Much work is put into constructing grant applications: sharing them would allow diffusing novel ideas faster, which I would expect it to be seen positively by researchers. Why aren't grant applications systematically published online, regardless of whether they are accepted?
Confidentiality is a real issue. In an ideal world, ideas could be shared freely, and this would enhance the flow of knowledge and the progress of science and other research endeavors. However, we live in a world with limited resources. In particular, there is a shortage of grant funding. Most grants do not get funded.
Consider the following scenario: A researcher has some very interesting preliminary data, and they want to get a chance to follow up and do a more full study. So they apply for a research grant from some agency. If their proposal is not funded, but the content of the proposal is made available online, other researchers may troll through and pick up on the idea. If another scientist with access to more resources sees the proposal and decides to do the follow-up work themselves, this robs the person who originally had the interesting idea and started the project of a fair shot at completing it and getting adequate credit for it.
Sometimes, having a good idea for something to study is the hardest part of the research process. I personally have some ideas for projects that I do not want widely disseminated, because I would like to do them myself at some point, but I don't have the time and resources I need right now. They do not involve urgent questions that really need to be answered, so there is no great loss if these avenues are not pursued promptly. Thus I feel justified in keeping some of my best ideas confidential.
I can think of many reasons why researchers are not prone to sharing (especially failed) grant proposals, but fundamentally I think your premise is flawed:
sharing [grant proposals independently of success] would allow diffusing novel ideas faster, which I would expect it to be seen positively by researchers.
This I think is wrong on two axis. First, the trivial one: most researchers are for pragmatic reasons not happy if their ideas are "diffusing" before they had the time and money to work on them themselves. This is usually called "being scooped," and it's not exactly something that researchers actively try to make happen. Yes, as researchers we love to make people aware of our work, but generally only if we already have results available or at least sufficient head start that getting scooped is unlikely.
Additionally, I am also unconvinced why sharing proposals would even "diffuse novel ideas faster". In many fields, there are nowadays so many finished, peer-reviewed papers to be read that adding more stuff that fundamentally proposes at the time of writing untested ideas and hypotheses to the discussion would likely mostly add noise, and hence get largely ignored. As I wrote in a comment above, at least in my field ideas are a dime a dozen, and unless you have thought of something particularly radical or mindblowing (which most proposals are not, just like most papers are incremental rather than radical), the content of a proposal alone is unlikely to get much attention.
The context of this answer is applied CS in Germany. It may or may not be applicable to other contexts.
I fully agree with
Much work is put into constructing grant applications
However, I think
sharing them would allow diffusing novel ideas faster
is unlikely and misses an important point: Grant applications often do not contain much in terms of actual ideas.
At least with grant applications directed at national ministries and connected organizations, the target audience of the grant applications (at least the part that will decide on whether or not to accept a given application) is not particularly knowledgeable about the field. They cannot possibly evaluate whether a particular approach is promising; they can only check whether a given grant application appears to be aimed at what was asked for (e.g. in a call for proposals), appears to be achievable based upon the applying entity's previous experience, and looks like a solid endeavour that can be completed as planned with the resources (money and time) provided.
At the same time, from the applying researchers' point of view, the goal is not to secure funding for a particular idea, but to secure funding. If possible, without any ties attached, but as that is usually not possible, the art of writing a grant application includes trying to make realistic promises that leave enough ways of interpretation so the actual direction the research takes is still relatively open.
Therefore, such grant applications focus on the following content:
- Defining the fields in which the applicant1 can provide their expertise. This builds upon the applicant's experience from other projects and previous publications. It touches upon solutions only on a per-topic basis (fictitious example: "Based upon the previous experiences in automated manufacturing processes, the applicant has the expertise to examine opportunities for using directed microwave radiation in manufacturing."), not with concrete ideas.
- Addressing the points/topics requested by the call for proposals, on a rather high level. Like the previous point, this will rather indicate possible combinations of topics with keywords from the call for proposals, but not elaborate a lot on concrete ideas for new developments.
- Explaining the general expected outcome from the results (such as scientific publications and practical applicability). The information related to this will often focus on the typical publication rate in a given field, and the current scientific activity level with respect to the topics asked for in the call for proposals, the topics the applicant has experience with, and - most importantly - the combination thereof.
- Justifying how it is feasible. This might go more in depth, but not about solutions. Rather than that, there may be information about operational and administrative details of conducting the research (usage of quality assurance tools; planned communication and decision processes in a consortium; ...).
So, while there is plenty of information in such a grant application, it does not necessarily describe any concrete ideas. And in the end, that is probably not too surprising, given that the start of a grant application is usually a somewhat vague idea about "several intriguing topics that should be combined", or "one intriguing topic that we should extend our expertise to and that sounds compatible with what we have been doing so far", while the actual ideas are developed only during the run of the grant.
1: This can be an individual researcher or a whole organization, or anything in between.
sharing them would allow diffusing novel ideas faster, which I would expect it to be seen positively by researchers
The lack of current availability suggests that most researchers would view it negatively. If you feel publicizing your grant proposals would help spread your ideas, you can easily post them on your web page. A few people do this for funded grant proposals, but it's rather unusual.
Specifically, in the areas I work in, I'd guess that perhaps 5% of funded proposals are directly posted by the PIs. Government-funded grants are generally available with some effort through FOIA or other open records laws, while other funding sources may not make them available at all. Unfunded proposals are never made available by the funding agency, and it's really rare to see them posted online by the PIs.
What this indicates is that almost all PIs either don't want their proposals publicly distributed any more so than they already are, haven't thought of distributing them online or don't know how, or feel it's not worth the very minor effort of adding them to a web page.
I agree with the other answers regarding the reason for this, but I want to emphasize that regardless of the underlying reasoning, you can tell most people aren't enthusiastic about distributing their own proposals.
Where would be the point in sharing grant applications? The only conceivable outcome (which you actually state so yourself) is that others will likely produce similar grant applications. It then basically becomes a random matter which of the applications will actually be granted, and that means that a potential grant may end up with someone who is not invested in the matter to a degree where he would have been even able to put together a well-researched grant application.
You don't apply for a grant out of the blue. You apply when there is a tangible chunk of research work you have whittled down to something that can be resolved by you with reasonable chance for success and/or progress in the scope of the grant.