Is it acceptable to write a reference letter for yourself?

I don't think this is a grey zone at all. Senior professionals in all domains, not just acadaemia, routinely sign letters that they did not pen:

• Directors signing press releases written by communications managers
• Vice-chancellors signing letters to government officials written by deputy vice chancellors
• Academics submitting journal articles that were written by a co-author, although the research is shared.

It's completely irrelevant who actually put the words together, unless there is a question over the copyright of the text - which there clearly isn't here. All that matters is that the person signing the letter stands by its content and takes responsibility for it.

If you write a reference letter for yourself that is over the top, they won't sign it. No ethical issues here whatsoever.

(I can't answer whether other institutions would frown on the practice for other reasons though.)

EDIT

Ok, some research.

Tip: Don’t ask the candidate to write a draft of the letter for you. Most heads of labs say this rarely saves time and sometimes leads to a weaker letter. It is better for the candidate to provide you with all the necessary information, from which you can then pick and choose as you write your letter.

... (and for the applicant)

You may be asked to write a draft of the letter. As mentioned on page 1, many heads of laboratories say this is not a good idea. However, if you are asked to do it, do it!

http://www.hhmi.org/sites/default/files/Educational%20Materials/Lab%20Management/letter.pdf

When requesting a letter of recommendation, don't be surprised if your instructor or supervisor hands the forms back to you and says, "Sure, why don't you go ahead and write the first draft yourself, and I'll revise it and sign at the bottom."

http://www.petersons.com/college-search/letter-recommendation-writing-yourself.aspx

The “draft” that you provide to the recommender to sign shouldn’t be a draft at all – it should be a perfectly polished letter ready to be signed, sealed and delivered. Don’t be insulted when your supervisor decides to edit. They’ll want to apply their own language and voice to your content. ... I’ve been asked to write lots of letters of recommendations for students, but I’ve never felt strongly enough to throw the responsibility back at someone. That’s not to say I haven’t managed some phenomenal students; I just haven’t had someone come along who I think is up to the task. So, when he or she does come along, that person will really be top notch.

http://studentbranding.com/writing-your-own-letter-of-recommendation/

While the standard practice is for references to write their own recommendation letters, it’s becoming increasingly common for time-strapped individuals to ask you to pen the first draft of a letter yourself.

http://www.vault.com/blog/job-search/how-to-write-your-own-recommendation-letter/

Question: Is it OK to Write My Own Letter of Recommendation? Answer: The only time it is acceptable to write your own letter of recommendation is when the person you requested the letter from asks you to do it. Even then, it is important to be honest in the letter. Don't write anything the other person wouldn't have written. When you have finished, ask the person to look over the letter, verify the information, and sign. You should never forge someone else's signature.

http://businessmajors.about.com/od/recommendationletterfaq/f/WritOwnRecLet.htm

I don't see much (any) evidence of any ethical quandaries in a professor requesting and submitting a letter of recommendation directly from the student. These examples aren't cherry-picked - they're the first few hits that came up when searching for phrases like "own letter of recommendation" or "letter of recommendation myself".

In summary:

Is it acceptable to write most of the reference letter and have the prof make minor edits?

Clearly, yes - if requested to do so.

Do academic institutions frown upon this practice?

No.

Would it be considered an academic offense if a student wrote a reference letter for themselves and had a prof sign it?

Hell no.

If a student needs a reference letter for graduate admissions or that sort of thing, some profs will ask the student to write a reference letter for themself before sending it off to the prof for minor edits and finalization.

Some profs lie, cheat and steal as well...occasionally in their professional life. Just because a practice is common does not make it right.

This seems to be a somewhat common practice given that some graduate schools ask for several letters of recommendation even when it is not reasonable for the student to have developed deep connections with that many professors, and that most professors are just too busy to write quality letters for all the students that ask them to.

It is certainly "reasonable" to ask for several letters of recommendation for graduate admissions. That many students will not have had substantial personal contact with faculty is something to keep in mind as one progresses through an undergraduate program. Also connections need not be especially deep in order to result in a good letter: if the writer can be confident that the student will succeed in the graduate program she is applying for, that is enough. Often a truly outstanding performance in a single course is sufficient.

Is it acceptable to write most of the reference letter and have the prof make minor edits?

No, this is a form of plagiarism and academic fraud. What you pass off as your written word must actually be yours except where you explicitly document to the contrary.

Do academic institutions frown upon this practice?

Many of them do, yes.

Would it be considered an academic offense if a student wrote a reference letter for themselves and had a prof sign it?

It depends on the institution and probably the nation in question but in the United States: yes, it certainly could be. If I found out that this happened in an application that I read, I would at the very least throw out the entire letter; I would probably be inclined to dismiss the entire application. I would probably not contact the faculty member because in my view they are equally culpable (if not more so because they should know better), but I would be much more skeptical of letters coming from that person and even that institution in the future.

The above takes a hard ethical line, as I am very frustrated with other answers to this and related questions that seem resigned that one must accept unethical behavior in this situation. But here is a different kind of answer:

A graduate admissions letter that a student writes for herself is going to be a bad letter compared to a "real" letter written by a qualified faculty member. A graduate admissions letter is a communication between one mature academic and another: how would a 22 year-old young adult know how to write such a letter in a convincing way? Without having read hundreds or thousands of other similar letters, how would she know what the faculty want to hear? She wouldn't. If you write your own letter, you are at best forging an ineffectively written letter. Surely you deserve a better one?

As for faculty being very busy: yes, we are. As for having lots of letters to write: yes, we do. But writing such letters is part of our job, so a faculty member who does not take time out to write a good letter is not a good faculty member, at least not in this aspect. Writing a good recommendation letter usually takes several hours and often more than one sitting.

How can you help your professor write a good letter (on their own!):

1) Give them lots of time to write the letter.

Academics are busy, and our schedules are uneven. If you give me something to do six weeks in advance, then maybe in week three I'll find a spare afternoon and be able to do it. Leaving much less than a month for someone to write a letter is getting off on the wrong foot and already implicitly asking for less than the best possible letter.

2) Provide information about yourself rather than waiting for the faculty member to ask.

You should not write the letter yourself, but you should certainly include all information that you think might be pertinent, and you are well within your rights to highlight certain information that you think might be especially pertinent. Preparing something like a CV but tailored for a good letter rather than a job would be ideal.

3) Do everything in your power to minimize the attendant clerical work in submitting the letter.

Faculty members are busy and also, honestly, a bit lazy/snooty about routine work. If you tell me to mail a letter to a certain address, then there is going to be a whole day in which I print out the letter and don't get around to correctly putting it in an envelope. If the letters still need to be mailed (fewer and fewer do, and most but not all faculty members prefer to do things electronically), it would be wise to provide a self-addressed stamped envelope. If the submission is electronic, again try to ensure that the faculty member needs to do as little as possible. Ideally we get a website and a password, we enter those in, and we immediately upload the letter. Much more than that is asking for trouble. My own university makes faculty members jump through many more hoops to submit a letter, and this worries the hell out of me.

3') If your letter needs to be sent to multiple locations in a way which requires the faculty member to do something multiple times, see if you can figure out a way for the faculty member to submit the letter only once. E.g., perhaps there is an administrative assistant (AA) at your institution that will agree to receive the letter and take care of the nitty gritty of sending it to various places. You may have to ask for this, and you should ask, as nicely as you possibly can. Given the choice between getting a faculty member to do this clerical work and getting an AA to do it, you want the AA to do it: they are superior in every way. Always remember to be extra nice to the AA's: you want in fact to be nicer to them than the average person they have to interact with, as then they will notice and do better work for you than the average person they have to interact with. If you're asking an AA to do something which it is not absolutely clearly part of their job description, go ahead and ask but be extra extra nice: a small gift at the end is a classy move.

4) Don't be shy about checking up on the faculty member to see whether the letter has been written.

I frankly expect this, to the extent that if you ask me for a letter (including giving me all the information) and never check back again, I almost believe that you changed your mind and didn't really need it after all. It is totally acceptable to ask multiple times for the faculty to turn in the letter. I'm afraid that there has been "email alert inflation" in recent years, to the extent that if I only get one email about something, it feels almost optional. Really important things have a way of resulting in multiple emails coming at shorter and shorter intervals. Even to get me to write my grant reports they are not shy about sending several. You should always be nice about this -- at any point the faculty member could in theory change their mind -- but we are grateful if it is your mental energy which is being spent on making sure it gets done.

There are certainly different degrees of flagrancy here. I know some good professors who will ask for a short paragraph just to "get the juices flowing", but will then modify and expand it into a full letter. I don't think that is wrong, although I wouldn't do it myself.

If the professor signs a letter written by the student without making any modifications, or if the final letter consists mostly of the student's words, I consider that fraudulent -- but it is primarily the professor's fraud. After all, the professor is the one signing his name.

When I write letters, I require the student to give me all the relevant records as well as a description of what they think I should emphasize in the letter. Not a single word of their description would ever be pasted into the letter, and I don't use it as a starting point for the text. It is simply their opportunity to remind me of the impressive things they have done, which I will then write about if I agree.

If I can't write a letter for them in my own words -- for whatever reason -- I will decline to write one at all.