Capital structure of a business refers to its sources of finance. The value of a company is independent of its capital structure and the cost of equity for a leveraged firm is equal to the cost of equity for an unleveraged firm (Miller and Modigliani, 1961). All this prevails if there is no transaction or bankruptcy cost and there is perfect information. The pecking order theory of Myers (1984) ranks internal sources such as debt and equity which paves the way for expansion of the concept of capital structure. Many of the previous studies reveal rising debts to have positive effect on the growth of economic activities of the business. On account of rising growth the business is likely to be overvalued; the benefits of which are accrued to the management. When the management is after these benefits, this creates agency costs. Agency costs provide relevance of capital structure. Unless free cash flow is given back to investors, management has an incentive to destroy firm value through empire building and perks. Increasing leverage imposes financial discipline on management. The studies of Black (1976), Allen and Michaely (2003) and Baker (1999) provide room and gap to explore link of agency cost, capital structure and dividends. This study is an academic contribution to fill the gap. This study investigates significance of agency costs in relationship of capital structure and dividend policy of Pakistani nonfinancial sector. Type and size of the industries, performance of business enterprises possess significance in capital structure. Industry-specific qualitative variables, market capitalization, performance indicators, risk factors are also included in the model. Finding interactive effects of the agency costs, capital structure and dividend policies are an important contribution of the present study. The study finds significance of agency costs which determines dividend policy at variable proportion. Certain leading theories helped in providing the scope of this study. ‘Dividend Puzzle’ of Black (1976); growth prospects of the industry, earnings of the firm, and dividend policy of Lintner (1956); capital structure irrelevance of Miller and Modigliani (1961); dividend policy perspective of Brealey and Myers (1991) & Brigham and Gapenski (1991); significance of agency costs for the determination of dividends (Lloyd, et al., 1985; Alli, Khan, & Ramirez, 1993; Dempsey, Laber, & Rozeff, 1993) are a few to mention. Objectives of the study include understanding of the environment and features of Pakistani business enterprises; dividend policies of the listed companies of nonfinancial sector of Pakistan; understanding significance of agency cost using nonlinear and interactive approach; bringing together various advanced econometric methods as measuring source of relationship of variables; exploring empirical effects of three different sizes and nine different types of industries on the dividend policy; investigating homogeneity/heterogeneity of management style of the Pakistani business enterprises. The present study employs qualitative response variable models and the panel data models using the variable modeling technique consulting Amemiya (1981); Daniel & Yu, (2000); Gujarati & Sangeetha (2007) and Greene (2005). Findings of the study reveal interesting results. Pakistani management is relatively efficient in meeting regulatory requirements as compared to other countries of the South Asian region. The highest degree of bureaucratic culture in tax department is observed in Pakistan. The highest degree of problem is in the sectors related to the utilities which make very important component of infrastructure. Descriptive statistics reveal inconsistent overall performance of nonfinancial sector of Pakistan and that 2004, 2005 and 2006 are the years of better performance of Pakistani business enterprises. However, there is no significant difference of the level of uncertainty across the Pakistani industries. The results also show increasing probability of paying dividend with market capitalization, sales growth and ROA of companies; increasing impact of debt-equity ratio xiv(agency cost) on the probability of paying dividend and that leverage compels managers to act more in the interests of shareholders. Liquidity has the least relevance with the dividend policy of the Pakistani nonfinancial sector. There are significant chances of receiving dividends from engineering, paper, transport and communication industries. There are fewer chances of receiving dividends from cement, chemical, fuel, and energy. There are significant chances of receiving dividends from sugar and textile industries, though less than that of miscellaneous group. Findings of the estimated models show significance of industry-specific qualitative variables in the determination of dividend policy and performance of the business enterprises. An agency cost significantly raises the probability of paying dividends. Significance of the risk factor is not so high in the determination of dividend policy and performance of the corporate sector of Pakistan. Performance (ROA; sales growth) raises the probability of paying dividends. Medium and large companies look more inclined towards payment of dividends. Structural interaction of size and the leverage has no significance in this study. Negative interactive effect of the agency cost on the dividend policy of cement, sugar and textile show overvaluation as per Pecking Order theory. Identical findings are observed for ROA and ROE in respect of dividend policy. However, risk-related ROE has statistically insignificant role in dividend policy. Significant performance of Cement, Engineering, Sugar, Textile, Fuel & Energy, Paper & Board, Transport & Communication, in the Pakistani NFS has been found in the this study. The better business operations in the form of cash flows may be considered as source of better performance in terms of their profit efficiency. The performance of small firms in the industry is better in terms of profit margins as compared to medium and large firms on account of management efficiency. Finding interactive effects of the agency costs, capital structure and dividend policies are an important contribution of the present study. The study finds significance of agency costs which determines dividend policy at variable proportion. Application of nonlinear methods is another contribution of the study. The management must be careful while deciding about leverage. This study also considers dividend policy instead of dividend payout which remains controversial in Pakistani context. Leverage as control variable has potential to resolve ‘dividend puzzle’. Variable Proportional relationship of agency cost and dividend policy is an academic contribution of the study. The study also provides guidance for the investors as to the choice of industries for the purpose of investment. The study recommends application of Almon Distributed lag models if quarterly data for the Pakistani businesses were available. Dynamic Panel data models could be another option for the further studies. Behavioral financial analysis through survey data can be another option of future research. Testing of the assumptions of market perfection in the framework of M-M model is desired. The future research can be extended to South Asia, as a comparative study. ‘Dividend Puzzle’ can be resolved by optimal utilization of leverage because there is variable proportional relationship of leverage and dividend policy. ‘Agency Cost’ does have significance through interactive relationship of other capital structure, performance, size and types of industries. The study augments to the worth mentioning contribution of Michel (1979); Baker et al. (1985); Dempsey et al. (1993); Pandey (2003); Black (1976); Baker (1999) who missed the interactive effects.