Are you writing a text extractor software which analyses the content streams?
Or using Acrobat to extract text?
Or using different software to extract text?
The effect of the -1, like the other elements, is to change how Trm is generated, and how the text spacing is calculated. It doesn't have a specific effect by itself. It will be being combined with the CTM, set by cm, which you didn't mention; my guess is that cm will also have a negative fourth component.
In all cases it's important to realise that spaces are inserted by guesswork and fuzzy logic and are often wrong. They may also depend on other factors including average text spacing elsewhere on the page.