In the many applications of Markowitz' theory, it is assumed that an investor can choose from a wide range of investments, but has no control over either the performance or the variability of the individual components (stocks, bonds, real estate, etc.). Planned investment is based on a study of risk compared to level of performance of particular components, such that the financial consequences of adverse conditions are controlled. In the security field, performance refers to the level of safety achieved for a community. Since absolute safety cannot be guaranteed, security selection is based on limiting the consequences of catastrophic but rare events and more likely but moderately disruptive events. Using Markowitz terminology, there is a protection "portfolio" (security policies and measures) that offer the greatest protection for the cost. This optimal protection does not eliminate damage, but rather minimizes and controls harm to an approximate level. This paper explains how Markowitz' portfolio management theory may be adapted to the situation of a small region that faces periodic and irregular disruptions in local and export markets. It uses a computer simulation model to show how the potential for major disruptions varies with economic policy, increasing rapidly with the risk-propensity of policymakers or policies that increase the likelihood of a set of connected events. The simulation shows how security strategies that reduce ambiguity in consequences or protect critical elements in evolving disasters provide more effective public-safety policy. 6 figures and 18 references