 Hello everybody, my name is Mackenzie Smith, and I'm the University Librarian at the University of California Davis Campus, and thank you for turning out in this gigantic room for the second to last session of the day. My presentation isn't a full hour, so we'll have that and then hopefully we'll have a little time for a discussion if you have questions or suggestions or ideas about this research and what to do next. So I'm going to be talking about a project that was completed a few months ago called Pay It Forward with funding from the Andrew W. Mellon Foundation. Thank you to them. And some of you have heard previous talks about this, so this is the final phase. We're done with the research. We've written the report. We're working on more human consumable versions of it to share, but this is sort of a complete accounting of what we found in the project and then some discussion about what conclusions we take from it. So the motivation for the project from the very beginning has been this observation that here in the US public policy is pushing us towards the so-called green road to open access where you deposit author manuscripts in institutional repositories to share them or disciplinary repositories like PubMed Central in some cases and that that replicates the literature that's in the journals itself and that that's how we achieve open access. Whereas in the UK and Europe and other parts of the world they're really looking at the journals themselves and making the articles openly accessible, the so-called gold road. And we in the University of California system said, well wait a minute, you know, we can't have both of these at the same time. It wouldn't be sustainable or affordable or sensible. So let's figure out, you know, which of these works better from a long-term sustainability point of view since for libraries this is still very much an economic issue as well as a social issue about providing access to the literature. So I'm sure most of you have heard about OA 2020, which is a manifesto, a proposal that was started at the Max Planck Institute in Germany to see if, well, can we flip the entire journal publishing system to gold? So article processing charge funding models by 2020 and just get this over with, right? So what would that look like if that actually happened? What would it mean for us? Particularly in large North American research universities who have a very different funding model than what they have in the UK, Europe and most of the rest of the world. So we set out to build a set of financial scenarios or what we call models to depict the financial implications of an article processing charge-based system of scholarly journal publishing for the conversion of the entire system for large North American research institutions, okay? So the perspective is very much us, you know, US and Canadian institutions and if the whole system flipped all at once. So we started out with quite a lot of quantitative data. This is the team that was working on that. Ivy Anderson was going to be here to co-present with me but she's had enough travel for a while, so she bigged out. And what we did was start by collecting a lot of data We got journal expenditure data from all of our partners and I think they were listed on the earlier slide, so that was all of the UC system, all 10 campuses and the University of British Columbia up in beautiful Vancouver, Ohio State University and Harvard. Those were the partners who provided us data. So we got journal expenditure data for a five-year period, 2009 to 2013 and corresponding publication data for the same five years and that means all of the authors from those institutions, what they published during those five years and that data came from the Web of Science and Scopus. All of this data is ridiculously hard to get and normalize across institutions, so that was a huge amount of work. We also tried to get grant data, so research expenditures at the partner institutions from the NSF Herd Survey. They don't have that in Canada, so it was limited to US institutions, but we also wanted to know how much grant funding is available at these institutions that we're looking at. And then finally we looked at article processing charges and publication costs across a set of publishers, all of those represented by the ALPSP organization. So we started out asking questions like, well, what does it actually cost to publish an article? And again, that is a very difficult question to answer. The range is, well, really from zero, but 500 of those who have any kind of revenue stream, it's between $500 and $4,000 today and people speculate that it could go much higher. A plausible minimum that we found looking at current publishers was around $1,000, including a small surplus, so they can reinvest in their infrastructure and so on. And looking at the journals that our actual authors were actually publishing, that number was a little higher, it was almost $2,000 per article. This is the cost to publish. However, we very quickly concluded that that's a really stupid question to ask, because it depends so much on the particular publisher's particular fixed effects. So are you in Sonoma or are you in downtown London? And so things like rent can really affect what it costs to publish, because that's based on the costs of the publishers' operations, and therefore things like how many journals you have have a huge impact. So it's really not a particularly useful thing to ask what it costs to publish, because it's all relative. So then we looked at what APCs might cost in the future. If we forget about what publishers actually earn today, how might they determine an APC? And what we found is that there's a very strong correlation between what they're charging now and the impact factor of the journal. So an impact factor, we didn't actually use impact factors. We use SNPs, the source normalized impact per paper, which is a different metric provided by Elsevier. But the reason we chose that is that it's a little more sensitive to disciplinary differences. We all know that impact factors really don't work across domains and disciplines. So we use SNPs and we found this very strong correlation. And that meant that we could do a regression analysis to figure out what a publisher would charge as the APC for a journal if we know what its SNP is. So we came up with this formula. The article processing charge would be $1,147 plus a margin that depends on the perceived quality of the journal, the SNP. And so that means that a baseline journal, something like plus one, would charge around $1,856 per article. Now, I'll admit that this is very, very crude, but it's something that nobody's tried to do before, is say, well, what will publishers charge? Because we all know that publishers don't charge either subscription fees or APCs based on what they need to bring in. They're making a profit, so they're charging as much as they can, typically, right? So this is more a question of what would a reasonable approach be for a publisher, and this is what we came up with. That means that we can start to look at things like, okay, well, what if all of the articles that I published in a sample year, 2013, had charged an APC? What if the library or the authors had been paying fees instead of subscriptions for that year? And so now I can look at a journal budget. This happens to be my institution. So my journal budget is approximately $4 million for scholarly journals. My institution published 3,593 scholarly research articles in that sample year. So then if I apply that formula to all of those 3,593 articles, I come up with what it would have cost me if I'd been paying APCs, and it was almost $7.5 million, which is significantly more than $4 million. Okay, so then you have to think about other sources of income to pay those APCs, and that's where the grants come in. So of the 3,593 articles in my institution in that year, quite a few of them, well over half, had a grant associated with them, right? So almost $2,500 with about 1,000 left over that didn't have a grant of any kind. So if we estimate that grants could help pay for the articles that have a grant associated with them, all of a sudden what I would have had to pay drops to $2.2 million, which is significantly less than my current journal budget. So you start to see how these factors play in, and we ran that same analysis for all of the partner institutions, and that's how we came up with this kind of well. For some institutions that are very research-intensive, which are down in the red end of the spectrum there, it's going to cost more. But for those in the green end of the spectrum who publish a little bit less, it's going to be a lot less than today. So you have your winners and your losers, and this represents a huge challenge for the community, because overall the costs are the same or lower, but who's paying changes dramatically under this new model. So once again, if it's just the library budget, nobody wins. We really can't afford it, even at the wealthiest of institutions. You know, that's Harvard, the little tiny one on the left there. But if you bring the grants into it, everybody, all of these institutions are bringing in enough grant money to more than cover the APCs for articles associated with that research, and we all make out quite well. So we conclude from this part of the research that APCs aren't perfectly predictable, and we haven't quite figured out how they vary across disciplines, but we can come up with some crude estimations that will get better over time for what, you know, these article processing charges will be, and it's something we can negotiate with publishers potentially. In North America, it's pretty clear that the library journal budgets alone, sorry, Slack wants to talk to me. Library budgets alone are not going to cover APCs for all that fairly small institutions, but authors who have grant funds will be just fine, and so the $64 million question quite literally is, you know, can we capture the grant funds that could pay those APCs, right? And unless APCs inflate uncontrollably, like the scholarly subscription market has, in the end, we may be better off. So the question is, would this be sustainable over time, and I realize this font is tiny, so I will read this to you. Funding a journal with AP, this is, sorry, let me back up. This is an article I read which was an interview between Richard Pointer in the UK and Sir Tim Gowers, who's a field medalist, mathematician, and a big open access proponent. And in this interview, Tim Gowers said, funding a journal with APCs is acceptable if authors do not have to pay the money themselves. In other words, that's a perfectly fine model, just don't make authors have to deal with this, right? Then he goes on to say, I think this open access big deals, which is starting to happen in the UK, is beginning to happen and that publishers are finding ways to create an APC-based market that will be as dysfunctional as the subscription-based market is. The basic problem with APCs is that publishers can charge whatever they like, knowing that if a university starts to tell academics that they have to publish in cheaper journals, there will be an uproar about the perceived threat to academic freedom. I have never seen a convincing explanation for how a properly free market in APCs can work. And I just laughed when I read this because these are completely contradictory statements, right? Authors shouldn't have to pay, but we don't have a free market, right? So that's an important thing to keep in the back of your mind. And now we turn to the qualitative research that we did in the project. And this was led by Dr. Carol Tenepere at the University of Tennessee in Knoxville. And she surveyed and talked to over 2,000 faculty, postdocs, and graduate students at the partner institutions about their attitudes towards publishing open access, who should pay, and a variety of other things. So one of the things she asked is, what are the most important factors to you when you decide where you're going to publish? And none of you are going to be shocked to hear that number one was the quality and reputation of the journal, right? That's what they care about. And then there was a list from there, you know, is my research a good fit with the journal, who is the audience for the journal, will I be reaching the right people, and so on, with open access dead last, okay? And in other questions, we discovered that they actually don't object to the concept of open access at all, most of them. It's just not an important criteria in deciding where to publish for the vast majority of authors. So her takeaway was that, you know, looking at this on balance, it's evident that reputation building within a specific field is at the heart of what matters to academic scholars. I'm sure none of you are shocked. So another thing we discovered is that authors are paying APCs today. We asked people, will have you ever published in an open access journal and fully a third of those 2000 survey respondents had published in an open access journal, and of those, the majority had paid in APCs. So this is not, you know, this is happening, okay? And I didn't mention it earlier, but when we looked at the percentage of the literature that's now open access, it's 14% and going up fast, okay? So this is happening kind of independent of what we think about any of this or whether it's sustainable. So we did talk about your willingness to pay from different sources of money, and it was very interesting. So we gave them a range of options from your own pocketbook to your department, discretionary research fund, and the library. And one thing we noted is that half of publications are from life sciences and medicine, okay? So that was an interesting fact. And then about a third of respondents from those disciplines are willing to pay quite a bit in open access fees. But we looked across all the disciplines and it varied widely, with, of course, the humanities being kind of at one end of the spectrum for how much they would consider. I mean, to them, the idea of paying to publish is just like insane because they haven't really been part of this experiment here too far. And then at the other extreme, you have the life scientists who are used to it. The majority of them have tried it and they're willing to pay quite a bit, relatively speaking. So interestingly and probably not surprisingly, the answer to the question of who should pay was well not me, you know. So if it's coming from their own pocketbook, they wouldn't be willing to pay more than $100, okay? And then it kind of goes down the list from there and the library was an interesting one because for the library, 32% said nothing. The library should pay nothing because of course they're worried that we'll stop buying books and other things like that, right? But the more nuanced ones were willing to let us pay $3,000 or more. So you had these two wild extremes around the same percentage but the bottom line is that authors are quite price sensitive. So when they think about what they're willing to pay depending on where the money is coming from, their answer is quite different and that's extremely important to keep in mind. So another slide that you can't read so I will walk you through this. We worked with an economist, Mark McCabe, who's been in this kind of area for quite a long time now and thinking about sustainability and what it's going to take to get us out of this situation that we're in now where there's a huge disconnect between the people who are benefiting from it and who's paying for it, kind of like healthcare. And the behavioral objective we want is for authors to really be like consumers where they're choosing the best platform for their article given the price, the funds that are available, the platform for readership, editorial quality and all the other things that they care about. But we want them to choose that and we want publishers to compete more for authors so that they're responding to elastic demand by competing for those authors' business. And the claim is, and this is simple market economics, that under ideal conditions, competition in an APC environment will lower the cost of scholarly communication over time. And then there are many mitigating factors to that like platform ownership concentration if there's only one platform obviously, competition drops off, delegation of APC payment responsibility. So if libraries are still paying the fees and the authors are not involved, then this doesn't work. So the implementation plan would be to give authors discretionary research funds to pay their APCs or other things. Okay, so this is very important and I'll give you an example. At University of California, Berkeley, down the road from me, they've instituted something called a bear account and their tenure-track faculty get small research budgets every semester. It's like $2,000 and you use it or lose it and you can use it for any valid research-related expense. They're incredibly popular and it isn't costing the institution all that much to do this. And so it was a real morale booster and a lot of faculty who don't have big grants just can't believe how fortunate they are and it gives them a way to pay the APCs. They can pay it from these small research grants. And then, of course, if that wasn't enough, then institutions would have to backfill and pay the difference to level the playing field. But this is the concept that authors would have control over some of the money so that they are behaving like consumers and they are making decisions about where to publish based on the cost-benefit ratio. So this introduces more of, you know, that sort of healthcare model again of a multi-payer system where you don't want to disadvantage anybody by forcing them to come up with the money, you know, from nowhere, but you want to introduce more of a little bit more responsibility on the part of the author in the decision about where to publish because otherwise, you know, they don't know what it costs so it can inflate ad infinitum like today. And there are lots of ways to accomplish this, by the way. The Berkeley example was just one, you could put money in startup packages, you could have a fund that the department administered, you know, there are many, many ways of making this happen. So in that model, what would it cost an institution where you have this multi-payer scenario? So back to my institution as an example, let's say I gave faculty authors a subsidy of up to the break-even point for my budget. Take my current budget divided by n number of papers, that's how much you get. You get $1,164. So that means that my share of the cost is exactly the same as today, $4 million. Grants available to authors who have them will cover another $2.5 million for all of the papers that have those. And the delta is about 1,000 papers that these discretionary funds would have to cover, which would amount to about $1 million in new money from the campus going to the authors to pay that. And so that's a 25% increase over current budget, which is my $4 million. But it's not an insane amount of money. In a different model, my subsidy is higher. My subsidy now goes up to a baseline journal. And then grants cover the rest, which is now slightly, or sorry, grants cover the articles that can, and it's under a million for those now. And then the discretionary funds drops down to $0.3 million. So the overall cost to the institution has gone up even higher because the library's budget went up significantly. Less money is coming out of grants. More money is coming from the library budget in this one. The point of these two examples is to show you that there are a million variations of this. These are different dials that you can turn. Library pays more, grants pay more, authors pay more. Some other source of income from the university pays more. So, but you can tweak these different levels to figure out where you're getting kind of that ideal behavior that you're looking for without disenfranchising anybody. And by the way, it doesn't have to be the same for every discipline. You could tune these dials differently for humanists versus life scientists and so on. So our prediction coming out of this project is that giving authors discretionary funding introduces price competition without interfering with their choice of where to publish, which is absolutely essential since we saw that ranked list of what they think about and you're not going to get authors to change their behavior in the short term. Over time, with competition, they may choose to publish elsewhere because it's a better deal for them. But in the short term, they're looking for that journal that is the highest impact factor or SNP in their field and that's where they're going to stop. So, giving them the money to do this without telling them where to publish is our absolute best chance to create a competitive market for journals which we've never had. Encourage authors to explore other options. Many of you are creating new publishing platforms that we would like to attract authors to, but as I said, they need some motivation to do that because otherwise they're just going to go right back to their default favorite journal and this is how we'll drive costs down over time. Now, I'd be the first to admit that there are lots and lots of gotchas with this model. There's a huge and valid concern that in this model, the rich would get richer. The harvards who have very large endowments could afford to give their authors 100% of the fees in every case if they wanted to, whereas a smaller institution with less money might struggle to find the money to publish the articles they're writing. We don't actually think that's what would happen because the amount that Harvard would have to find is really significant because they publish a lot, but this is a valid worry that we don't want to make it so that if you don't have money, you can't publish. So that means that we, as the scholarly communication professionals in the world, would have to come up with ways to level the playing field, like endowments that people could apply to for the funds. And that is essentially what's happening today with APC waivers that publishers have and society subsidies and that kind of thing. Another issue is, of course, many, many disciplines that don't get research grant funding. At my institution happens to be very STEM heavy, so lots and lots of researchers have grants that they wouldn't mind applying in this way, but in other institutions that aren't a STEM heavy, they just don't have that level of grant funding. So that's a huge differentiator in North America. And then you've got young and independent scholars who don't have access to grants or institutions. They're not part of a university necessarily, or they're just starting out, so they don't really have access to funds. An issue that comes up every time we have this conversation is about the global South, all the developing countries in the world who have much, much less research funding than the US and Europe. And the issue there, again, is, well, how do we know they'd be worse off? It depends on how much they're publishing. That's something that we need to look at in the same kind of way and say, okay, take a typical research university in Brazil. What's their library budget? How much are they publishing and what access do they have to research funds? Would it be better or worse for them? And then you've got liars and cheaters. So part of this is depending on authors being willing to tap into their grant funds to pay fees. And we, most of us, don't have really good mechanisms to track who have research grants that might be eligible to pay APCs for a particular article. So there's a lot of trust involved in this, and you can come up with policies, but whether we'd have any way to enforce those is a pretty good question. And then finally, the library has historically played a really critical stewardship role in scholarly communication, so we're the ones who worry about things like long-term preservation. And if in this new world, libraries somehow get disintermediated, nobody's going to care about that, right? Authors aren't going to be demanding that publishers worry about long-term preservation. So we have to figure out a way to make sure that the institution particularly the library is still involved in this in a significant way. So, you know, on balance, I think we're fairly optimistic that this OA 2020 proposal could work. It would be financially sustainable with a bunch of caveats that we need to work through. But I just want to end with a little story because I think a lot of us are thinking, you know, how dramatic a change this would be to the world. But I've been looking at the history of page fees recently. And in high-energy physics in particular, there's this very interesting phenomenon where prior to 19... Well, about, you know, World War I era, it was subscription-based. That's how journals were funded. But after World War I, there was this explosion of research in physics. So the publishers couldn't keep up. They were starting new journals, you know, every other month. And they couldn't bring in enough subscription revenue to really cover their costs. So they introduced this concept of a page fee that authors would pay, or really what their idea was is that the so-called research patrons would pay, i.e. funding agencies and universities that hire these physicists. But that was a really revolutionary concept to a lot of physicists and authors. So initially, some grant agents... So sorry, some foundations stepped up, like Rockefeller, and they paid the page fees for a while so that the idea could get socialized and become acceptable to actual scientists in the field. And the fees were actually voluntary for the first 10 years or so, but became more and more normative in the field over time. And this is all back... This is back when AIP was first formed. So they were really leading this effort. And a lot of it had to do with the rhetoric around what an article is. Is an article really for the reader, or is it for the author? Is it a public good, or is it part of your research record? So the point is articles have multiple purposes. One is to convey the research to an audience of fellow researchers. The other is to promote your own research and your institution. It's a promotional tool. So what happened is that over time, the balance of these two things, the page fees and the subscription fees, kind of shifted. And by the 1950s, they had come up with this interesting concept where... Remember, this is the print era, right? So in that time, there was heavy infrastructure cost to just produce the journal in the first place. And then there was an almost equal infrastructure cost to get those journals out to the people who wanted them, to mail them to everybody. So they said, okay, well, authors should pay for the infrastructure to produce the journal, and readers should pay for the infrastructure to get it to the reader. And that kind of made sense in the print era. So by the 1960s, you see that page charges were the larger part of the revenue for these publishers, and subscriptions were the minority. And then something changed. The tax law changed around mail, physical mail, and a few other things changed so that it became a little more expensive to... Sorry, I think it was a little less expensive to mail things to readers. And that also was about the time that commercial publishers swooped in and started buying up scholarly journals. So this is when Elsevier started to grow enormously. So through the 70s and 80s, it shifted back to subscriptions as the majority source of revenue for publishing. And this idea that articles are a private commodity, not a public good, kind of went back into vogue. And that led to a rapid rise of subscription rates, which is where most of us came into the picture, right? 70s, 80s, 90s. And so to us, we didn't remember the era of the page fees. We just remember the era of the ever-rising subscription fees. But then you have the internet, where all of a sudden the cost of distribution drops to zero. So now almost all the cost is on the infrastructure to publish side, not to distribute. And that's when we start seeing open access come back into... Or this idea of the article as a public good come back into vogue. And all of a sudden, we're talking about open access. So in a way, what we're experiencing now isn't anything new at all. It's just this cycle we go through of whether the article is a public good or a private commodity, and whether it's in the best interest of the author or the reader. And we're seeing the balance of that shift back to reader public good. And it's not that what's happening today is exactly the same as page fees, but if you think about it, an APC is just basically a page fee. It's the same idea. So really, this isn't new. It's just that we didn't do it in every discipline. And the amount we're talking about would be a little bit higher than it historically was. And in the OA 2020 manifesto, we flipped entirely to page fees away from subscription fees, unless there are other ways we can capture some of that revenue from non-research producing institutions. So what do corporations pay who don't actually publish anything but are huge beneficiaries of the research literature? So with that little history lesson, I think I'll stop and point you to the report, which is a 200-page doorstop, if you would like to read that. And we are working on, as I said, a couple of articles that kind of boil it down to the essence of what I've told you today. But some of the things I'm interested in talking about are what to do next. We've talked about applying this kind of thinking to monographs, which are also going through potential conversion to author-funded payments, this issue about the global south, and whether we want to look at how this model would affect other parts of the world. So there are lots of things we could do to follow on this work, but for now, let me just stop and thank you for your attention. In the usual way of seeing, I have no people here to help with this. There are microphones. There's one right there. I can probably hear you just fine and repeat your question if you have any. And I realize it's a huge room, but please, I'd love your ideas. I'm Mackenzie. Hi, Peter. I'm Peter McDonald.