Summary

An expression is in tail context if its value will be returned immediately following its evaluation.
A tail call is a function or method call that appears in a tail context.
Whether tail calls are faster and more space-efficient than non-tail calls depends on the implementation of your programming language.
If the implementation is properly tail recursive, tail calls are more space-efficient than non-tail calls, and will probably be faster as well.
If the implementation is not properly tail recursive, but performs tail call optimization, a tail call might be more efficient than a non-tail call, but you can't rely on that.
Some programming languages, notably Scheme, require implementations to be properly tail recursive.
In most implementations of most programming languages, tail calls are implemented so inefficiently that tail calls are just as slow and consume just as much space as non-tail calls.
The inefficiency of those implementations is unlikely to be fixed in the near future.
In languages such as Java or C, whose implementations are not properly tail recursive and may not even perform tail call optimization, you'll need to rewrite performance-sensitive tail calls using special loop syntax.
Rewriting tail calls to use special syntax for loops is not always possible, because the special syntax is usually confined to a single function or method body. With tail recursion, loops can cross function/method boundaries.