Opening up to global trade and investment is often thought to trigger institutional improvement by raising the expected benefits of institutional reform and reducing incumbents' incentives and ability to preserve the status quo. However, recent experience is not entirely consistent with this conventional wisdom. We suggest an explanation based on variation across countries in firms' reliance on ambient institutions. Large, well-established firms depend less on an economy's institutions than do small and incipient firms. Multinational firms likewise can use their global organizations to sidestep weak local institutions. Firm heterogeneity of this sort can thus contribute to markedly different institutional responses to liberalization—institutional development is better in locations where firms and potential entrants benefit more from such development. Our framework also suggests that institutional development might occur in stages. In an economy whose basic institutions are sound, individuals rationally invest in entrepreneurial capability and firms rationally invest less in institution substitutes. Economies with firms that rely more on ambient institutions or with more potential entrants who would rely on those institutions are more likely to experience further institutional improvement following accession to the global economy. Economies with fewer firms or potential entrants dependent on sound institutions, in acceding to the global economy, may exhibit scant institutional improvement, and perhaps even institutional deterioration. Political rent-seeking is not necessary for the latter outcome, but expands the range of conditions under which it ensues.