On most joint papers the corresponding author is whoever created the submission account. It looks like an administrative detail, and in the mechanics of submission it mostly is.
In evaluation systems it is not. In several countries — Vietnam among them — corresponding authorship carries specific weight in promotion files, grant scoring and institutional reporting. A choice made for convenience in one lab becomes part of a colleague's career record in another.
What the role actually involves
Four things, and only the first is administrative:
Handling submission and correspondence with the editorial office — cover letter, files, revisions, proofs.
Confirming that all authors have approved the submitted version. This is a real responsibility and journals increasingly ask for explicit confirmation.
Being the contact point after publication — for readers, for data requests, and for anything that goes wrong.
Being the person the journal writes to if a correction or retraction becomes necessary, sometimes years later.
That last point is worth stating plainly at the outset. The role does not end at publication.
Why it is counted differently in different places
Many evaluation systems distinguish first author and corresponding author from the rest, and some count them at a substantially higher weight. Where that is the case, the position is not a courtesy — it is a measurable credit.
Systems also differ on whether a paper can have more than one corresponding author, and whether all of them count. Some journals allow two or three; some evaluation frameworks recognise only the first listed.
The practical consequence for a collaboration: do not assume the role means the same thing on both sides. Ask your partner directly how their institution counts it. This is a short conversation and it prevents a long one.
How to decide
Four criteria, roughly in order of usefulness:
Who will actually do the work of corresponding? Handling revisions across a time zone gap while teaching a full load is real labour. Assigning the role to someone who cannot do it slows the paper down.
Who needs it more? An early-career researcher whose promotion file depends on corresponding-author counts benefits far more than a senior professor with two hundred papers. This is the criterion most worth applying deliberately.
Who is best placed to answer questions after publication? Usually whoever holds the data.
Whose institution will host the data or materials long-term.
Where the journal permits co-corresponding authors and both sides need the credit, that is often the clean answer — but check first that both institutions recognise it, because in some systems only the first listed counts and the arrangement then helps only one side.
Agree these five things before submission
Not after acceptance, when the positions are already fixed in the proof:
Who corresponds, and whether there will be more than one.
Who holds the submission account and who has the password — losing access when someone leaves the institution is a common and avoidable problem.
Who drafts the response to reviewers, and the internal deadline for comments from co-authors.
Who handles the data availability statement and where the data will sit.
What happens if the corresponding author changes institution before publication — the contact email should be one that will still work.
The institutional email problem
This is worth its own section because it causes trouble regularly.
Corresponding author addresses are usually institutional, and institutional addresses stop working when people move. A paper published today may need a correspondence in five years, and by then the address in the header may be dead.
Two practical steps: add an ORCID identifier for every author, which remains valid regardless of affiliation; and where the journal allows a second contact address, use it.
When the roles are contested
It happens, and it is usually not about ego. It is usually about two evaluation systems that both require the same slot.
Three approaches that work:
Split across papers rather than within one. If the collaboration will produce several outputs, agreeing who corresponds on which one is easier than splitting a single paper.
Use co-corresponding authorship where both systems recognise it.
Make the contribution statement explicit. Where a contribution statement is available, a detailed one can carry weight in evaluation even when the position does not.
What does not work: deciding by seniority alone, or leaving it to whoever raises it last. Both produce a result that one side experiences as a loss, and collaborations rarely survive two of those.
A short note on the other end
Submission is also where the affiliation list gets fixed, and affiliations matter for institutional reporting on both sides. Check that every author's affiliation is written the way their institution needs it — including whether a laboratory or centre must be named separately. Corrections to affiliations after publication are possible but tedious, and some evaluation systems will simply not count a paper whose affiliation string does not match.
Why does corresponding authorship matter beyond administration?
Because several evaluation systems, including Vietnam's, count it at a higher weight in promotion files and grant scoring — so it is a measurable credit, not a courtesy.
How should the role be assigned?
By who will actually do the work, who needs the credit more, who can answer post-publication questions, and whose institution holds the data.
Is co-corresponding authorship a good solution?
Often, but check first that both institutions recognise it — some count only the first listed, which helps only one side.
What should be agreed before submission?
Who corresponds, who holds the submission account, who drafts reviewer responses, who handles data availability, and what happens if someone changes institution.
