For decades, business press and researchers have observed and investigated the premiums/discounts on closed end investment companies. Proposed explanations for the phenomenon have been poor performance, high expenses (due to agency relationships), inefficient internal capital markets and excess volatility in the returns of the shares of the closed-end investment companies. Some, but not conclusive, empirical evidence support these theories. Most empirical evidence is based on American data. This study uses British and Swedish data on closed end investment companies. Some, but not conclusive, empirical evidence support these theories. Most empirical evidence is based on American data. This study uses British and Swedish data on closed end investment companies from 1972 – 2004 to investigate the premiums/discounts. Three areas of explanations are examined: performance, agency costs and diversification. In contrast to previous studies this study uses detailed data on quoted and unquoted securities respectively to investigate the relationship between performance and premiums/discounts. Evidence is found for a relationship between the performance on unquoted securities and premiums/discounts, but not for quoted securities. Indications that measurement biases in unquoted securities are properly priced are also found. The agency problem is analyzed in two ways, formal and controlling power, to investigate if actions taken by the company substantiating agency behavior have additional effects on prices. Such actions are measured as large investments in other portfolio companies (controlling power).  The empirical evidence suggests that the existence of formal power creates additional discounts. The marginal effect on discounts is even deeper when proposed agency actions are identified. Diversification is argued to decrease the value of a portfolio of securities when heterogeneous beliefs are present. This study provides evidence that portfolio diversification deepens discounts. / Diss. Stockholm : Handelshögskolan, 2007