There is a large literature – theoretical and empirical – on how patents incentivize (or not, in the case of “downstream” innovations) new inventions. Less well-known is the effect of patent law on distorting new invention choice. Moser uses a dataset from World’s Fairs in 1851 and 1876, which provide a useful natural experiment, to examine this issue. In the 19th century, there were massive differences in patent length and protection, ranging from no patents in the Netherlands after 1869 to fairly broad protection in the US in both years. In the absence of patents, inventors can recoup the fixed cost of innovation by keeping a method secret (as in the Coca-Cola recipe) or by standard first-mover advantage. The World’s Fair records clearly show that countries with weaker patent protection saw relatively more inventions in classes of goods which are difficult to reverse engineer – i.e., goods whose production method can easily be kept secret.