt ratio, which is the predicted MDR from OLS regression. L3MDR is the three-year trailing average MDR. * indicates significance at the 10% level, ** indicates significance at the 5% level, and *** indicates significance at the 1% level.

Panel A: Subsidiary
	(1)	(2)	(3)	(4)	(5)
lagMDRit		0.7881 ***	0.5080 ***	0.7950 ***	−0.5316 ***
		(0.000)	(0.000)	(0.000)	(0.000)
EBIT_AT	−0.2841 ***	−0.1478 ***	−0.1791 ***
	(0.000)	(0.000)	(0.000)
M/B	−0.0002 **	−0.0003 **	−0.0002
	(0.010)	(0.040)	(0.132)
DEP_AT	−0.2895	0.1077	0.2951
	(0.138)	(0.375)	(0.126)
LnAT	0.0241 ***	0.0053 **	0.0281 ***
	(0.000)	(0.010)	(0.000)
PPE_AT	0.2123 ***	0.0334 **	0.0272
	(0.000)	(0.024)	(0.569)
R&D_AT	0.0038	−0.1348 **	−0.0550
	(0.891)	(0.009)	(0.644)
TDROLS				0.3095 ***
				(0.000)
L3MDR					1.5642 ***
					(0.000)
Fixed Effects?	No	No	Yes	No	No
No. of Obs	1114	1002	1002	1002	896
R2 Adj.	0.1705	0.7149	0.6392	0.7113	0.8619
Panel B: Parent
	(1)	(2)	(3)	(4)	(5)
lagMDRit		0.8029 ***	0.5888 *	0.8003 ***	−0.5152 ***
		(0.000)	(0.000)	(0.000)	(0.000)
EBIT_AT	−0.5075 ***	−0.1623 ***	−0.1453 *
	(0.000)	(0.000)	(0.000)
M/B	0.0000	0.0000	0.0000
	(0.981)	(0.694)	(0.497)
DEP_AT	0.5328 ***	0.2933 ***	0.3204 *
	(0.003)	(0.006)	(0.031)
LnAT	0.0130 ***	0.0055 ***	0.0111 *
	(0.000)	(0.000)	(0.000)
PPE_AT	0.0779 ***	−0.0013	−0.0015
	(0.001)	(0.925)	(0.965)
R&D_AT	−0.8243 ***	−0.1363 ***	−0.2057
	(0.000)	(0.001)	(0.008)
TDROLS				0.2471 ***
				(0.000)
L3MDR					1.5267 ***
					(0.000)
Fixed Effects?	No	No	Yes	No	No
No of Obs	1748	1646	1646	1646	2529
R2 Adj.	0.2067	0.7432	0.7294	0.7413	0.8905

.

In column (1) of Panel A, we present a basic cross-sectional specification commonly used in earlier research to identify corporate optimal leverage. The coefficient estimates closely align with the determinants from previous studies. Higher earnings, elevated market-to-book ratios, and greater depreciation expenses are each linked to lower target leverage levels. while larger firm size (total assets) and higher levels of fixed assets are linked to higher target leverage. This specification imposes a zero constraint on the coefficient of the lagged market debt ratio (MDR), suggesting that the current debt ratio fully aligns with its target. However, the results in column (2) challenge this assumption. Introducing the lagged dependent variable yields a statistically significant coefficient of 0.7881, highlighting the presence of a key explanatory factor omitted by the simple cross-sectional model. Consequently, the exclusion of firm fixed effects is not warranted.

Column (3) employs a partial adjustment model toward the target capital ratio, incorporating firm-level fixed effects. The resulting coefficient estimates diverge notably from those in the more basic specification presented in column (1). For instance, the coefficient on EBIT_AT in column (1) differs substantially from its long-run counterpart in column (3), while the long-run estimates for LnAT and PPE_AT are markedly lower. These discrepancies underscore the impact of omitted variables in column (1), which significantly distort the identification of key determinants of target leverage.

Column (4) applies a partial adjustment model using a target debt ratio (TDROLS) derived from column (1). The estimated adjustment speed of 30.95% aligns with findings from other studies using target proxies without fixed effects, although the TDROLS coefficient remains below theoretical expectations. column (5) assesses whether L3MDR—a three-year trailing average MDR—serves as an effective proxy for the firm’s target capital ratio.

In Panel B, column (1), the inclusion of the lagged dependent variable produces a significant coefficient of 0.8029, indicating strong persistence in capital structure. Column (3) reintroduces the partial adjustment framework with firm fixed effects. Among parent firms, the coefficient on EBIT_AT in column (1) diverges notably from its long-run counterpart in column (3), while the long-run estimates for LnAT and PPE_AT remain relatively stable across specifications. Column (4) estimates a partial adjustment model using TDROLS, revealing a slower adjustment speed of 24.71%, significantly less than that of subsidiaries. However, the TDROLS coefficient closely matches that of subsidiaries. Finally, column (5) finds that L3MDR is a poor proxy for target capital structure.

5.4. Pecking Order Explanations of Book Debt Ratio

Prior studies have developed models in which the financing deficit—central to the pecking order theory—and the weighted average of historical market-to-book ratios—reflecting market timing—compete with trade-off theory variables in explaining capital structure decisions. Frank and Goyal (2003) contend that, under the pecking order framework, the financing deficit should fully explain changes in capital structure, thereby minimizing the relevance of alternative explanatory factors. However, if the financing deficit represents just one among several considerations in firms’ financing choices, this interpretation aligns more closely with an expanded view of the trade-off theory—a perspective also endorsed by Baker and Wurgler (2002).

To evaluate this claim, we employ a book-value specification of the main model, defining the book debt ratio (BDR) as total debt (long-term plus short-term) divided by total assets. The financing deficit (FINDEF) follows the definition provided by Frank and Goyal (2003): (dividends + investments + change in working capital + internal cash flow) divided by total assets. For subsidiary firms, we adapt FINDEF to include only dividends and changes in working capital. We then estimate a regression of BDR on this modified FINDEF alongside other control variables. The central question is whether FINDEF significantly affects the coefficients of the control variables X_i,t or the lagged dependent variable. Evidence consistent with the pecking order theory is presented in

Table 6. This table presents Pecking order explanations of the BDR: BDR_i,t+1 = (ʎβ)X_i,t + (1 − ʎ)BDR_i,t + ɣFINDEF_i,t+1 + ɛ_i,t+1. ΔBDR_i,t+1 = (ʎβ)X_i,t − ʎBDR_i,t + ɣFINDEF_i,t+1 + ɛ_i,t+1. MDR is the book value of total debt over market value of assets. BDR is the value of total debt over total assets. EBIT_AT is the earnings before interest and taxes divided by total assets. M/B is the market-to-book ratio. DEP_AT is Depreciation per total assets. LnAT is the natural logarithm of total assets. PPE_AT is Property, Plant & Equipment per total assets. R&D_AT is Research and development expenses per total assets. FINDEF is the financing deficit as defined: (dividend payments + change in working capital)/(total assets). * indicates significance at the 10% level, ** indicates significance at the 5% level, and *** indicates significance at the 1% level.

Panel A: Subsidiary
	(1) BDR	(2) ΔBDR	(3) MDR	(4) ΔMDR	(5) MDR
lagBDRit	0.6966 ***	−0.3717 ***
	(0.000)	(0.000)
lagMDRit			0.5112 ***	−0.4888 ***
			(0.000)	(0.000)
EBIT_AT	−0.2032 ***	−0.1755 ***	−0.1823 ***	−0.1823 ***	−0.2511 ***
	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)
M/B	−0.0001	−0.0001	−0.0002	−0.0002	0.0000
	(0.492)	(0.328)	(0.140)	(0.140)	(0.579)
DEP_AT	−0.1812	−0.2691	0.2839	0.2839	0.0169
	(0.315)	(0.117)	(0.142)	(0.142)	(0.938)
LnAT	0.0217 ***	0.0149 **	0.0288 ***	0.0288 ***	0.0370 ***
	(0.000)	(0.001)	(0.000)	(0.000)	(0.000)
PPE_AT	0.0000	−0.0986 ***	0.0394	0.0394	0.0174
	(1.000)	(0.027)	(0.420)	(0.420)	(0.755)
R&D_AT	−0.1503		−0.0555	−0.0555	0.0094
	(0.179)		(0.641)	(0.641)	(0.738)
FINDEF		−0.2178 ***	0.0241	0.0241	−0.0626 **
		(0.000)	(0.359)	(0.359)	(0.029)
Fixed Effects?	Yes	Yes	Yes	Yes	Yes
No. of Obs	1004	999	997	997	1102
R2 Adj.	0.7551	0.1119	0.6396	0.1075	0.1199
Panel B: Parent
	(1) BDR	(2) ΔBDR	(3) MDR	(4) ΔMDR	(5) MDR
BDRit	0.7421 ***	−0.2601 ***
	(0.000)	(0.000)
lagMDRit			0.5774 ***	−0.4226 ***
			(0.000)	(0.000)
EBIT_AT	−0.0359	0.0491 **	−0.1381 ***	−0.1381 ***	−0.2133 ***
	(0.191)	(0.026)	(0.000)	(0.000)	(0.000)
M/B	0.0001 **	0.0001 *	−0.0001	−0.0001	−0.0002 **
	(0.034)	(0.071)	(0.343)	(0.343)	(0.019)
DEP_AT	−0.0602	1.1978 ***	0.2969 **	0.2969 **	0.5070 ***
	(0.655)	(0.000)	(0.047)	(0.047)	(0.004)
LnAT	0.0078 ***	0.0110 ***	0.0088 ***	0.0088 ***	0.0024
	(0.002)	(0.000)	(0.003)	(0.003)	(0.481)
PPE_AT	0.0245	−0.0528 **	−0.0220	−0.0220	−0.1127 ***
	(0.442)	(0.040)	(0.540)	(0.540)	(0.007)
R&D_AT	0.0265		−0.2181 ***	−0.2181 ***	−0.4583 ***
	(0.707)		(0.006)	(0.006)	(0.000)
FINDEF		0.1246 ***	−0.0286	−0.0286	−0.1043 ***
		(0.000)	(0.120)	(0.120)	(0.000)
Fixed Effects?	Yes	Yes	Yes	Yes	Yes
No. of Obs	1646	2590	1611	1611	1708
R2 Adj.	0.8008	0.0691	0.6799	0.1252	0.1490

.

Column (1) of Table 6 establishes the baseline for the BDR regressions. It presents results from a model that explains BDR using only the standard explanatory variables X_i,t, firm and year fixed effects, and a lagged BDR. The partial adjustment model demonstrates a strong fit, with the coefficient on the lagged BDR indicating a rapid adjustment speed—approximately 31% per year. Moreover, the variables intended to reflect target leverage are statistically significant and exhibit the expected signs.

In column (2), the model specification is extended to explain changes in BDR based on variations in the typical explanatory factors and the firm’s financing deficit. The coefficient on the financing deficit (FINDEF) is positive but does not materially affect the signs or significance levels of the other variables. This suggests that pecking order considerations are embedded within a broader trade-off framework rather than acting as an independent determinant of financial leverage.

A similar pattern emerges for parent companies, as shown in Panel B of Table 6. In this case, the coefficient of FINDEF is smaller and statistically insignificant.

5.5. Regression Results on the Welch (2004) Specification

Table 7. This table presents results on Welch (2004) specification. Regression results for the models: (1) MDR_i,t+1 = α₀ + α₁MDR_i,t + α₂IDR_i,t+1 + μ_i,t+1. (2) MDR_i,t+1 = α₀+ (1 − ʎ₁)IDR_i,t+1 + (ʎ₁β) X_i,t + μ_i,t+1. (3) MDR_i,t+1 = α₀ + (1 − ʎ₁)MDR_i,t + (1 − ʎ₂)SPE_i,t+1 + (ʎ₁β) X_i,t + μ_i,t+1. (4) MDR_i,t+1 = α₀ + α₁MDR_i,t + α₂MDR_i,t+1 + α₂SPE_i,t+1 + μ_i,t+1. MDR is the book value of total debt over market value of assets. IDR is spe+lagmdr. EBIT_AT is the earnings before interest and taxes divided by total assets. M/B is the market-to-book ratio. DEP_AT is Depreciation per total assets. LnAT is the natural logarithm of total assets. PPE_AT is Property, Plant & Equipment per total assets. R&D_AT is Research and development expenses per total assets. SPE is calculated as the change in MDR (SPEt+1 = implied debt ratio-MDRt). * indicates significance at the 10% level, ** indicates significance at the 5% level, and *** indicates significance at the 1% level.

Panel A: Subsidiary
	(1)	(2)	(3)	(4)
lagMDRit	0.8500 ***		0.7875 ***	0.5081 ***
	(0.000)		(0.000)	(0.000)
IDR	0.0032	0.0147 ***
	(0.345)	(0.007)
EBIT_AT		−0.3981 ***	−0.1462 ***	−0.1791 ***
		(0.000)	(0.000)	(0.000)
M/B		−0.0008 ***	−0.0003 **	−0.0002
		(0.002)	(0.043)	(0.133)
DEP_AT		−0.4113 **	0.1003	0.2954
		(0.040)	(0.410)	(0.126)
LnAT		0.0176 ***	0.0053 **	0.0282 ***
		(0.000)	(0.010)	(0.000)
PPE_AT		0.1811 ***	0.0359 **	0.0271
		(0.000)	(0.016)	(0.571)
R&D_AT		−0.4923 ***	−0.1337 ***	−0.0550
		(0.000)	(0.009)	(0.644)
SPE			0.0036	0.0002
			(0.282)	(0.950)
Industry			−0.0009
			(0.316)
Fixed Effects?	No	No	No	Yes
No of Obs	1006	1005	1002	1002
R2 Adj.	0.7000	0.2178	0.7149	0.6392
Panel B: Parent
	(1)	(2)	(3)	(4)
lagMDRit	0.8782 ***		0.7918 ***	0.5855 ***
	(0.000)		(0.000)	(0.000)
IDR	−0.0062 ***	0.0003
	(0.001)	(0.937)
EBIT_AT		−0.5216 ***	−0.1671 ***	−0.1427 ***
		(0.000)	(0.000)	(0.000)
M/B		0.0000	0.0000	0.0000
		(0.889)	(0.708)	(0.496)
DEP_AT		0.5615 ***	0.2587 **	0.3264 **
		(0.002)	(0.016)	(0.028)
LnAT		0.0134 ***	0.0056 ***	0.0110 ***
		(0.000)	(0.000)	(0.000)
PPE_AT		0.0771 ***	0.0008	−0.0036 *
		(0.001)	(0.955)	(0.917)
R&D_AT		−0.7915 ***	−0.1237 ***	−0.2074 **
		(0.000)	(0.003)	(0.008)
SPE			−0.0033	−0.0034
			(0.132)	(0.107)
Industry			−0.0018 **
			(0.010)
Fixed Effects?	No	No	No	Yes
No of Obs	2596	1653	1646	1646
R2 Adj.	0.7598	0.2032	0.7443	0.7299

presents the regression outcomes following the specification outlined by Welch (2004) for this data.

The results show that the coefficient of the lagged debt ratio is positive, while the coefficient of the implied debt ratio is close to zero. This suggests that firms allow changes in share prices to permanently affect their leverage, rather than offsetting those changes. As a result, the firm’s target leverage completely reflects the fluctuation in leverage at the beginning of period t + 1. In Column (2), rather than assuming firms adjust toward the previous period’s debt ratio, we employ a model that estimates the firm’s target debt ratio. The IDR comprises two components: the Share Price Effect (SPE) and the Market Debt Ratio (MDR). Firm fixed effects are excluded from the first three columns, although they are critical for precise estimation of the speed of adjustment. Column (4) introduces a baseline specification that includes SPE. Comparing this to Table 4’s final column shows that adding SPE has little impact on the baseline coefficients. The MDR_i,t coefficient of 0.30 indicates that firms close about 30% of the gap toward their target each period, suggesting limited responsiveness to contemporaneous share price changes. Nonetheless, SPE is gradually incorporated into MDR_i,t over time, being offset at roughly 34% annually.

5.6. Estimates Across Forecast Periods

Firm behavior is represented in approximate form by the standard partial adjustment model without specifying the time interval between observations. In Equation (4), we analyze intervals of one to five years to assess whether the estimated adjustment rates are consistent with the assumption that firms move toward their target leverage at a constant proportion each period.

Table 8. This table presents regression results for different forecast horizons for MDR_i,t = (ʎβ)X_i,t + (1 − ʎ) MDR_i,t + ẟ_i,t+1 using different intervals. MDR is the book value of total debt over market value of assets. EBIT_AT is the earnings before interest and taxes divided by total assets. M/B is the market-to-book ratio. DEP_AT is Depreciation scaled by total assets. LnAT is the natural logarithm of total assets. PPE_AT is Property, Plant & Equipment per total assets. R&D_AT is Research and development expenses per total assets. * indicates significance at the 10% level, ** indicates significance at the 5% level, and *** indicates significance at the 1% level.

Panel A: Subsidiary
	2 Years	3 Years	4 Years	5 Years
lagMDRit	0.0633	0.2659 ***	0.2511 ***	0.2363 ***
	(0.572)	(0.000)	(0.000)	(0.000)
EBIT_AT	−0.0357	−0.0862	−0.1379 *	−0.1646 **
	(0.755)	(0.358)	(0.067)	(0.013)
M/B	0.0002	0.0003	0.0002	0.0004
	(0.404)	(0.282)	(0.391)	(0.124)
DEP_AT	0.3339	1.3935 ***	1.2585 ***	0.7390 **
	(0.561)	(0.000)	(0.000)	(0.011)
LnAT	0.0772 **	0.0554 **	0.0474 **	0.0457 ***
	(0.028)	(0.019)	(0.010)	(0.003)
PPE_AT	−0.0682	−0.3363 *	−0.1092	0.0482
	(0.825)	(0.080)	(0.434)	(0.682)
R&D_AT		−0.1573	−0.1757	−0.0238
		(0.554)	(0.449)	(0.897)
Fixed Effects?	Yes	Yes	Yes	Yes
No. of obs	208	297	374	441
R2 Adj	0.1183	0.1556	0.3301	0.4052
Panel B: Parent
	2 Years	3 Years	4 Years	5 Years
lagMDRit	−0.0960	0.2043 **	0.2540 ***	0.3257 ***
	(0.231)	(0.007)	(0.000)	(0.000)
EBIT_AT	−0.1393 *	−0.0803	−0.0558	−0.0642
	(0.057)	(0.210)	(0.289)	(0.181)
M/B	0.0001	0.0004	0.0000	−0.0001
	(0.801)	(0.550)	(0.984)	(0.773)
DEP_AT	−0.3523	0.6677 **	0.6886 **	0.3703
	(0.265)	(0.040)	(0.012)	(0.124)
LnAT	0.0460 ***	0.0118	0.0129	0.0111
	(0.001)	(0.381)	(0.218)	(0.197)
PPE_AT	−0.1774	−0.1518	−0.1133	−0.0588
	(0.242)	(0.169)	(0.175)	(0.418)
R&D_AT	−0.0480	−0.1284	−0.1271	−0.1117
	(0.370)	(0.606)	(0.478)	(0.407)
Fixed Effects?	Yes	Yes	Yes	Yes
No. of obs	314	286	374	464
R2 Adj	0.0132	0.4479	0.5854	0.7047

reports the results across these forecast horizons.

According to the first column of Table 8, 93.67% (1–0.0633) of the initial deviation (computed as 1–0.378) would be corrected within two years. The results over longer time frames align with the idea of a continuous adjustment process. Given the strong alignment between our empirical estimates and the theoretical predictions, we find evidence that a continuous partial adjustment model accurately reflects the observed data variation.

5.7. Alternative Definitions of Leverage

Leverage is defined in various ways in prior literature. We use three different ways to define the market debt ratio:

MDR₁ = (Long Term Debt + Short Term Debt)/(Total assets-Book Equity + Market Equity)

MDR₂ = Total Liabilities/(Total Liabilities + Market Equity)

MDR₃ = Long Term Debt/(Total assets − Currrent Liabilities − Book Equity + Market Equity).

The results using alternative definitions of leverage is shown in

Table 9. This table presents regression results for MDR_i,t = (ʎβ)X_i,t + (1 − ʎ) MDR_i,t + ẟ_i,t+1 using three alternative definitions of the market debt ratio. MDR is the book value of total debt over market value of assets. EBIT_AT is the earnings before interest and taxes divided by total assets. M/B is the market-to-book ratio. DEP_AT is Depreciation scaled by total assets. LnAT is the natural logarithm of total assets. PPE_AT is Property, Plant & Equipment per total assets. R&D_AT is Research and development expenses per total assets. MDR1 is the ratio of long-term and short-term debt to total assets net of book equity plus market equity. MDR2 is the ratio of total liabilities to the sum of total liabilities and market equity. MDR3 is the ratio of long-term debt to total assets net of current liabilities and book equity plus market equity. * indicates significance at the 10% level, ** indicates significance at the 5% level, and *** indicates significance at the 1% level.

Panel A: Subsidiary
	MDR1	MDR2	MDR3
lagMDRit	0.3984 ***	0.5603 ***	0.4206 ***
	(0.000)	(0.000)	(0.000)
EBIT_AT	−0.1565 ***	−0.3304 ***	−0.1323 ***
	(0.000)	(0.000)	(0.000)
M/B	−0.0001	−0.0005 **	−0.0003 **
	(0.301)	(0.008)	(0.021)
DEP_AT	0.1182	0.3103	0.2462
	(0.417)	(0.166)	(0.123)
LnAT	0.0236 ***	0.0242 ***	0.0260 ***
	(0.000)	(0.000)	(0.000)
PPE_AT	0.0067	0.0410	−0.0001
	(0.853)	(0.458)	(0.999)
R&D_AT	−0.0979	0.0124	−0.0636
	(−1.09)	(0.928)	(0.519)
Fixed Effects?	Yes	Yes	Yes
No. of obs	1002	1002	999
R2 Adj	0.6039	0.5875	0.5954
Panel B: Parent
	MDR1	MDR2	MDR3
lagMDRit	0.3793 ***	0.5205 ***	0.4780 ***
	(0.000)	(0.000)	(0.000)
EBIT_AT	−0.0888 ***	−0.4211 ***	−0.0961 ***
	(0.000)	(0.000)	(0.000)
M/B	0.0000	−0.0001 *	−0.0001
	(0.313)	(0.081)	(0.278)
DEP_AT	0.1114	0.2175	0.2402 *
	(0.297)	(0.224)	(0.053)
LnAT	0.0123 ***	0.0199 ***	0.0110 ***
	(0.000)	(0.000)	(0.000)
PPE_AT	0.0562 **	0.0461	0.0088
	(0.026)	(0.276)	(0.764)
R&D_AT	−0.1695 ***	−0.1556 *	−0.1535 **
	(0.002)	(0.097)	(0.017)
Fixed Effects?	Yes	Yes	Yes
No. of obs	1646	1649	1612
R2 Adj	0.6364	0.5771	0.6345

.

Our findings indicate that conclusions regarding target leverage and adjustment speeds remain robust across alternative definitions of leverage. The estimated adjustment speeds range from 44% to 60.16% per year, and the variables influencing target leverage typically exhibit statistically significant coefficients with expected signs.

5.8. Robustness Check

The Hausman test results for parents and subsidiaries are presented in

Table 10. Hausman Test: The Hausman test results for parents and subsidiaries are presented in Table 10.

Hausman Test
	Subsidiary	Parent
	Difference in Coefficients	Difference in Coefficients
lagMDRit	−0.2166	−0.2141
	(0.0163)	(0.0149)
EBIT_AT	−0.0152	0.0170
	(0.0188)	(0.0189)
M/B	−0.0000	−0.0000
	(0.0000)	(0.0000)
DEP_AT	0.1841	0.0270
	(0.1342)	(0.1031)
LnAT	0.0193	0.0056
	(0.0042)	(0.0026)
PPE_AT	−0.0191	−0.0003
	(0.0439)	(0.0323)
R&D_AT	0.1045	−0.0694
	(0.1012)	(0.0656)

with standard errors in parentheses. The Null Hypothesis is that the efficient model (e.g., random effects) is consistent and efficient. The alternative hypothesis is that the efficient model is inconsistent, favoring the consistent model (e.g., fixed effects). As prob > chi2 is small (e.g., <0.05): the null is rejected, indicating that the consistent model (fixed effects) is preferred.

The endogeneity test results for instrumental variable regressions for parents and subsidiaries are presented in

Table 11. Tests of Endogeneity: The endogeneity test results for IV regression in Table 4 are presented in Table 11.

Test of Endogeneity	Null Hypothesis	Results for Subsidiaries	Results for Parents
Durbin (score)	H0: Variables are exogenous	χ2(1) = 227.545 (p = 0.0000)	χ2(1) = 318.669 (p = 0.0000)
Wu–Hausman	H0: Variables are exogenous	F(1,993) = 291.756 (p = 0.0000)	F(1,2679) = 393.015 (p = 0.0000)

. The small p-value leads to the rejection of the null hypothesis, indicating that the instrumental variables are endogenous. This finding highlights the importance of identifying more appropriate instruments in future research.

The baseline regression results using the merger sample for the model MDR_i,t+1 = (ʎβ)X_i,t+1 + (1 − ʎ) MDR_i,t + ẟ_i,t+1 are presented in

Table 12. Regression results using merger sample for the model MDR_i,t+1 = (ʎβ)X_i,t+1 + (1 − ʎ) MDR_i,t + ẟ_i,t+1 where MDR is the market debt ratio. The (lagged) ‘‘X’’ variables determine a firm’s long-run target debt ratio; T-statistics are reported in parentheses. MDR = book value of total debt over market value of assets. The regression results are recorded for mergers during the sample period. The merged firm sample was selected by using distribution code from CRSP events and then the baseline regression was run on this sample. ** indicates significance at the 5% level, and *** indicates significance at the 1% level.

	(1)	(2)	(3)	(4)	(5)	(6)
	Baseline	Fixed Effect Panel	Fixed Effect Panel (with Year Dummy)	Fixed Effect Panel (with Industry Dummy)	Instrumental Variable Panel	Random Effects Panel
lagMDRit	0.8084 ***	0.5236 ***	0.5334 ***	0.5237 ***	1.1377 ***	0.7320 ***
	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)
EBIT_AT	−0.0700 ***	−0.0851 ***	−0.0848 ***	−0.0851 ***	−0.0076	−0.0819 ***
	(0.000)	(0.000)	(0.000)	(0.000)	(0.209)	(0.000)
M/B	−0.0000 ***	−0.0000	−0.0000	−0.0000	−0.0000	−0.0000 ***
	(0.000)	(0.123)	(0.385)	(0.123)	(0.215)	(0.000)
DEP_AT	−0.1292 ***	−0.1185 ***	−0.1359 ***	−0.1185 ***	−0.0184	−0.1631 ***
	(0.000)	(0.000)	(0.000)	(0.000)	(0.415)	(0.000)
LnAT	0.0069 ***	0.0257 ***	0.0311 ***	0.0257 ***	−0.0035 ***	0.0110 ***
	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)
PPE_AT	0.0395 ***	0.1168 ***	0.1028 ***	0.1168 ***	−0.0097 **	0.0477 ***
	(0.000)	(0.000)	(0.000)	(0.000)	(0.029)	(0.000)
R&D_AT	−0.1209 ***	−0.0819 ***	−0.0868 ***	−0.0819 ***	0.0330 ***	−0.1333 ***
	(0.000)	(0.000)	(0.000)	(0.000)	(0.005)	(0.000)
F-F industry factors	0.0019 ***
	(0.000)
Fixed Effects?	No	Yes	Yes	Yes	No	No
No. of obs	17781	17781	17781	17781	17781	17781
R2	0.7392	0.6748	0.6783	0.6748	0.6527	0.7364

. The merged firm sample was selected by using distribution code 5523 from CRSP events, and then the firm-level data was collected from Compustat annual to run a regression on this sample.

The parameter estimates in the two-step GMM are reported in

Table 13. The dynamic panel-data estimation, two-step system GMM results are presented in Table 13. * indicates significance at the 10% level, ** indicates significance at the 5% level, and *** indicates significance at the 1% level.

		Panel A: Subsidiary		Panel A: Subsidiary
		Coefficient		Coefficient
	lagMDRit	0.7861 ***		0.7997 ***
		(0.000)		(0.000)
	EBIT_TA	−0.1473 ***		−0.1613 ***
		(0.003)		(0.000)
	MB	−0.0003 **		0.0000
		(0.017)		(0.657)
	DEP_TA	0.0985		0.2396
		(0.618)		(0.134)
	LnTA	0.0060 *		0.0059 ***
		(0.051)		(0.001)
	PPE_TA	0.0426		−0.0004
		(0.117)		(0.987)
	R&D_TA	−0.1364 *		−0.1476 **
		(0.051)		(0.002)
	N	956		1639
	Instrument	BDR		BDR
Test	Statistic (df)	Value	p-Value	Notes
Arellano–Bond AR(1) (first diff.)	z	−3.99	0.000	Significant, expected sign
	z	−5.03	0.000	Significant, expected sign
Arellano–Bond AR(2) (first diff.)	z	−0.89	0.375	Not significant
	z	−0.00	0.999	Not significant
Sargan test (overid. restrictions)	χ2 (947)	926.89	0.674	Not robust; not weakened by many instruments
	χ2 (1443)	1490.55	0.187	Not robust; not weakened by many instruments
Hansen test (overid. restrictions)	χ2 (947)	101.66	1.000	Robust; weakened by many instruments
	χ2 (1443)	98.44	1.000	Robust; weakened by many instruments
Diff-in-Hansen: GMM levels	χ2 (828) excl.	94.20	1.000	Instruments valid
	χ2 (119) diff.	7.46	1.000	Exogeneity not rejected
	χ2 (1297) excl.	102.05	1.000	Instruments valid
	χ2 (146) diff.	−3.61	1.000	Exogeneity not rejected
Diff-in-Hansen: IV(bdr)	χ2 (946) excl.	96.71	1.000	Instruments valid
	χ2 (1) diff.	4.95	0.026	Exogeneity rejected
	χ2 (1442) excl.	98.89	1.000	Instruments valid
	χ2 (1) diff.	−0.45	1.000	Exogeneity not rejected

, they are consistent and efficient under the model assumptions. The lagMDR coefficients are significant while control variables show different signs than baseline regression. Hansen Test (or Sargan Test)are testto check the validity of the instruments. AR(1) and AR(2) Tests check for serial correlation in the residuals. AR(1) is significant (expected in first-differenced models), while AR(2) is not significant (indicating no second-order serial correlation).

The comparative Lasso Linear Model results for the Subsidiary and Parent are presented in

Table 14. Lasso linear model results are presented in Table 14.

	Subsidiary	Parent
Covariates	7	7
CV folds	10	10
Selected λ (by CV)	0.0009	0.0009
No. of nonzero coefficients (selected λ)	7	5
Variables retained at selected λ	lagmdr, ppebyta, ebitbyta, lnta, rdbyat, mkttobook, depbyta	lagmdr, ebitbyta, lnta, rdbyat, depbyta
Variables excluded at selected λ	None	mkttobook
R-squared (selected λ)	0.7119	0.7406
CV mean prediction error (selected λ)	0.0081	0.0061
No. of Observations	1002	1646

. The regularization strength used in this model was set to 0.0009. Smaller coefficients suggest less influence on the target variable. The models retain the features with non-zero coefficients. The number of nonzero coefficients (selected λ) is 7 for subsidiary and 5 for parent. The magnitude of these coefficients indicates their relative importance. Both models have high explanatory power. None of the variables were excluded for subsidiary. Features with coefficients shrunk to zero are deemed irrelevant and removed. Larger coefficients suggest higher importance, but interpretation depends on feature scaling. When features are highly correlated, Lasso may arbitrarily select one while shrinking others to zero. A higher alpha increases sparsity but may exclude relevant features.

6. Conclusions and Future Implications

Evidence suggests that subsidiary firms tend to target specific capital ratios throughout the sample period. Our findings indicate that a partial adjustment model, which incorporates the firm fixed effects, aligns closely with our observed data. Consistent with prior research, we find that target leverage ratios are influenced by well-established characteristics of the firms. Firms that deviate from their target leverage—whether under- or overleveraged—adjust their leverage ratios relatively quickly to reduce the discrepancy. Parent companies involved in corporate spin-offs exhibit similar behavior but adjust at a slower pace. This targeting behavior is evident in both market-based and book-based leverage measures for parent and subsidiary firms alike. Variables representing the pecking order theory do not show statistically significant effects. Share price fluctuations appear to affect market debt ratios only in the short term. Depending on how leverage is defined, the estimated adjustment speeds vary considerably, ranging from 44% to 60.16% annually. Longer adjustment horizons also support the notion of a continuous adjustment process. The role played by firm variables in making capital structure decisions after a spin-off is of interest to corporate managers. Therefore, this study can serve the interest of managers in assessing the role played by firm characteristics in target leverage attainment.

Many large US corporations are undertaking or examining the possibility of future corporate spin-offs. This could be necessitated by the fact that some companies have previously diversified and grown in unrelated business segments. Some companies might actually be broken up into parts because of anti-trust laws and Government decisions. The stock market also seems to value the sum of the parts of the company greater than the whole, as is evident with the recent share price increase in GE after spinning of GE Healthcare (GEHC) and that of the Chinese tech-giant Alibaba (BABA). There is a suggestion that Google might spin off YouTube as a separate company, and Amazon can be broken down into Amazon Web Services and a separate e-commerce retail business. So Corporate divestiture or spin-off might be a forthcoming trend. In that case, it would be pertinent to examine the capital structure decisions of the spin-off firms because they would be accessing the capital markets for fundraising after their spin-off.