Digital economy has free lunch

Introduction

Unlike oil, data is non-rival. That means data can be used multiple times without inherently diminishing its value. Data can be duplicated and shared at a relatively low cost, so the production of data does not have to be balanced with the uses of that data.  

Source Mandel M. (2017), “The Economic Impact of Data: Why Data is not like Oil”, Progressive Policy Institute, Washington, DC, https://www.progressivepolicy.org/wp-content/uploads/2017/07/PowerofData-Report_2017.pdf.

Let’s go in detail 

The same data source can be easily applied to multiple questions, resulting in a brought variety of responses and uses. This also implies that the transition from empiric to an abstracted model can be performed in a cheap manner.  

Mathematical view:

Based on models the required data fusion, data analysis, and business expertise can be easy, on labeled data sets also automatized, applied for validation. 

Li, W C Y and B H Hall (2018), “Depreciation of Business R&D Capital”, Review of Income and Wealth, https://voxeu.org/article/no-such-thing-free-lunch-digital-economy 

Examples:

Li, W C Y and B H Hall (2018), “Depreciation of Business R&D Capital”, Review of Income and Wealth, Excerpt from https://voxeu.org/article/no-such-thing-free-lunch-digital-economy 

Why almost free lunch

The “no free lunch theorem” (NFL) states that an algorithm – resulting in a model – outperforms in one case, but fail on another case: 

• Model are simplifications of reality
• Simplifications imply assumption bias
• Assumptions fail in unknown situations

In respect to our purpose, it reflects that data fusion and data analytics requires adaptations between different applications. These required adaptations also cause effort – but the ratio compared to the gained benefit is low. Oil can only be burned once, data multiple times – multiple data-driven engines can be run in parallel with the same data “oil”. 

Almost like free lunch for the data-driven business models.