Ask HN: How to justify time spent on updating dependencies to managers?

Can be time-consuming and benefits are not as apparent as adding new features. For example going from Rails 3 to Rails 5 is a huge deal. Especially if you have 500 lines long Gemfile (list of dependencies) where all of the gems also have their own dependencies.

Without thinking much I can say:

1. It is easier to work with relatively fresh versions of modules. Blog posts are being written about them. Generally frameworks and libraries get easier and easier to work with. New features are being added making previously difficult things easy.

2. Fewer bugs and security vulnerabilities. Bugs discovered by you are more likely to get fixed upstream. If you use old version and find a bug that is present everywhere you have to fix it yourself for the unsupported version.

3. Developers from the market would consider working with legacy versions a disadvantage and less likely to join the company.

4. Better performance.

Reasons to say no:

1. Time spent on updating can be spent on features which are easier to connect to revenue, thus making it more lucrative for people who'll get the credit. Career advancement. And often features are actually more important than everything else including security, performance, code quality, and things like that.

2. You will introduce regressions.

3. Some applications are working well enough anyway. Your update can mess them up.

4. Expensive.

5. Most applications and programs don't need to be technically excellent. They can be mediocre with strong brand.

How would you justify such updates? How do you make sure that you're not putting your own desire for freshness above company's true goals of earning money?