Grant writing, wasted time, and Red Queen effects

I was recently having tea with a philosopher who is the head of an interdisciplinary research group. We talked about grant proposals. My interlocutor said he devoted a lot of his working hours (at least 1/3 in his estimation) to writing grant proposals. He also knew someone personally, who was not a philosopher, but someone from an empirical discipline, who devoted as much as 70% of his time on grant writing. That latter person even said that he can now scarcely keep up with the literature in his highly specialized field - let alone contribute original research. But, given that his research group (comprising many PhD students and postdocs) depended on his ability to secure grants, there was no other option but to devote more and more time to the grant writing process. Since grant schemes often ask for unrealistically elaborate timetables and detailed projected results, many experienced grant writers have turned to this heuristic (they have admitted this freely to me, and are unabashed about it - I haven't tried this for myself, but the practice is widespread):

1. Write a grant proposal that describes the work you have recently done (let's call this research project X).
2. If your proposal gets funded, you start doing the research you really want to do (research project Y)
3. If asked for a report of results, you simply mention the papers that are now in press, undergoing review or are recently published from project X; you do not mention the actual work that is now going on in your centre or lab, project Y.
4. About 1 year off from the completion of your current grant, you start developing a new grant proposal, this time detailing how you will carry out project Y (which, of course, is already about completed), allowing you in the future to pursue project Z.

And so on. This practice illustrates, I believe, that there is something deeply wrong with the grant making process as it is currently practiced.

This problem has been described using the tragedy of the commons, but I think the Red Queen effect might be a closer analogy. The Red Queen effect in evolutionary biology describes how organisms must constantly adapt and evolve simply because other organisms do so as well. As a result, in some cases, the organisms in time t+1 have traits that are more harmful to them than in t, but merely in order to keep up, such traits get enhanced in subsequent generations. The burdensome tail of a peacock is a standard example - it's a serious handicap to the owner in terms of parasites and predation risk, but without it, a peacock would not be able to secure a mate.

Similarly, it is not uncommon for directors of centers in the life and physical sciences to decrease their personal research productivity to almost zero in order to enhance the overall productivity of their lab. This tendency is exacerbated as funding is being increasingly taken away from universities and concentrated in grant schemes that promise to award excellence in research. In The Netherlands, the policy was to concentrate money for research in their ambitious NWO Veni, vidi, vici scheme. Similar trends are happening in the UK.

Overall, this is translated in more postdoc opportunities and less faculty positions, a situation that has pros and cons. A recent opinion piece in the Guardian highlights the problems in a stark way. The piece summarizes empirical research on grant writing and its costs. A summary of the findings:

• "Scientists in Australia spent more than five centuries' worth of time preparing research-grant proposals for consideration by the largest funding scheme of 2012. Because just 20.5% of these applications were successful, the equivalent of some four centuries of effort returned no immediate benefit to researchers and wasted…" Link to paper in Nature here
•  "We show that the $40,000 (Canadian) cost of preparation for a grant application and rejection by peer review in 2007 exceeded that of giving every qualified investigator a direct baseline discovery grant of $30,000 (average grant). This means the Canadian Federal Government could institute direct grants for 100% of qualified applicants for the same money. See here. 
• If there is no cap on individual effort (time spent in writing grant proposals), the total net returns of grants is lower than if grants were simply given at random, rather than to those who have the most promising grant proposals. Only if there is a fixed cap of 10% time investment in grant writing does the system perform better than in the random scenario (preprint here). The problem with the scenario in 3 is that the incentive to defect is high: under our current system, spending time at writing grants does improve chances and given the benefits of grants, many of us spend more than 10%. Indeed, given that others probably spend more than 10%, we have little choice but to do the same (Red Queen effect). I would also like to add 2 additional problems of the current model 
• A culture where most research money comes from grants increases the Matthew effect: people who got lucky with an initial grant have a better track record (e.g., in pubs, but, ironically, also a better track record of grants), and so are more likely to get even more grants 
• Grant proposals are rarely anonymized. This may give rise to unconscious forms of bias, unfairly disadvantaging some people over others. A grant application from a prestigious department might be given more serious consideration than one from a less prestigious one. Women might be subject to implicit bias. Indeed, the ERC (European Research Council) has a useful summary of statistics showing that the percentage of women who get awarded ERC starting grants and advance grants has been consistently smaller than the percentage of women who have applied for grants. This corresponds well with recent studies showing that research abstracts that have a woman's name on them are evaluated as less good than those with a man's name by scientists. Also, women who do get grants consistently get smaller grants than men applying for the same funding agencies. 

Given these problems, we could think of radical and less radical solutions:

1. Give grants to everyone who asks them. Eric Schwitzgebel says it as follows "Give money to researchers for their research without their having to apply, and let them spend it on any reasonable research expenses." This equal distribution would of course mean smaller grants for everyone, but it would mean an enormous increase in research productivity.  
2. Allocate grants purely randomly (so not for everyone, but for a random selection of researchers). Anyone who wants to have a grant puts their name in a hat, and the grants are allocated on a lottery basis. 
3. Give grants on the basis of past track record alone. Those who have received grants in the past and have a good output in terms of quality (venues published in), and quantity (no of pubs) can get new ones, others do not. Of course, this would also result in a Matthew effect, but strategy 3 could be improved as follows: provide a fixed percentage of all money (say, 50%, but the percentage might be different for optimal results) for those who have done good things with their grant in the past, and allocate 50% randomly to those who have never received a grant before but put their name in a hat. 
4. Make the process of grant writing less time inefficient, e.g., evaluate grant proposals only on the basis of an abstract of 500 words, and make the review process double blind to avoid Mathilda/Matthew effects. An abstract, no matter how carefully crafted, doesn't take up that much time and allows for some peer review. Also, it would get rid of the obligation of writing detailed projected results and timetables, which as experienced grant writers know, are mostly conjecture anyway. 

There are doubtless other strategies better than the current one, but I think many of them would fare a great deal better than our current model.