Generating intellectual property in the form of patents enriches both inventor and the general public. New drugs, faster computer chips, more productive crops, and safer cars can result from a bright idea or years of sustained, methodical research, brought to market with the expectation of profit.

Looking at the 50 states and the District of Columbia over the past ten years of patent grants from 2005 to 2014 shows that the top three states in terms of per capita patents granted are Vermont, Idaho and Massachusetts. California, with its dynamic Silicon Valley, came in fourth in the per capita state rankings. Alaska brings up the rear in the rankings with Vermonters generating almost 18 times the patents per capita as Alaskans.

Reasons for the wide variation in patent generation at the state level are hard to ascertain. Individual companies can account for a huge portion of the patents granted. For instance, Idaho benefits by being the home to Micron Technology MU +0.00%, a semiconductor device design and manufacturing firm that often accounts for 98 percent or more of the patents granted in the state of 1.6 million people known more for the raw ingredient in potato chips than for its microchips.

Among 106 nations, Taiwan leads the way in patent production per capita over the past decade, followed closely by Japan and the United States. Israel is in fourth place, but it’s a distant fourth. Of the 106 nations tracked, ten had no patents filed at all in the past ten years, eight of which were in Africa and two in Asia.

A multivariate regression was run testing international per capita patent generation in 106 nations against eight variables, five of them ratings from the Fraser Institute’sEconomic Freedom of the World index, including: Legal System & Property Rights; Size of Government; Sound Money; Freedom to Trade Internationally; Regulation; along with Purchasing Power Parity; Transparency International’s Corruption Perceptions Index; and the Adult Literacy rate. Of these factors, the Corruption Perceptions Index was the most predictive, with greater levels of corruption equating to less invention. The next most important factor was the per capita purchasing power parity of a nation. Surprisingly, the incidence of adult literacy had no predictive value at all.

As would be expected, the differences in per capita patent production is far greater between nations than between the states in the U.S. Breaking the analysis into upper and lower tiers, with the top 50 nations with robust research universities in the top tier and other nations in a lower tier adds further detail to the analysis.

For developing nations, the sole factor with predictive value for patent production was the Corruption Perceptions Index. Per capita wealth approached significance statistically. This reinforces the notion that rule of law and freedom from corruption is a vital first step in lifting nations out of poverty.

For the top 50 nations with a robust higher education system, none of the eight variables were predictive in determining per capita patent productivity. Only when adding a ninth variable, the Universitas 21 rankings of national higher education systems, can patent generation be predicted with some reliability—though the model still has significant uncertainty with an adjusted R square of only 0.2899. This suggests that other factors may be significant, such as culture, access to capital, and comparative advantage. Economic comparative advantage may explain why nations that are relatively poor in natural resources dominate the top ten: 1st in per capita patents Taiwan; 2nd place Japan; 4th place Israel; 6th place Switzerland; and 7th place Finland.