Strategic Management Journal Strat. Mgmt. J., 30: 577–594 (2009) Published online 26 January 2009 in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/smj.752 Received 17 July 2006; Final revision received 11 December 2008

DIVERSIFICATION STRATEGY, CAPITAL STRUCTURE, AND THE ASIAN FINANCIAL CRISIS (1997–1998): EVIDENCE FROM SINGAPORE FIRMS† ELIZABETH NGAH-KIING LIM,1 * SHOBHA S. DAS,2 and AMIT DAS2 1 2

School of Business, University of Connecticut, Storrs, Connecticut, U.S.A. Nanyang Business School, Nanyang Technological University, Singapore

We use agency theory to predict the influence of related and unrelated product diversification on a firm’s level of debt financing. Further, we argue that the link between diversification and capital structure is moderated by the environment in which firms operate. Using SAS PROC MIXED, we fit a mixed-effects model to our unique six-year longitudinal dataset (1995–2000) of 245 publicly listed Singapore firms. Our data spans the period of the Asian Financial Crisis (1997–1998). We find that firms pursuing unrelated product diversification take on less debt financing in stable environments, but more debt financing in dynamic environments. Using longitudinal structural equation modeling, we find a reciprocal relationship between a firm’s product diversification strategy and its debt financing level. Copyright  2009 John Wiley & Sons, Ltd.

INTRODUCTION In their groundbreaking work, Modigliani and Miller (1958) showed that under conditions of absence of taxes, perfect information among players, and negligible transaction costs in capital markets, a manager’s choice between debt and equity financing has no influence on firm value. Since these conditions are rarely attained in the real economy, strategy researchers have relaxed these assumptions to investigate the link between capital structure and strategy. In this study, we use agency theory to examine how diversification strategy influences financing decisions. To add to the existing literature on this Keywords: agency theory; contingency theory; diversification strategy; debt financing; Asian financial crisis; environment

∗ Correspondence to: Elizabeth Ngah-Kiing Lim, School of Business, Department of Management, University of Connecticut, 2100 Hillside Road, Unit 1041, Storrs, CT 06269-1041, U.S.A. E-mail: [email protected] † Elizabeth Lim will join the University of Texas-Dallas in the Fall 2009

Copyright  2009 John Wiley & Sons, Ltd.

topic (e.g., Barton and Gordon, 1988; Kochhar, 1996; Kochhar and Hitt, 1998), we consider the role of the environment (Simerly and Li, 2000) as a moderator of the relationship between strategy and capital structure. We propose that the influence of diversification on debt financing is contingent on whether the environment within which firms operate is stable or dynamic. We suggest that the standard predictions of agency theory might not hold under dynamic conditions when business risks, unpredictability, and uncertainty are high. Our research, therefore, extends existing theoretical explanations of the diversification-capital structure linkage by showing how managerial decision making differs under dynamic versus stable environments. We chose agency theory as our theoretical lens because the agency theory of debt has had a strong influence on strategic management research (Rumelt, Schendel, and Teece, 1994). Agency theory is also suited to this study because of its focus on principal-agent problems under uncertainty. We

578

E N-K. Lim, S. S. Das, and A. Das

use a Singapore sample because in terms of maturity of institutions and capital markets, Singapore falls midway between developed economies, such as the United States and the emerging economies of East Asia (International Monetary Fund, 2008; World Bank, 2008), providing a testing ground for generalizing agency-theoretic arguments beyond the U.S. context. To test our theoretical arguments, we sampled 245 publicly listed Singapore firms for the period 1995–2000, and drew upon the Asian Financial Crisis (1997–1998) as a natural experiment. We analyzed our unique six-year longitudinal dataset using a mixed-model approach (implemented with SAS PROC MIXED), which is more flexible and powerful than simple regression in capturing heterogeneous trends in time-series data. We respond to the calls of Hitt, Gimeno, and Hoskisson (1998) and Shook, Ketchen, Cycyota, and Crockett (2003) to exploit the use of longitudinal designs, sophisticated statistical tools, and analytical techniques capable of capturing complex strategic phenomena. Unlike most prior diversification studies, we also used structural equation modeling to examine the simultaneous longitudinal relationship of diversification and financing, thus casting light on the causal direction between the two constructs. Next, we review some of the literature related to diversification, capital structure, and the environment. Then we briefly describe the Asian Financial Crisis, a salient event that shaped the behaviors of firms during our period of study.

LITERATURE REVIEW Though prior studies in economics and strategy suggest that capital structure has wide-ranging implications for firms, the effect of strategy on capital structure is not conclusive. Barton and Gordon (1988) found that undiversified and relatedstrategy firms (i.e. firms pursuing related diversification) have lower debt levels, while those pursuing unrelated diversification have higher debt levels. Taylor and Lowe (1995) replicated and supported these findings. Yet more support was found by Kochhar and Hitt (1998), who showed that related diversification is commonly associated with equity financing, and unrelated diversification with debt financing. They also found Copyright  2009 John Wiley & Sons, Ltd.

that diversification via acquisition is often associated with public financing, while internal development is usually funded through private financing. Lowe, Naughton, and Taylor (1994) found that highly diversified firms had debt-equity ratios, and characteristics of growth, risk, and cash flow different from less diversified firms. While Singh, Davidson, and Suchard (2003) found a negative relationship between geographical diversification and firm leverage, Low and Chen (2004) found a positive relationship between product diversification and leverage. Interestingly, Men´endez-Alonso (2003), using a sample of 480 Spanish manufacturing firms, did not find support for the effect of firm diversification on debt levels. Jahera and Lloyd (1996) reported that asset type, firm diversification, and availability of tax shields are the strongest determinants of firm leverage. Booth, Aivazian, Demirguc-Kunt and Maksimovic (2001) found that the more profitable the firm, the lower the debt ratio, and that asset tangibility affects total and long-term debt decisions differently. Allayannis, Brown, and Klapper (2003) suggest that the lower cost of foreign currency debt and the need to access deeper foreign capital markets consistently explain the use of debt by East Asian nonfinancial firms. At a more detailed level, Vicente-Lorente (2001) found that specific and opaque strategic resources such as research and development (R&D) and human capital relate inversely to leverage; however, reputational assets are positively correlated with debt ratio. There is mixed support for firm-specific determinants of capital structure. Balakrishnan and Fox (1993) found that firm-specific characteristics strongly affect leverage. On the contrary, Harris (1994) tested investment and financing decisions at firm-specific levels in a sample of long-standing Fortune 500 firms and found evidence contrary to Williamson’s (1985) transaction cost economics (TCE) theory of capital structure. Crutchely and Jensen’s (1996) study supports Jensen’s (1986) free cash flow theory of leverage. Consistent with the agency costs hypothesis, Berger and diPatti (2006) showed that higher leverage is significantly associated with higher profit efficiency. Berger, Ofek, and Yermack (1997) found that entrenched managers seek to avoid debt, and Morellec (2004) demonstrated that manager-shareholder conflicts explain low debt levels. However, Kochhar (1996), using content analysis, suggests that debt versus equity financing is important for governance of Strat. Mgmt. J., 30: 577–594 (2009) DOI: 10.1002/smj

Diversification and Capital Structure in the Asian Financial Crisis firm resources, but not for governance of free cash flows. Crises, such as the Asian Financial Crisis of 1997–1998, often impact the capital structure of firms. Ho (2004) showed that the financial turmoil in East Asia in 1997–1998 worsened the debt ratio of many firms. Firms in Hong Kong and Singapore that had lower leverage were therefore less exposed; however, firms in Thailand, South Korea, Indonesia, and Malaysia that had higher leverage before the crisis suffered more. Braun and Larrain (2005) found that recessions have greater impact on industries that naturally rely more on external funds and soft assets. Based on a study of South Korean firms during 1992–2001, Fattouh, Scaramozzino, and Harris (2005) showed that the share of debt in the capital structure before and after the Asian Financial Crisis evolved differently across firms. Clearly, a large body of work has studied the determinants of capital structure, with most researchers finding a significant link between diversification strategy and debt financing. While equivocal results were obtained for firm-specific determinants, the Asian Financial Crisis has been generally found to have significant effects on capital structure in the financial economics literature. Therefore, it seems a natural progression to study the moderating effect of the environment on the relation between diversification and debt financing level. Heeding Thompson’s (1967) call to examine constraints and contingencies residing within and outside organizational boundaries, we posit that the interaction of environmental conditions and managerial choice with respect to diversification strategy is a key driver of heterogeneity in the level of debt financing across firms. Scholars have long argued that both managerial choice and the environment determine firm behavior (Chandler, 1962; Child, 1972; Cyert and March, 1963). Maintaining the same strategy under different environmental conditions might not allow firms to compete effectively for capital resources. As the environment changes, we expect managers to adjust their strategy accordingly. In the empirical context, Venkatraman (1989) argued that the way ‘fit’ is conceptualized in a research effort has important implications for the formulation and testing of relationships. In the present study, fit is conceptualized as a theoretically defined match between diversification strategy and the environment that influences the amount of debt firms Copyright  2009 John Wiley & Sons, Ltd.

579

assume in their capital structure. In the next section, we describe key features of the Asian Financial Crisis (1997–1998) that characterize it as a period of environmental dynamism as conceived by Dess and Beard (1984).

THE ASIAN FINANCIAL CRISIS (1997–1998) In the decade prior to 1997, East Asian economies achieved rapid economic growth, with average annual gross domestic product (GDP) growth of close to eight percent in ASEAN-5.1 Asian economies engaged in ‘gradual financial market liberalization, and maintained effectively pegged nominal exchange rates’ (Harvey and Roper, 1999: 32). Firms invested heavily in tangible fixed assets and often borrowed money to finance their expansion. To avoid diluting management control, debt financing was preferred over equity. As a result, debt levels skyrocketed to a record high in the mid-1990s, with ‘the average [publicly] listed firm in South Korea and Thailand [having] a debt-toequity ratio of 3.5 and 2.3, respectively, relative to ratios in the United States and Germany of 1.1 and 1.5, respectively’ (Allayannis et al., 2003: 2671). In addition to heavy debt loads, rapid annual growth rates of 30–40 percent stretched both managerial capacity and distribution and marketing channels of firms (Pomerleano, 1998). The Asian Financial Crisis was the most severe shock to hit Asian countries and firms in 50 years (Singh and Yip, 2000). It started on 2 July, 1997 with a 20 percent devaluation of the Thai baht from its peg to the U.S. dollar. The contagion quickly spread throughout Asia and to the United States, Latin America, and Europe, resulting in loss of confidence. Asian stock markets were affected adversely, with the Singapore Stock Exchange Straits Times Index (STI) plummeting sharply. The Singapore dollar fell from a high of SGD1.43 per U.S. dollar on 1 July, 1997, to SGD1.75 per U.S. dollar on 7 January, 1998, a decline of 18.3 percent in six months (Ngiam, 2000). Other regional currencies depreciated from 35 percent (the Philippine peso) to 70 percent (the Indonesian rupiah). Although the Singapore dollar depreciated against the U.S. dollar (making imports costlier), 1 The five original members of the Association of South East Asian Nations (ASEAN) are Indonesia, Malaysia, the Philippines, Singapore, and Thailand.

Strat. Mgmt. J., 30: 577–594 (2009) DOI: 10.1002/smj

580

E N-K. Lim, S. S. Das, and A. Das

it appreciated against the regional currencies (decreasing the competitiveness of regional exports). General business and consumer sentiment in Singapore’s domestic market was adversely affected by the regional downturn (Chia, 2000). Singapore’s GDP growth rate fell from 8.8 percent in 1997 to 0.5 percent in 1998 (Stein and Lim, 2004) as a result of the crisis-induced fall in regional and global demand for electronics, its major export. Business slowdowns and demand uncertainties led to a decline in corporate earnings, a downsizing of business operations, disruptions in the product and financial markets, problems with buyer and supply chains, and an increase in financial and political risks associated with doing business in the region. By June 1998, the tumultuous crisis started to recede, exchange rates began to stabilize, and equity markets rebounded by a whopping average of 75.4 percent (Allayannis et al., 2003). The region’s domestic currencies started to appreciate against the U.S. dollar, and exchange rate volatility decreased. How did this period of turmoil affect the capital structure decision of Singapore firms, especially in regard to debt financing? How did firms’ strategy with respect to diversification, related and unrelated, affect their appetite for debt? We derive specific hypotheses about these issues in our theory section. But, first, we conceptualize the Asian Financial Crisis as a period of environmental dynamism associated with uncertainty and volatility. A period of environmental dynamism Most theorists agree that the environment is multidimensional, with multiple and different effects on organizational characteristics (Keats and Hitt, 1988). In particular, Aldrich (1979) developed a range of six environmental dimensions: capacity, homogeneity, stability, concentration, consensus, and turbulence. Dess and Beard (1984) later collapsed five of these dimensions into a parsimonious set of three: munificence, dynamism, and complexity. Munificence is defined as resource abundance with slack and capacity to support growth. Dynamism is concerned with instability and volatility. Complexity reflects heterogeneity and dispersion. We focus on a single aspect of Dess and Beard’s (1984) environmental dimensions, that of Copyright  2009 John Wiley & Sons, Ltd.

dynamism, for two reasons. First, environmental dynamism captures the unpredictability and volatility of the Asian Financial Crisis, a period marked by instability and difficult-to-predict discontinuities (Aldrich, 1979; Dess and Beard, 1984). Second, scholars have highlighted the importance of distinguishing between the rate and unpredictability of environmental change (Miles, Snow, and Pfeffer, 1974; Jurkovich, 1974). Dess and Beard (1984) point out that the study of dynamism should be restricted to change that is hard to predict and lacks a pattern, and thus increases the level of uncertainty. We believe that the unpredictable, unsystematic, and uncertain changes that resulted from the Asian Financial Crisis characterize a dynamic environment. Conversely, the higher predictability, lower uncertainty, and lower volatility of the period leading up to 1997 (i.e., before the Asian Financial Crisis) represent a stable environment. Unpredictable and volatile, the Asian Financial Crisis was a period of very low munificence because resource scarcity and business slowdowns stunted growth in Singapore and the region. At the same time, complexity was low because organizations streamlined their operations and reduced the range and heterogeneity of their activities (Child, 1972; Thompson, 1967; Duncan, 1972) in response to falling consumer demand. Thus, the Asian Financial Crisis clearly fits Dess and Beard’s (1984) characterization of a dynamic environment. In what follows, we draw upon agency theory to hypothesize relationships between different types of diversification strategy and debt financing level. Then, we use contingency theory to refine the predictions of agency thinking under the moderating influence of the environment. We provide reasons why standard agency predictions that hold in stable environments might be reversed under environmental dynamism, thus testing the boundary conditions for agency theory in dynamic environments.

THEORY AND HYPOTHESES Agency theory, which emerged in the 1970s, extended our earlier understanding of economic relationships under conditions of uncertainty and imperfect information (Berle and Means, 1932; Coase, 1937). Agency logic suggests that, under the separation of ownership and control, conflicts arise between managers and owners of the firm Strat. Mgmt. J., 30: 577–594 (2009) DOI: 10.1002/smj

Diversification and Capital Structure in the Asian Financial Crisis due to divergence of goals and interests and to the different risks borne by shareholders and managers (Jensen and Meckling, 1976). Agency problems are expected to arise when principals contract agents on their behalf. As principals of the firm, owners seek to maximize profit, while managers hold their own self-interest uppermost. In the presence of information asymmetry, owners cannot obtain, assess, and interpret all information on opportunistic managerial behavior. An example of managerial self-interest is the pursuit of unrelated diversification to reduce employment risk (Amihud and Lev, 1981), increase compensation levels (Tosi and Gomez-Mejia, 1989), or build empires for managerial aggrandizement (Simerly and Li, 2000). To control and align managers’ interests with those of the principals, the latter establish incentive structures and monitoring mechanisms. The difference between the payoffs to owners when interests are aligned and when they are misaligned is referred to as agency cost. As it is costly to achieve alignment, it is rational to incur agency costs up to a point (Jensen and Smith, 1985). Free cash flow is defined as ‘cash flow in excess of that required to fund all projects that have positive net present values when discounted at the relevant cost of capital’ (Jensen, 1986: 323). Jensen (1986) argued that debt financing can reduce the agency costs related to free cash flows by reducing the discretionary funds available to managers, who are contractually bound to make timely interest payments on the debt capital raised (Kochhar, 1996). Unrelated diversification Jensen’s (1986) free cash flow thesis predicts that managers of firms with excess cash undertake lowreturn activities against the interests of shareholders instead of paying out dividends, because such payouts reduce the resources they control, diminish their power, and put them at the mercy of the external capital markets (see Rozeff, 1982; Easterbrook, 1984). Managers are likely to diversify their business when there is excess capacity of productive factors (Penrose, 1959). Instead of returning money to shareholders, self-interested managers are more likely to redirect free cash flows toward unrelated diversification, which focuses on financial synergies using the economies of the internal capital market. While managers might not embrace Copyright  2009 John Wiley & Sons, Ltd.

581

bad projects that could jeopardize their reputation or their worth in the labor market, it is reasonable to expect managers to allocate resources to projects that bring them personal economic benefits. As Hambrick and Finkelstein (1995) point out, managers in management-controlled firms strive to maximize their economic interests while appearing to adhere to basic conventions of contemporary business practices. Repayment of debt reduces the level of free cash flow. Monitoring and control by holders of the firm’s debt demand that managers be prudent in their expenditures. Under the watchful eyes of debt holders, managers have limited opportunities to splurge on high-ticket items or engage in shirking and free-riding behaviors; they must focus on activities consistent with the profit-maximizing interests of shareholders. Default on interest and principal repayments entitles debt holders to legal redress, such as bankruptcy court. Debt holders can force managers to trim their expansion programs and to divest or sell units to make timely repayments. Further, banks and bond markets might be less willing to lend funds to firms that follow an unrelated diversification strategy, especially if they believe their investment is not safeguarded (Kochhar, 1996). Even if banks agree to lend, the higher interest rates demanded might adversely impact the firm’s free cash flows. Thus, we predict that firms pursuing a strategy of unrelated diversification are less likely to take on additional debt to fund their expansion. Hypothesis 1: Unrelated diversification strategy is negatively associated with debt financing level. Related diversification Firms pursuing related diversification are more likely to finance their projects with debt. Related diversification requires the sharing of activities and the transfer of skills across businesses to increase firm value from operational synergies. Funds obtained from banks and corporate bond markets can be used to facilitate operations across units, build interdependencies, and generate synergies across businesses to create value. As managers focus on achieving operational synergies and cost savings, debt holders may be more willing to lend money to the firm and may be less likely to scrutinize and interfere in firm operations. Strat. Mgmt. J., 30: 577–594 (2009) DOI: 10.1002/smj

582

E N-K. Lim, S. S. Das, and A. Das

Hypothesis 2: Related diversification strategy is positively associated with debt financing level. Our base hypotheses link diversification strategy to debt financing level based on Jensen’s (1986) free cash flow thesis. By differentiating between related and unrelated diversification, we offer more nuanced explanations for the relationships between diversification strategy and debt financing level. Next, we build upon the main relationships by accounting for the moderating role of the environment. The moderating role of the environment Lawrence and Lorsch (1969) argue that an organization’s internal states and processes should be consistent with external demands. Using their argument, we develop an analytic framework to anticipate the contingent influence of the external environment on the relationship between diversification and debt financing level. In other words, we argue that variations in debt financing level depend on the interactions between levels of diversification (related and unrelated) and environmental conditions. Unrelated diversification and environment Free cash flow can be channeled by self-interested managers to unrelated diversification, even if it potentially reduces firm value. In a stable environment, debt holders observe and recognize the strategies of managers to control agency problems (Simerly and Li, 2000) and may impose controls that limit managerial discretion. To avoid scrutiny and interference by debt holders, managers undertaking unrelated diversification are likely to shun debt funding. Dynamic environments characterized by uncertainty and instability (Keats and Hitt, 1988) make it difficult for debt holders to assess the quality of managerial decisions. In such environments, management decisions involve more intense information processing (Galbraith, 1973), and unpredictability clouds the relation of these decisions to their consequences. Under such fluid conditions, debt holders are less effective in their control and monitoring roles than in a stable environment. A dynamic environment also increases managerial discretion (Haleblian and Finkelstein, 1993), influence, and control over the organization at the Copyright  2009 John Wiley & Sons, Ltd.

expense of debt holders. Where efficiency is difficult to achieve (Keats and Hitt, 1988), managers can pursue unrelated diversification without being heavily penalized by boards of directors or debt holders. Thus, managers can engage in organizational tactics such as buffering, collusion, and long-term contracts (Dess and Beard, 1984) without fear of censure by debt holders. The debt monitoring mechanism designed to control agency problems is less effective due to insufficient knowledge for critical decision making (Milliken, 1987), and self-serving managers are emboldened to raise debt financing to pursue unrelated diversification. In sum: Hypothesis 3: The relationship between unrelated diversification and debt financing level is negatively moderated by stable environments, but positively moderated by dynamic environments. Related diversification and environment Related diversification is often viewed as value enhancing in the hope that firms might exploit synergies with existing operations (Rumelt, 1974; Montgomery and Wernerfelt, 1988). In stable environments, such firms might face a lower cost of capital when they choose debt financing because debt holders are more confident of recovering capital and interest from their investments. However, environmental dynamism can deter such firms from assuming and servicing additional debt. High environmental uncertainty increases the cost of debt financing across the board for all borrowers. Thus, firms pursuing related diversification may choose to wait out environmental instability or pursue other financing options. Hypothesis 4: The relationship between related diversification and debt financing level is positively moderated by stable environments, but negatively moderated by dynamic environments. Reciprocal interdependence Some researchers (Kochhar and Hitt, 1998; Chatterjee and Wernerfelt, 1991) have argued that a firm’s capital structure may affect diversification strategy, which is opposite to the causal direction suggested by Hypotheses 1 and 2. Repayment on debt decreases free cash flows available for Strat. Mgmt. J., 30: 577–594 (2009) DOI: 10.1002/smj

Diversification and Capital Structure in the Asian Financial Crisis wasteful investments by managers, which results in lower agency costs. Higher monitoring and control by debt holders also deters managers from undertaking value-decreasing unrelated diversification (Jensen, 1986), and encourages value-increasing related diversification. Taken together with the logic underlying our base hypotheses, these arguments suggest a reciprocal interdependence between diversification strategy and debt financing, wherein the choice of diversification strategy influences debt financing, while debt financing constrains diversification decisions.

METHODS Sample We developed a unique dataset from an initial population of 370 publicly listed firms in Singapore to test our hypotheses. These firms represented all 10 industries listed on the Main Trading Board of Singapore (SGX) and eight industries in the Second Trading Board (SESDAQ, or the Stock Exchange of Singapore Dealing and Automated Quotation system), ensuring our results generalize across industries (see Table 1). Consistent with prior diversification research, we excluded firms in the financial services industry, which are subject to government regulations that might place constraints on corporate strategy. We also removed firms that did not have annual Table 1.

583

reports in the International Resource Center (formerly known as the SGX Library) in Singapore or for which segment information and Singapore Standard Industrial Classification (SSIC) data were not available for the period 1995–2000. Our longitudinal dataset for the period 1995–2000 (the data before 1995 contained too many missing values) includes information on sales of product segments and SSIC data for 245 firms, which was collected through a highly resource-intensive process. Publicly listed firms incorporated in Singapore are required by the Statement of Accounting Standard Number 23 (issued by the Institute of Certified Public Accountants of Singapore) to reveal segment information in their financial reports (Chen and Ho, 2000). Using annual reports, we manually gathered sales data by product segments for each company in each fiscal year of our study.

Independent variables We used an SIC-based entropy index to measure diversification, a measure with demonstrated construct validity (Hoskisson et al., 1993) that has been extensively used in strategic management research (eg., Palepu, 1985; Baysinger and Hoskisson, 1989; Hitt, Hoskisson, and Kim, 1997). We chose an entropy index over Rumelt’s (1974, 1982) subjective classification with specialization ratio, or SIC count measure, because it avoids

Industry breakdown of Singapore firms listed on SGX and SESDAQ

Industry

Commerce Construction Engineering Hotels/restaurants Investment holding/trading Information technology Manufacturing Multi Properties Services Transport/storage/communications Others Total

Counts SGX

SESDAQ

35 14 2 12 2 1 79 14 19 11 0 1 185

7 3 0 1 1 0 22 0 0 2 19 0 60

Total counts

Percentage

42 17 2 13 3 1 101 14 19 13 19 1 245

17.14% 6.94% 0.82% 5.33% 1.23% 0.41% 41.22% 5.71% 7.76% 5.33% 7.76% 0.41% 100%

Notes: SGX = Main Trading Board of Singapore; SESDAQ = Second Trading Board. Percentages do not add up to 100% due to rounding errors. Copyright  2009 John Wiley & Sons, Ltd.

Strat. Mgmt. J., 30: 577–594 (2009) DOI: 10.1002/smj

584

E N-K. Lim, S. S. Das, and A. Das

aggregation biases where ‘the lines of business created may reflect researcher values (biases) more than managerial identified lines of business’ (Hoskisson et al., 1993: 232). Also, as Baysinger and Hoskisson (1989) state, an entropy-based diversification measure applies to both related and unrelated diversification, aligning well with our research questions. Entropy measures the diversity of a firm’s business with respect to the number and relative size of segments. The entropy index, a continuous measure of diversification, is computed as
Pi ln(1/Pi ), where Pi is the sales attributed to segment i, and ln(1/Pi ), the logarithm of the inverse of sales, is the weight for each segment i. Separate indices DR and DU are computed for related and unrelated diversification, respectively; the total diversification index, DT, is the sum of these two. Industries in Singapore are classified using fivedigit SSIC codes. Since firm-level SSIC data are limited in Singapore, we mapped each firm’s product segments into SSIC codes for each study year using information from three different sources.2 We recruited college students on a part-time basis to help with data collection and coding tasks. They were thoroughly trained by the researchers before being allowed to work on their own. To ensure accuracy and reliability, two student coders independently examined all products associated with 30 companies, selected at random, for one year. The interrater agreement for coding 225 product segments was 97.3 percent (i.e., the coders arrived at the same SSIC codes for 219 of the 225 product segments). After resolving the coding differences

on the remaining six product segments, the coders worked on different sections of the dataset. Product segments were classified as related if they contained the same first three digits of the SSIC codes, indicating the segments belonged to the same industry subcategory, and were classified as unrelated if they did not. Related diversification (DR) has a minimum value of zero; the higher the DR value, the greater the relatedness. Similarly, unrelated diversification (DU) has a minimum value of zero; the higher the DU value, the greater the unrelatedness. Entropy could not be defined at the four-digit SSIC code level due to numerous missing observations, while measurement at the two-digit SSIC code level proved too coarse. Three-digit SSIC codes largely mitigated these problems. To measure environment, the moderating variable, we collected time-series data on the weekly levels of the Singapore Stock Exchange Straits Times Index (STI) from the Yahoo Web site (http://sg.finance.yahoo.com).3 Since our firm-level variables were collected on a yearly basis, the STI data were averaged for each calendar year. Dependent variable Firm leverage, or the level of debt financing, was operationalized as the ratio of long-term debt to total capital (Geringer, Tallman, and Olsen, 2000). We transformed the debt financing variable into its natural logarithm to remove skewness and obtained an approximately normally distributed variable. Control variables

2

The Singapore Standard Industrial Classification (SSIC) codes and the corresponding product descriptions of all businesses for each firm during 1995–2000 were collected from multiple sources. These included firm annual reports, the Singapore Standard Industrial Classification (SSIC) (2000) published by the Singapore Department of Statistics at the Ministry of Trade and Industry, and an online SSIC search Web site managed by Statistics Singapore (http://www.singstat.gov.sg/stats/resources.html). Unlike the United States, the Stock Exchange of Singapore (SES) does not use SIC codes to classify listed firms (Chen and Ho, 2000). Data are limited because the sources do not provide complete SSIC code information associated with the various product lines or businesses of each firm during the study period. For example, firm annual reports present product descriptions but not the corresponding SSIC codes. The SSIC (2000) publication contains information on SSIC codes and product descriptions for each industry, but not the firm. The online SSIC Web site provides information on SSIC codes and product descriptions for each industry, but not the study year. By triangulating these sources, we compensate missing data in one source with data from the other two sources. Copyright  2009 John Wiley & Sons, Ltd.

Control variables accounted for characteristics of Singapore firms and any systematic influence of extraneous factors. We controlled for firm size (the natural logarithm of total assets) because larger firms tend to have more resources to undertake diversification (Lu and Beamish, 2001; Tallman and Li, 1996; Geringer et al., 2000), and for firm 3 Given that the Singapore currency was strongly affected by the Asian Financial Crisis, we also measured environment using monthly Singapore dollar (per USD) currency exchange rate data for the period 1995–2000 collected from the International Financial Statistics (IFS) database. Our analyses revealed that the findings remained robust. Henceforth, we chose stock market STI for data reporting because we have weekly data points (52 weeks × six years), which capture more timely environmental fluctuations that can be averaged to provide more accurate annual estimates.

Strat. Mgmt. J., 30: 577–594 (2009) DOI: 10.1002/smj

Diversification and Capital Structure in the Asian Financial Crisis age (natural logarithm of firm age) to account for the fact that younger firms may have more fluid capital structures (Preece, Miles, and Baetz, 1998; Chen and Ho, 2000). We used return on assets (ROA, log-transformed to achieve normality) to control for firm performance (Beamish and daCosta, 1984; Tallman and Li, 1996; Hitt et al., 1997) because it correlates well (r ≥ 0.9) with other measures such as return on equity (ROE) and return on sales (ROS). Because almost 60 percent of Singapore firms belong to two broad industries (see Table 1)—manufacturing and commerce—we dummy-coded firm industry type using three levels: 1 = ‘manufacturing,’ 2 = ‘commerce,’ and 3 = ‘others.’ Other than segment-related data, all firm-level data were gathered from the Research Insight database and various volumes of Financial Highlights of Companies on the Singapore Exchange published in 2000 by the Center for Business Research and Development of the Faculty of Business Administration at the National University of Singapore.

ANALYSES SAS PROC MIXED modeling Given that the Asian Financial Crisis occurred in 1997–1998, we split the total period of the study (1995–2000) into a precrisis period to reflect a stable environment (1995–1997), and a crisis and recovery period to reflect a dynamic environment (1998–2000). Separate models were estimated for the two subperiods. To assess whether a mixed modeling approach was appropriate for our longitudinal analyses, we first determined whether we needed a randomeffect model. We ran a null model containing only one parameter, a random intercept, which partitions the total variation in the data into withinfirm and between-firm components. The intraclass correlation (ICC) computed from the null model assesses whether a random-effect model is required. The numeric formula for the ICC is I CC =

σˆ v20 σˆ v20 + σˆ 2

where σˆ v02 is the variance explained by the null model, and σˆ 2 is the residual variance. Copyright  2009 John Wiley & Sons, Ltd.

585

Computation of covariance parameter estimates indicated ICC values for the stable and dynamic environments as 1.50/(1.50 + 1.08) = 0.58, and 2.51/(2.51 + 1.09) = 0.70, respectively, suggesting that 58 percent and 70 percent of variation is explained by allowing the intercept to vary across firms. Because the statistically significant values for within-firm variations suggested that the structure of our data is best captured with a randomeffect model, we allowed the intercept to vary. Since we have no a priori theoretical reasons to assume any of the other predictors in the model are randomly distributed, we employed SAS procedure PROC MIXED (SAS Institute Inc., 1992), which estimates and tests for statistical significance in the fixed-effect predictors of interest and the random intercept. The general form of the mixed model is: Yi = Xi β + Zi γi + εi where, for firm i, Yi is a vector of p responses, Xi is a k × p fixed-effects design matrix, Zi is a random-effects design matrix, γi is a vector of random effects associated with Zi , and εi is vector of random errors. The model is ‘mixed’ because it contains both a fixed effect β and a random effect γi . While fixed effects are assumed to be measured with no error, random effects are assumed to be samples from a distribution of values. We also assume γi. and εi are independent, where γi. is normally distributed with mean γ and variance-covariance matrix , and εi is normally distributed with mean 0 and variance-covariance matrix Ri . Compared to traditional regression models, mixed-effects models are more flexible, incorporating the influence of both fixed and timevarying independent variables on the dependent variable. They account for part of the ‘error’ variance in purely fixed-effects models, leading to ‘more efficient estimates and more powerful tests’ (Bagiella, Sloan, and Heitjan, 2000: 13). In addition, mixed-effects models handle missing observations and unbalanced designs more effectively than MANOVA and repeated-measures ANOVA, which require deletion of cases with missing observations. Parameters of mixed-effects models are estimated using a maximum likelihood method that examines all available data, including cases with missing observations (Dempster et al., 1984; Laird and Ware, 1982). They are also parsimonious in Strat. Mgmt. J., 30: 577–594 (2009) DOI: 10.1002/smj

586

E N-K. Lim, S. S. Das, and A. Das

that they ‘estimate the minimum number of parameters necessary to obtain an adequate fit’ (Bagiella et al., 2000: 15), thereby freeing up more degrees of freedom for testing fixed effects. Finally, the maximum likelihood method of estimation captures the variance in the data more accurately, leading to more precise estimates of coefficients (i.e., narrower confidence intervals). This study also responds to the calls of Ramanujam and Varadarajan (1989) and Keats (1990) for more time-variant diversification research. While prior studies have proxied uncertainty with R&D intensity (Dean and Meyer, 1996), ROA change (Miller and Leiblein, 1996), and sales growth (Dean and Meyer, 1996), we measured environment more directly by using the STI stock exchange index. To address potential serial autocorrelation in the independent and dependent variables, we specified a homogeneous first-order autoregressive AR(1) covariance structure, which assumes that measures close together in time are more highly correlated than those distant in time. In particular, AR(1) assumes that a measure in period one is more strongly correlated with the measure in period two than with the value in period three or in more distant periods. The AR(1) model is more realistic than assuming independence across time periods and, therefore, has more power to detect significant changes. We implemented a three-stage AR(1) approach: Yt = Øo + Ø1 Yt−1 + δt ∼ N(0, σ 2 ) ˆ o and Ø ˆ 1 to OLS regression4 1) Fit Ø ˆ ˆ t−1 2) δˆt = Yt − Øo + Ø1 Y T 3) σˆ 2 = (
t=2 δt2 )/((T − 2 + 1) − 2). We used time-series analysis to obtain environment values. When running the PROC MIXED analyses, we used the mean squared error (MSE) from the AR(1) model for the environment variable. The MSE is based on the predicted values which, in turn, are based on the coefficient estimates. The ordinary least squares (OLS) is routinely used to estimate the coefficient(s) of the AR time series. For stationary AR(1) the data is Gaussian and, therefore, the OLS estimator is expected 4 For stationary AR(1) data, the error term satisfies the GaussMarkov assumptions: (1) E(εi ) = 0, (2) Var (εi ) = σ 2 < ∞ (all errors have same variance, i.e., homoscedasticity), and (3) Cov (εi , εj ) = 0, for i = j (i.e., uncorrelatedness).

Copyright  2009 John Wiley & Sons, Ltd.

to be BLUE (best linear unbiased estimator) by the Gauss-Markov theorem. We checked our data for normality (using Shapiro-Wilks and KolmogorovSmirnov tests) and heteroscedasticity (using scatter plots) and found the data to be well behaved in both respects. To test our base hypotheses, we first ran one PROC MIXED model in SAS version 9.1 to estimate the main effects of diversification strategies (related and unrelated) on debt financing level. Next, we ran two PROC MIXED models, one for the stable environment (1995–1997) and one for the dynamic environment (1998–2000). In each of these two models, our predictors of debt financing levels are diversification strategy (related and unrelated), the measure of environment over the three-year window, and the interactions of these two variables.5 Structural equation modeling Some researchers have suggested that the effects of capital structure and corporate strategy should be examined simultaneously. We examined the causal direction between them with four longitudinal structural equation models implemented using AMOS version 6.0: two models for related diversification (one constrained and one unconstrained), and two for unrelated diversification (again, one constrained and one unconstrained). Constraints, in this context, refer to setting the path coefficient from lagged-strategy to debt and from lagged-debt to strategy as equal. The unconstrained models are free from this requirement. Figure 1 shows the coefficient estimates for the related diversification strategy (values from the unconstrained model in parentheses), and Figure 2 shows the coefficient estimates for the unrelated diversification strategy (values from the unconstrained model in parentheses). The chi-square value for each constrained model is compared against the chi-square value for the corresponding unconstrained model. A significant chi-square difference between the constrained and unconstrained models rejects the hypothesis that 5 We also ran all three PROC MIXED models separately for related and unrelated diversification. Given that the parameter estimates and significance levels remained robust across the related and unrelated diversification modes, we report, for the sake of parsimony, the results of models where both forms of diversification are present.

Strat. Mgmt. J., 30: 577–594 (2009) DOI: 10.1002/smj

Diversification and Capital Structure in the Asian Financial Crisis

587

-0.03 (-0.05)

e1

Debt financing 95–96

0.49*** (0.48***) a=-0.00† (0.06)

0.04 (0.04)

e2 R2=0.24 (0.24)

Debt financing 97–98

0.73*** (0.74***) c =0.00 (0.04)

b=-0.10† (-0.10†) Related strategy 95–96

0.50*** (0.50***)

R2=0.51 (0.51) Debt financing 99–00

d=0.02 (0.02)

Related strategy 97–98

0.58*** (0.58***) R2=0.25 (0.25)

Related strategy 99–00 R2=0.21 (0.21) e4

e3

-0.15 (-0.15)

Figure 1. Longitudinal structural equation model showing standardized coefficient estimates for the constrained (and unconstrained) relationships in Model 1 (and Model 2), respectively, between related strategy and debt financing Notes: Debt financing 95–96 refers to firm debt financing level averaged over two years during 1995–1996, Related strategy 95–96 refers to related strategy averaged over two years during 1995–1996, and so forth. ‘a’ and ‘b’ as well as ‘c’ and ‘d’ in bold are equality constraints in Model 1. Model 2 is the same as Model 1, except that no constraints are imposed on its parameters. Values for unconstrained relationships in Model 2 are in parentheses. ∗∗∗ p < 0.001, ∗∗ p < 0.01, ∗ p < 0.05, †p < 0.10

the constrained path coefficients (from laggedstrategy to debt, and from lagged-debt to strategy) are equal. In other words, a significant chi-square difference tentatively indicates that one variable exerts more influence than the other, and that the relationship is unidirectional. A non-significant chi-square difference, on the other hand, leads to the conclusion that the variables are reciprocally interdependent.

RESULTS Table 2 shows descriptive statistics and bivariate correlations for the study variables. The natural logarithm of the mean age of Singapore publicly listed firms in the sample is 3.06, with a standard deviation of 0.65 (mean firm age of 21.35 years, and standard deviation of 1.92 years). As shown in Table 1, the majority of these young Copyright  2009 John Wiley & Sons, Ltd.

firms (41.22%) are in the manufacturing industry, followed by commerce (17.14%). The natural logarithm of the mean firm ROA stands at 1.23, with a standard deviation of 1.14 (mean firm ROA of 3.41, and standard deviation of 3.12). The logarithm of the mean firm size, measured by total assets, is 12.28. Using an exchange rate of SGD1.60 per one U.S. dollar (an average level for the study period), the average firm in our sample holds total assets of about U.S.$134.2 million (or SGD214.7 million). Hypothesis 1 states that unrelated diversification strategy is negatively associated with firm debt financing level, and Hypothesis 2 states that related diversification strategy is positively related with firm debt financing level. Table 3 indicates that the main effects of unrelated diversification and related diversification on firm debt financing levels are insignificant. Therefore, Hypotheses 1 and 2 are not supported. Strat. Mgmt. J., 30: 577–594 (2009) DOI: 10.1002/smj

588

E N-K. Lim, S. S. Das, and A. Das -0.03 (-0.04)

e1

Debt financing 95–96

0.49*** (0.46***) a=-0.01* (0.11*)

0.27*** (0.27***)

e2

R2= 0.24 (0.25) 0.73*** (0.74***) Debt financing c=0.00 97–98 (0.01)

d= 0.04 (0.04)

b=-0.09* (-0.09*) Unrelated strategy 95–96

R2=0.51 (0.51) Debt financing 99–00

Unrelated strategy 99–00

Unrelated strategy 97–98

0.73*** (0.73***)

0.77*** (0.76***) R2=0.50 (0.51)

R2=0.43 (0.43) e4

e3

-0.20* (-0.20*)

Figure 2. Longitudinal structural equation model showing standardized coefficient estimates for the constrained (and unconstrained) relationships in Model 3 (and Model 4), respectively, between unrelated strategy and debt financing Notes: Debt financing 95–96 refers to firm debt financing level averaged over two years during 1995–1996, Unrelated strategy 95–96 refers to unrelated strategy averaged over two years during 1995–1996, and so forth. ‘a’ and ‘b’ as well as ‘c’ and ‘d’ in bold are equality constraints in Model 3. Model 4 is the same as Model 3, except that no constraints are imposed on its parameters. Values for unconstrained relationships in Model 4 are in parentheses. ∗∗∗ p < 0.001, ∗∗ p < 0.01, ∗ p < 0.05, †p < 0.10

Table 2. Descriptive statistics and bivariate correlations Variables 1 2 3 4 5 6 7 8

Debt financing (Ln) Total assets (Ln) ROA (Ln) Industry type Average age (Ln) Related strategy Unrelated strategy Environment

Mean

SD

2.24 1.74 12.28 1.52 1.23 1.14 2.00 0.91 3.06 0.65 0.080 0.18 0.45 0.41 3092.92 1451.30

1

2

3

4

5

6

7

8

— 0.26∗ −0.22∗ 0.14∗ 0.16∗ 0.06 0.14∗ 0.00

— −0.14∗ 0.23∗ 0.21∗ 0.01 0.13∗ 0.02

— −0.11∗ −0.17∗ 0.04 −0.24∗ −0.02

— 0.21∗ 0.08∗ 0.15∗ 0.00

— −0.02 0.21∗ 0.00

— −0.06∗ 0.03

— 0.01

—

∗ (p < 0.05); n (max) = 1470 (245 firms multiplied by six years). Number of observations varies for each variable due to missing cases.

Hypothesis 3 predicts that the relationship between unrelated diversification and debt financing level is negatively moderated by stable environments, but positively moderated by dynamic environments. Consistent with Hypothesis 3, the Copyright  2009 John Wiley & Sons, Ltd.

statistically significant results reported in Table 3 indicate that, for unrelated strategy, interaction with stable environments negatively affects debt financing level, but interaction with dynamic environments positively affects debt financing level. Strat. Mgmt. J., 30: 577–594 (2009) DOI: 10.1002/smj

Diversification and Capital Structure in the Asian Financial Crisis

589

Table 3. Mixed-effects model for main effects of diversification strategy (related and unrelated) and its interactions with environmental conditions on debt financing levels Type of model

Main effects

Time Window

1995–2000

Parameter

Estimate

Total assets (ln) ROA (ln) Industry type Average age (ln) Unrelated strategy Related strategy Environment (stable) Environment (dynamic) Unrelated strategy × (stable) environment Related strategy × (stable) environment Unrelated strategy × (dynamic) environment Related strategy × (dynamic) environment Degrees of freedom

0.17∗∗ −0.25∗∗ 0.03 0.00 0.19 0.53

∗∗

Interactions 1995–1997

Standard error 0.05 0.06 0.14 0.19 0.19 0.42

Estimate

1998–2000

Standard error

0.08 −0.31∗∗ 0.17 0.10 1.13∗ 1.45 0.38∗

0.07 0.09 0.15 0.24 0.50 1.29 0.16

−0.47∗ −0.47

0.24 0.75

489

Estimate

Standard error

0.31∗∗ −0.27∗∗ −0.04 −0.09 −2.28∗ −0.32

0.08 0.10 0.18 0.24 0.99 2.51

−0.33∗

0.15

0.57∗∗ 0.18

0.21 0.55

203

273

p < 0.01, ∗ p < 0.05, †p < 0.10

Table 4.

Indices for longitudinal structural equation models

Model description Constrained debt financing-related strategy (Model 1) Unconstrained debt financing-related strategy (Model 2) Constrained debt financing-unrelated strategy (Model 3) Unconstrained debt financing-unrelated strategy (Model 4)

χ 2 (df)

CFI

NFI

RMSEA

χ 2 ( df)

5.14 (6) 3.30 (4) 15.40∗ (6) 11.10∗ (4)

1.00 1.00 0.98 0.99

0.99 0.99 0.97 0.98

0.00 0.00 0.08 0.09

— 1.84 (2) — 4.30 (2)

N = 245 firms. CFI = Comparative fit index; NFI = Normed fit index; RMSEA = Root mean square error of approximation. ∗∗ p < 0.01, ∗ p < 0.05, †p < 0.10

Hypothesis 4 states that the relationship between related diversification and debt financing level is positively moderated by stable environments, but negatively moderated by dynamic environments. For related strategy, Table 3 shows that its interactions with stable and dynamic environments have no significant effects on debt financing levels. Thus, Hypothesis 4 is not supported. Findings also indicate that the control variables—firm performance as proxied by log(ROA), and firm size as measured by log(total assets)— are, by and large, significant and have negative and positive coefficients, respectively. It suggests, in general, that firm profitability decreases debt financing level and firm size increases debt financing level. As shown in Table 4, there are no significant chi-square differences between Model 1 and Model Copyright  2009 John Wiley & Sons, Ltd.

2 and between Model 3 and Model 4. Hence, these findings suggest that debt financing level and (related and unrelated) diversification strategy are simultaneous and reciprocally interrelated. Our longitudinal structural equation models support both causal directions.

DISCUSSION Since the seminal work of Modigliani and Miller (1958), capital structure has been extensively studied in financial economics. However, most finance research on capital structure has emphasized shareholder wealth maximization (Simerly and Li, 2000), assuming perfect rationality, complete information, and efficient markets (Bromiley, 1990); firm behavior is not regarded as important. Most Strat. Mgmt. J., 30: 577–594 (2009) DOI: 10.1002/smj

590

E N-K. Lim, S. S. Das, and A. Das

researchers who study capital structure and strategy investigate the influence of firm capital structure on strategy, rather than the influence of strategy on capital structure. We complement these studies by examining the influence of diversification strategy on debt financing level (the other causal direction in a reciprocal relationship) under different environmental conditions. Our results do not show any significant main effects of unrelated and related diversification on debt financing level, suggesting that diversification strategy in the product market does not directly affect capital structure decisions, at least for the sample of Singapore firms during the study period. Although this contradicts some findings for U.S. firms, these results are consistent with Men´endezAlonso (2003), who found no significant effect of diversification strategy on firm capital structure in a sample of 480 Spanish manufacturing firms during 1991–1994, despite using different debt ratios and two proxies of firm diversification (a revenuebased Herfindahl index and an entropy measure). A possible explanation is that country-level differences may have more effect on the firm’s capital structure. Second, as Chen and Ho (2000) note, the average Singapore firm, in terms of book value of total assets, is about one-third the size of a U.S. firm. Smaller firms might experience a different relation between diversification and capital structure than their larger counterparts. This conjecture finds support from the work of Jordan, Lowe and Taylor (1999), which extends the product diversification and capital structure link to small- and medium-sized firms (SMEs) and shows that, contrary to the results found using large firms, corporate strategy has no effect on the capital structures of SMEs. With separate models for stable and dynamic environments, we observe significant interactions for unrelated diversification, suggesting that environmental conditions matter, at least for Singapore firms. We conclude that the same type of diversification strategy may have opposing impact on debt financing level under stable or dynamic environmental conditions. In particular, our results indicate that unrelated diversification negatively influences debt financing level under stable environmental conditions, which conforms to agencytheoretic logic. In stable environments, managers pursuing an unrelated diversification strategy are less likely to prefer debt financing because of repayment obligations and the increased scrutiny Copyright  2009 John Wiley & Sons, Ltd.

of debt holders. However, the interaction of a dynamic environment and unrelated diversification positively affects debt financing level. Managers appear to exploit environmental dynamism by raising more debt, perhaps because debt holders are less effective in dictating managerial behavior in a dynamic environment. Thus, consistent with contingency theory, we find that environment significantly moderates the relationship between diversification strategy and debt financing. Unrelated diversification in stable versus dynamic environments is associated with different levels of debt financing. While we find significant effects for unrelated diversification on debt financing level under varying environmental conditions, we did not find any interactive effects for related diversification strategy on debt financing level, regardless of environmental conditions. Perhaps related-diversified Singapore firms are able to raise equity capital more inexpensively than debt, and so prefer equity financing over debt financing. Our longitudinal structural models reveal the bidirectional relationship between capital structure and corporate strategy. Thus, prior findings showing the effect of debt financing on diversification (Kochhar, 1996; Kochhar and Hitt, 1998) can be reconciled with our findings, which revealed the influence of diversification on debt financing level. Synthesizing both sets of results demonstrates that diversification decisions affect capital structure choices, which in turn drive diversification strategy. We believe our findings offer several theoretical insights and practical implications. First, the strategic management paradigm as proposed by Schendel and Hofer (1979) has highlighted the importance of the external environment on firm-level decisions. Our study provides a comprehensive theoretical explanation for how the external environment affects decisions regarding capital structure. We suggest that standard agency-theoretic predictions might need to be contextualized to specific environmental conditions, especially in times of crisis. We further contribute to extant literature by building upon the few prior studies that relate capital structure to the environment. For instance, Simerly and Li, who suggest that ‘environmental dynamism moderates the relationship between leverage and performance’ (Simerly and Strat. Mgmt. J., 30: 577–594 (2009) DOI: 10.1002/smj

Diversification and Capital Structure in the Asian Financial Crisis Li, 2000 : 45), did not consider corporate strategy. We extend their study by introducing diversification strategy into the equation. By examining different diversification strategies under varying environmental conditions, we show that both strategy and environment matter in debt financing decisions. From a practical standpoint, our study points to the importance of considering different environmental conditions in managerial decisions about diversification and firm financing strategies. In particular, overlooking the environmental contingencies in product diversification decisions could lead to misleading conclusions about debt financing. Our bidirectional causal relationship between debt financing and diversification strategy poses an important methodological consideration. Since strategic and financing decisions go hand in hand, methods that handle endogeneity, such as twostage least squares (2SLS) or three-stage least squares (3SLS) regression (see Kennedy, 1987, and Zellner and Theil, 1962, for more details), are recommended when the two sets of decisions are studied together. In spite of these important contributions and insights, the present study has several limitations. First, while we examine the effect of diversification on debt financing level, we also recognize that there are bigger questions, such as the relationship between capital structure and strategy. Similarly, environmental dynamism is only part of a bigger issue that considers the influence of institutional effects on firm strategic behavior. Second, we did not investigate firm value or performance as a dependent variable. While we focused on capital structure, we found firm performance (as a control variable) to be negatively associated with firm leverage. This relationship can be examined in far greater depth, theoretically and empirically. Third, both left- and right-censoring in our data exist because data collection for Singapore firms for three years before and after the Asian Financial Crisis (1995–2000) was a very resource-intensive process. Fourth, we tested our hypotheses on a sample of firms from Singapore. While Singapore represents an Asian market affected by the crisis, our results generalize best to firms in countries similarly affected by an economic crisis. Fifth, although we did not control for some factors that might bear upon the diversification strategy in our sample, such as market growth, corporate culture, ownership structure, and Copyright  2009 John Wiley & Sons, Ltd.

591

incentives, we included control variables such as age, size, performance, and industry type in our analysis.

CONCLUSION We explored linkages between strategic management and finance, specifically investigating how different types of diversification relate to capital structure, as well as the moderating role of the environment on these relationships. Our results are consistent with agency theory under stable environments. However, under environmental dynamism, the negative relationship between unrelated diversification and debt financing is reversed, suggesting that standard agency-theoretic predictions about capital structure might need to be contextualized to account for environmental contingencies. In addition, our longitudinal structural equation models support the bidirectional causal relationship between diversification and debt financing. Our research thus emphasizes the importance of examining diversification and financing decisions concurrently to enhance our understanding of the complex relationship between corporate strategy and capital structure.

ACKNOWLEDGEMENTS We are grateful for helpful suggestions from Donald Bergh, Michael Levine, and Gayla Olbricht. We also express our gratitude to Editor Ed Zajac and two anonymous SMJ referees for their many insightful comments. We acknowledge research funding from the Nanyang Business School, Singapore.

REFERENCES Aldrich HE. 1979. Organizations and Environments. Prentice-Hall: Englewood Cliffs, NJ. Allayannis G, Brown GW, Klapper LF. 2003. Capital structure and financial risk: evidence from foreign debt use in East Asia. Journal of Finance 58(6): 2667–2710. Amihud Y, Lev B. 1981. Risk reduction as a managerial motive for conglomerate mergers. Bell Journal of Economics 12: 605–616. Bagiella E, Sloan RP, Heitjan DF. 2000. Mixed-effects models in psychophysiology. Psychophysiology 37: 13–20. Strat. Mgmt. J., 30: 577–594 (2009) DOI: 10.1002/smj

592

E N-K. Lim, S. S. Das, and A. Das

Balakrishnan S, Fox I. 1993. Asset specificity, firm heterogeneity and capital structure. Strategic Management Journal 14(1): 3–16. Barton SL, Gordon PJ. 1988. Corporate strategy and capital structure. Strategic Management Journal 9(6): 623–632. Baysinger BD, Hoskisson RE. 1989. Diversification strategy and R&D intensity in large multiproduct firms. Academy of Management Journal 32: 310–332. Beamish PW, da Costa RC. 1984. Factors affecting the comparative performance of multinational enterprises. Proceedings, European International Business Association Conference, Rotterdam, The Netherlands. Berger AN, di Patti EB. 2006. Capital structure and firm performance: a new approach to testing agency theory and an application to the banking industry. Journal of Banking and Finance 30: 1065–1102. Berger PG, Ofek E, Yermack DL. 1997. Managerial entrenchment and capital structure decisions. Journal of Finance 52(4): 1411–1438. Berle AA, Means GC. 1932. The Modern Corporation and Private Property. Macmillan: New York. Booth L, Aivazian V, Demirguc-Kunt A, Maksimovic V. 2001. Capital structures in developing countries. Journal of Finance 56: 87–120. Braun M, Larrain B. 2005. Finance and the business cycle: international, inter-industry evidence. Journal of Finance 60(3): 1097–1128. Bromiley P. 1990. On the use of finance theory in strategic management. In Advances in Strategic Management, Vol. 6, Lamb R, Shrivastava P (eds). JAI Press: Greenwich, CT; 71–98. Chandler AD Jr. 1962. Strategy and Structure: Chapters in the History of the American Industrial Enterprise. MIT Press: Cambridge, MA. Chatterjee S, Wernerfelt B. 1991. The link between resources and type of diversification: theory and evidence. Strategic Management Journal 12(1): 33–48. Chen SS, Ho KW. 2000. Corporate diversification, ownership structure, and firm value: the Singapore evidence. International Review of Financial Analysis 9(3): 315–326. Chia SY. 2000. Singapore’s responses to the Asian Crisis and the challenges of globalisation. In Restoring East Asia’s Dynamism, Masuyama S, Vandenbrink D, Chia SY (eds). Nomura Research Institute and Institute of Southeast Asian Studies: Tokyo and Singapore; 163–193. Child J. 1972. Organization structure, environment and performance: the role of strategic choice. Sociology 6: 1–22. Coase RH. 1937. The nature of the firm. Economica 4: 386–405. Crutchely CE, Jensen MRH. 1996. Changes in corporate debt policy: information asymmetry and agency factors. Managerial Finance 22(2): 1–15. Cyert RM, March JG. 1963. A Behavioral Theory of the Firm. Prentice-Hall: Englewood Cliffs, NJ. Dean TJ, Meyer GD. 1996. Industry environments and new venture formation in U.S. manufacturing: A conceptual and empirical analysis of demand Copyright  2009 John Wiley & Sons, Ltd.

determinants. Journal of Business Venturing 11(2): 107–132. Dempster AP, Selwyn MS, Patel CP, Roth AJ. 1984. Statistical and computational aspects of mixed model analysis. Applied Statistics 33: 203–214. Dess GG, Beard DW. 1984. Dimensions of organizational task environments. Administrative Science Quarterly 29: 52–73. Duncan RB. 1972. Characteristics of organizational environments and perceived environmental uncertainty. Administrative Science Quarterly 17: 313–327. Easterbrook FH. 1984. Two agency-cost explanations of dividends. American Economic Review 74: 220–230. Fattouh B, Scaramozzino P, Harris L. 2005. Capital structure in South Korea: a quantile regression approach. Journal of Development Economics 76: 231–250. Galbraith J. 1973. Designing Complex Organizations. Addison-Wesley: Reading, MA. Geringer JM, Tallman S, Olsen DM. 2000. Product and international diversification among Japanese multinational firms. Strategic Management Journal 21(1): 51–80. Haleblian J, Finkelstein S. 1993. Top management team size, CEO dominance, and firm performance : the moderating roles of environmental turbulence and discretion. Academy of Management Journal 36(4): 844–863. Hambrick DC, Finkelstein S. 1995. The effects of ownership structure on conditions at the top: the case of CEO pay raises. Strategic Management Journal 16(3): 175–193. Harris FH. 1994. Asset specificity, capital intensity and capital structure: an empirical test. Managerial and Decision Economics 15: 563–576. Harvey CR, Roper AH. 1999. The Asian bet. In The Crisis in Emerging Financial Markets, Harwood A, Litan RE, Pomerleano M (eds). Brookings Institution Press: Washington, D.C.; 29–115. Available at; http://www.duke.edu/∼charvey/curvit.htm. Hitt MA, Gimeno J, Hoskisson RE. 1998. Current and future research methods in strategic management. Organizational Research Methods 1: 6–44. Hitt MA, Hoskisson RW, Kim H. 1997. International diversification: effects on innovation and firm performance in product diversified firms. Academy of Management Journal 40(4): 767–798. Ho TW. 2004. The foreign exchange exposure of capital structure: the 1997 Asian crises revisited. Applied Financial Economics 14: 497–505. Hoskisson RE, Hitt MA, Johnson RA, Moesel DD. 1993. Construct validity on an objective (entropy) categorical measure of diversification strategy. Strategic Management Journal 14(3): 215–235. International Monetary Fund. 2008. World economic outlook database, April 2008: nominal GDP list of countries, data for the year 2007. http://www.imf.org/ external/pubs/ft/weo/2008/01/weodata/index.aspx. Accessed on 13 September 2008. Jahera JS, Lloyd WL. 1996. An empirical assessment of factors affecting corporate debt levels. Managerial Finance 22(2): 29–38. Strat. Mgmt. J., 30: 577–594 (2009) DOI: 10.1002/smj

Diversification and Capital Structure in the Asian Financial Crisis Jensen MC. 1986. Agency cost of free cash flow, corporate finance and takeovers. American Economic Review 76(2): 323–329. Jensen MC, Meckling W. 1976. Theory of the firm: managerial behavior, agency costs, and ownership structure. Journal of Financial Economics 3: 305–360. Jensen MC, Smith CW Jr. 1985. Stockholder, manager and creditor interests: applications of agency theory. In Recent Advances in Corporate Finance, Altman E, Subrahmanyam M (eds). Irwin: Homewood, IL; 93–131. Jordan J, Lowe J, Taylor P. 1999. Strategy and financial policy in UK small firms. Journal of Business Finance and Accounting 25: 1–27. Jurkovich R. 1974. A core typology of organizational environments. Administrative Science Quarterly 19: 380–394. Keats BW. 1990. Diversification and business economic performance revisited: issues of measurement and causality. Journal of Management 16: 61–72. Keats BW, Hitt MA. 1988. A causal model of linkages among environmental dimensions, macro organizational characteristics, and performance. Academy of Management Journal 31(3): 570–598. Kennedy P. 1987. A Guide to Econometrics (2nd edn). MIT Press: Cambridge, MA. Kochhar R. 1996. Explaining firm capital structure: the role of agency theory vs. transaction cost economics. Strategic Management Journal 17(9): 713–728. Kochhar R, Hitt MA. 1998. Linking corporate strategy to capital structure: diversification strategy, type and source of financing. Strategic Management Journal 19(6): 601–610. Laird MN, Ware JH. 1982. Random-effects models for longitudinal data. Biometrics 38: 963–974. Lawrence PR, Lorsch JW. 1969. Organization and Environment. Irwin: Homewood, IL. Low PY, Chen KH. 2004. Diversification and capital structure: some international evidence. Review of Quantitative Finance and Accounting 23: 55–71. Lowe JA, Naughton A, Taylor P. 1994. The impact of corporate strategy on the capital structure of Australian companies. Managerial and Decision Economics 15: 245–257. Lu JW, Beamish PW. 2001. The internationalization and performance of SMEs. Strategic Management Journal , June-July Special Isue 22: 565–586. Men´endez-Alonso EJ. 2003. Does diversification strategy matter in explaining capital structure? Some evidence from Spain. Applied Financial Economics 13: 427–430. Miles RE, Snow CC, Pfeffer J. 1974. Organizationenvironment: concepts and issues. Industrial Relations 13: 244–264. Miller KD, Leiblein MJ. 1996. Corporate risk-return relations: returns variability versus downside risk. Academy of Management Journal 39: 91–122. Milliken FJ. 1987. Three types of perceived uncertainty about the environment: state, effect, and response uncertainty. Academy of Management Review 12: 133–143. Copyright  2009 John Wiley & Sons, Ltd.

593

Modigliani F, Miller MH. 1958. The cost of capital, corporate finance and the theory of investment. American Economic Review 48: 261–297. Montgomery CA, Wernerfelt B. 1988. Diversification, Ricardian rents, and Tobin’s q. RAND Journal of Economics 19(4): 623–632. Morellec E. 2004. Can managerial discretion explain observed leverage ratios? Review of Financial Studies 17(1): 257–294. Ngiam KJ. 2000. Coping with the Asian financial crisis: the Singapore experience. Visiting Researchers Series No. 8, Institute of Southeast Asian Studies, Singapore. Available at: http://www.iseas.edu.sg/vr82000.pdf (Accessed 1 June, 2006). Palepu K. 1985. Diversification strategy, profit performance and the entropy measure. Strategic Management Journal 6(3): 239–255. Penrose E. 1959. The Theory of the Growth of the Firm. Basil Blackwell: London, UK. Pomerleano M. 1998. The East Asia crisis and corporate finances: the untold microeconomic story. Emerging Markets Quarterly 2(4): 14–28. Preece SB, Miles G, Baetz M. 1998. Explaining the international intensity and global diversity of earlystage technology-based firms. Journal of Business Venturing 14(3): 259–282. Ramanujam V, Varadarajan P. 1989. Research on corporate diversification: a synthesis. Strategic Management Journal 10(6): 523–551. Rozeff M. 1982. Growth, beta, and agency costs as determinants of dividend payout ratios. The Journal of Financial Research 5: 249–259. Rumelt RP. 1974. Strategy, Structure and Economic Performance. Division of Research, Harvard Business School: Boston, MA. Rumelt RP. 1982. Diversification strategy and profitability. Strategic Management Journal 3(4): 359–369. Rumelt RP, Schendel D, Teece DJ. 1994. Fundamental Issues in Strategy: A Research Agenda. Harvard Business School Press: Boston, MA. SAS Institute Inc. 1992. SAS/STAT Software: Changes and Enhancements, Release 6.07 . SAS Technical Report P-229. SAS Institute, Inc.: Cary, NC. Schendel DE, Hofer CW. 1979. Strategic Management: A New View of Business Policy and Planning. Little, Brown and Company: Boston, MA. Shook CL, Ketchen DJ Jr, Cycyota CS, Crockett D. 2003. Data analytic trends and training in strategic management. Strategic Management Journal 24(12): 1231–1237. Simerly RL, Li M. 2000. Environmental dynamism, capital structure and performance: a theoretical integration and an empirical test. Strategic Management Journal 21(1): 31–49. Singh K, Yip GS. 2000. Strategic lessons from the Asian crisis. Long Range Planning 33: 706–729. Singh M, Davidson WN, Suchard J. 2003. Corporate diversification strategies and capital structure. Quarterly Review of Economics and Finance 43: 147–167. Stein JL, Lim GC. 2004. Asian crises: Theory, evidence, warning-signals. Working paper no. 1159, CESIFO Strat. Mgmt. J., 30: 577–594 (2009) DOI: 10.1002/smj

594

E N-K. Lim, S. S. Das, and A. Das

GmbH, Munich, Germany. Available at SSRN: http://ssrn.com/abstract=528185. Tallman S, Li J. 1996. Effects of international diversity and product diversity on the performance of multinational firms. Academy of Management Journal 39(1): 179–196. Taylor P, Lowe J. 1995. A note on corporate strategy and capital structure. Strategic Management Journal 16(5): 411–414. Thompson JD. 1967. Organizations in Action. McGrawHill: New York. Tosi HL Jr, Gomez-Mejia LR. 1989. The decoupling of CEO pay and performance: an agency theory perspective. Administrative Science Quarterly 34(2): 169–189. Venkatraman N. 1989. The concept of fit in strategy research: toward verbal and statistical correspondence. Academy of Management Review 14: 423–444.

Copyright  2009 John Wiley & Sons, Ltd.

Vicente-Lorente JD. 2001. Specificity and opacity as resource-based determinants of capital structure: evidence for Spanish manufacturing firms. Strategic Management Journal 22(2): 157–177. Williamson OE. 1985. The Economic Institutions of Capitalism: Firms, Markets, Relational Contracting. Free Press: New York. World Bank. 2008. Gross domestic product data for the year 2007. World Development Indicators database, World Bank, Washington, D.C. Available at: http://siteresources.worldbank.org/DATA STATISTICS/Resources/GDP.pdf (Accessed 13 September 2008). Zellner A, Theil H. 1962. Three-stage least squares: simultaneous estimation of simultaneous equations. Econometrica 30(1): 54–78.

Strat. Mgmt. J., 30: 577–594 (2009) DOI: 10.1002/smj