This dissertation investigates which functions are efficiently combined in one institution. Part I investigates the conflicts of interest ascribed to universal banking, bank equity stakes in issuing firms, and underwriter affiliation with venture capitalists, finding a positive relationship between universal-bank structure and initial IPO1 returns (underpricing). In the secondary markets, however, universal-bank underwritten and specialized-bank underwritten stocks are indistinguishable, suggesting that underpricing compensates for potential conflicts of interest. The paper also finds that pre-existing bank relationships rather than issuer characteristics determine the underwriter choice.|Part II studies investor valuation of U.S. conglomerates throughout a period of three years at the end of the 1960s. Recent research finds that conglomerates had greater market-to-book ratios than combinations of comparable single-segment firms during 1966-1968 and that diversifying acquisitions generally earned positive abnormal returns in the 1960s. During the 1970s and 1980s, however, conglomerate performance declined sharply. Previous explanations of the conglomerate merger wave fail to account for the conglomerates' initial popularity. This paper argues that investors assign value to corporate structure as such, having systematically overvalued the conglomerate corporate structure during the 1960s and then systematically updated their evaluation. The conglomerates' initial popularity and later decline can be seen as evidence of the systematic struggle to determine the value of corporate structures. I find some evidence of such structural effects that are not explainable by a capital asset pricing model. Firms with the conglomerate structure are clearly related to each other.|Although conflicts of interest can arise and the conglomerate structure appears to have been largely inefficient, neither observation calls for government regulations. Investors are clearly aware of potential conflicts involved with universal banking and require a risk premium as compensation. The study of investor valuation of conglomerates reveals that investors assign value to corporate organization, but that the value of any given structure is hidden and needs to be learned. Only an unrestricted market mechanism can provide the information needed to infer the best allocation of resources.