There  is a long explanation over at medium

committing code is a small part of what software engineers do. There are many ways in which you can contribute technically without writing much code. In fact, the best senior engineers I know are great at prioritization, delegation, and simplifying. Their pull requests are tiny because they’ve made the code modular and straightforward. … Writing negative code is a point of pride! If you refactored a complex piece of logic, picked a good library to solve a problem, or eliminated unnecessary features, then you’ll have a negative commit history. ..  fine-tune configs, research arcane issues, and dive deep into underlying architecture. Working on these jobs can mean days of research just to make a few-line change.

The same applies to amount of funding received by a scientist (although frequently used for calculation of salariy). If I can achieve the same impact at a fraction of the costs, isn’t that much better than producing more costs?