DIVIDEND, INVESTMENT AND LEVERAGE POLICY BASED ON AGENCY THEORY IN INDONESIA MANUFACTURING COMPANIES

The research is aimed to explore Dividend, Investment & leverage policy based on agency thoeory at manufacturing companies in Indonesia. Having the Agency theory approach such as irving fisher theory and peking order theory using Lisrel software support for 195 manufacturing companies that listed on Indonesia stock exchange of 2014 till 2018 period as samples. The result shown that Agency cost variable has no significant positive impact towards leverage or on the other words that it has negative impact towards leverage level of the company. Bankcruptcy has a negative effect on the company's leverage level. Investment however has positive effect on the leverage level. Whereas Dividend has a negative impact on company's investment. Dividend gave a positive effect on leverage level and it capable to mediate the investment's effect towards agency cost and leverage level of the enterprise. The outcomes of this trial would bring consequences of the company which has high growth level that requires bigger funds for the investment implementation. Based on pecking order theory, stated that funds requirements should be prioritized on internal fund sources such as retained earnings. Since most of the funds are alocated for the requirement of investment funds, therefore the paid dividend would then reduced. This descriptive proof supported the pecking order theory as this outcomes gave countenance to Fama and French that by using pecking order theory can be simply drawn that debts generally would increased if the investment exceeded the retained earnings, and it would decreased when investment is lower than retained earnings. This study discusses policies, Dividends, Investment and leverage based on theories such as the Irving Fisher theory and the Peking order theory and research conducted in manufacturing companies in Indonesia which still very rarely reveals this topic and also provides an explanation of this problem using the SEM model.