This paper develops new data on mega-regions. It takes issue with theories of activity is being more widely decentralized—‘ globalization which argue that global economicthe world is flat’. We use a global dataset of night-time light emissions to produce an objectively consistent set of mega-regions for the data to estimate globe. We draw on high-resolution populationthe population of each of these regions. We then process the light data in combination with national gross domestic product (GDP) to produce rough but useful regional estimates of economic activity. We also present estimates of technological and scientific innovation.