How should ownership rights be allocated in public-good settings? We report data from a laboratory experiment with 480 participants that was designed to test a public-good version of the property rights theory based on incomplete contracting. Consider two parties, one of whom can invest in the provision of a public good. The parties value the public good differently. In the literature it has been argued that more investments will be made if the high-valuation party is the owner, regardless of whether or not this party is the investor. Our experimental results cast some doubts on the robustness of this conclusion.