Turning Academics into Entrepreneurs

If an academic invents something while working at a university, who should own the patent? Who should be entitled to the profits? And how does this affect the rate of entrepreneurship in an economy?

These are questions within the debate about how we can get more of the benefits from academic research to translate into private startups and economic growth. While no one would claim that all academic research should have commercial considerations, I think it’s fair to say that some of it should. Some academics will want to create startups to commercialise their discoveries (‘spinouts’), while others may want to patent ideas, or license technologies for use in the private sector.

In this post, I’ll be considering what can be learned from the experience with ‘professor’s privilege’. This is the concept that academics should own the rights to any technologies they create or patents they file in the course of their employment. Today, Sweden is the only country in the world that has this policy. But several countries used to have it, and in a paper from 2018, Hans Hvide and Benjamin Jones exploit a fascinating natural experiment: the end of professor’s privilege in Norway in 2003. Toward the end of the post, we will turn our attention to the possible lessons for Ireland.

Fostering more spinouts is an acknowledged policy objective for the current government. The National Development Plan 2021–2030 profiles a successful spinout, and aims to increase the “transfer and commercialisation of research”.1 Impact 2030, the national innovation strategy, has a target to increase the rate of start-up generation from publicly funded research by 25 percent.2 In their announcement of their new CEO, Taighde Éireann – Research Ireland cites among his qualifications extensive experience with “licensing intellectual property and scaling new company creation”. Spinouts are just one way to encourage more startups, but it’s worth taking seriously what’s blocking them.

Subscribe now

The professor’s privilege

Before 2003, Norwegian academics owned the full rights to any inventions they patented. They did not have to give up any equity or revenue to the university if they founded a company while working there. But after a new law was passed in 2003, any patents filed would be owned by the university, and the institution would be entitled to two-thirds of licensing revenue from startups. Norway was one of five European countries to abolish the professor’s privilege in the 2000s, along with Denmark (2000), Germany (2002), Austria (2002), and Finland (2007).3 Italy (discussed later) is the only other country to have ever had professor’s privilege. After the reforms, Europe’s intellectual property (IP) system became broadly similar to that of the United States.

The primary finding of Hvide and Jones is that the end of professor’s privilege led to a 50 percent decline in the rate of university start-up formation. They also find negative effects on the quality of startups and patents associated with academic institutions (growth rates of startups are lower, and patents are cited less often). The decline in startups was most pronounced in the most technologically-intensive areas: before the reform, 27 percent of startups were in higher technology; after, 17 percent were. Among those with a science or engineering qualification employed by a university, rates of startup formation fell 63 percent.4

This effect is absolutely gigantic. One reason to suspect that it might be overstated is the phenomenon of a ‘sneakout’. This is where the same innovations occur, but academics hide them from the university bureaucracy, by registering patents under the name(s) of family members or cofounders, or otherwise having a legal structure that avoids any formal association with the institution. Sneakouts make it difficult to interpret any statistic about spinouts. But Hvide and Jones conclude, for various reasons, that there cannot possibly be enough sneakouts to explain more than a tiny fraction of the observed effect in Norway.5

Ironically, the purpose of the European IP law changes of the 2000s was to encourage more spinouts.6 The idea was that, if universities owned their patents, they would be incentivised to teach their employees about entrepreneurship, and to set up structures for patenting. But the policy had exactly the opposite of its intended effect.

The methodology the authors used to reach their conclusion is a bit technical, so I will leave details for those interested in a footnote.7 But in summary, this is about as clean a natural experiment as you ever get in social science. I am quite impressed by this paper. If you took these results at face value, it would imply that many European economies could effectively double the rate of innovation from universities by returning to the pre-2000s system of intellectual property law. That is not small potatoes.

We also have (weaker) evidence from other countries. In Germany, patenting and innovation declined after professor’s privilege was abolished. Italy adopted professor’s privilege for the first time in 2001. This seems to have had ambiguous effects, given that it took place at the same time as other sweeping reforms, which may well have cancelled out any effect. Professor’s privilege was finally abolished there in 2023, so it will be interesting to follow in the coming years whether Italy will see a similar decline to Norway in spillovers from academia. In professor’s privilege countries, the share of patents going to academics is or was relatively high (Sweden 6 percent, Finland 8 percent). In countries that didn’t have it, the academic patent share is or was relatively low (France 3 percent, Italy and the Netherlands 4 percent).

Sweden, recall, still has the professor’s privilege. In both Sweden and the United States, there is an ‘academic penalty’: PhDs in academia are less likely to start companies than non-academics with the same educational qualification. But in Sweden, the academic penalty is only half what it is in the United States.8

Ultimately, I don’t think that spinouts have ever been a large enough part of the economy to be a primary factor in explaining why European companies have slipped behind their American peers. But it does add nuance to the story. It certainly looks like much of Europe previously had a better system for dealing with academic IP. Now, we have inherited some of the failings of the Americans, but more so.

How spinouts work: the Bayh-Dole Act

No discussion of university tech policy would be complete without a mention of America’s Bayh-Dole Act. This has been extensively studied and debated, and is one of the most important examples to understand in all of innovation policy. Bayh-Dole has generally been well-regarded, although there are some dissenting voices.

Since this policy was introduced in 1980, patents filed by academics have been owned by the university – rather than the government – and the inventors have been entitled to about a third of the licensing revenues from spinouts. The remainder of the licensing fee is split equally between the inventor’s department and the university as a whole. (Whether the revenues go to a department, rather than the university it’s housed within, might sound trivial. But it would be interesting to see if this policy has succeeded in creating competition between departments to encourage more spinouts.) This split is on licensing fees, so it applies regardless of whether a new company is created, or if an existing company is paying to use the technology.

In addition to sharing licensing revenue, universities generally hold stock (equity) in their spinouts. In America, the share of the equity owned by the institution is typically low (in the 1–5 percent range), and lower at the more prestigious institutions. The fraction of the company owned by the university is much higher in Europe.

This is all getting quite complicated. What follows is a toy example, using realistic figures for two example universities: University College Dublin and Stanford University. The licensing fee and equity share can vary significantly between and sometimes within institutions, depending on the exact terms negotiated.

Stanford is widely regarded as the gold standard for spinout policy. You may have even heard of one or two Stanford spinouts, including Google, which began as a commercialisation of the PageRank algorithm.

Let’s suppose that an idea for a new drug has been discovered in a biology lab. A startup has been created to sell the product that is now worth $100 million, and a larger company now wants to acquire it. Let’s also imagine that investors originally put forward $50 million. Here is how the money generated from that exit would be distributed:

Stanford

• $3 million (3 percent) in licensing fees to the university. This is distributed equally:

  • $1 million to the Stanford biology department

  • $1 million to Stanford University as a whole

  • $1 million to the original inventor

• $2 million (2 percent) is the equity stake that goes to Stanford.

• $95 million remaining for the rest of the investors.

Rate of return for investors: 90 percent.9

University College Dublin

• $15 million (15 percent) goes to the university’s equity stake.

• $3 million (3 percent) in licensing fees. This is distributed on a sliding scale, starting at a 70 percent share for the original inventor, and getting closer to the American one-third share at higher revenues. Using rough numbers, that might look like:

  • $1.5 million to the original inventor

  • $1.5 million to the university

• $82 million for the remaining investors.

Rate of return for investors: 64 percent.

As you can see, the rate of return for investors on the same company is dramatically different, depending on the country the university is in. And Ireland is relatively ‘founder-friendly’ compared to most of Europe. In my example, the returns are quite high in either case. But in a narrower-margin investment, the difference between Stanford and UCD would be more apparent. One experienced venture capitalist I spoke with said their firm and most others they knew of considered startups associated with Irish academia to be “uninvestible” – among other reasons, because the fraction of equity owed to the university is so substantial. We will get back to this.

Interestingly, the UK is a major outlier among advanced economies in terms of spinouts. 34 percent of UK spinout deals involve the university owning over 30 percent of the company. In a not-insignificant number of cases, over half of the company is owned by the university. You can browse these spinout deals in a helpful open database. Thanks largely to a highly successful campaign by Air Street Capital, these equity shares have come crashing down in recent years. The UK government has now advised universities to implement their recommendations. Perhaps a similar effort can work elsewhere.

Share

The rise of technology transfer

As mentioned, the purpose of the European IP reforms of the 2000s was to encourage universities to invest in technologies and infrastructure to support patenting. As part of this, they established many more tech transfer offices (TTOs), and greatly expanded their role. TTOs are university departments tasked with supporting students and academics who want to commercialise their research. They help with patents, legal advice, and may also provide other services, including (say) introductions to investors.

Opinions vary on the usefulness of TTOs. Some are good. My sense is that the TTO within NovaUCD is well regarded. But TTOs are frequently criticised by founders for being slow, bureaucratic, and demanding unreasonable and/or confusing terms for spinouts. Anecdotally, one hears lots of horror stories. One person I spoke with tried to commercialise research he did while at Oxford University. The university initially asked for half of his equity, for a level of support he described as “nothing”. He negotiated them down to 3 percent of revenue, but for a low-margin tech business, this was still unviable. He eventually abandoned the project, and re-wrote all the code for his product from scratch, severing any formal link with the university.

This inspires two thoughts:

First, there may be adverse selection in the quality of founders that remain in academia. The best people might get fed up and leave the system entirely.

Second, spinout policy often seems more designed for deeptech and highly capital-intensive research, not (say) software. A biochemistry academic whose company benefited from free use of expensive and sophisticated lab equipment would probably agree that it was reasonable for her university to take a substantial chunk of the equity. A cynic would say that, if the equity shares are low, the Bayh-Dole system is a form of wealth redistribution away from the taxpayer – who funded the research to begin with – and toward private investors.

Hvide and Jones conducted interviews with Norwegian academics who had involvement with TTOs. While there were certainly some positive comments, the majority of the feedback was negative. Some representative quotes from their collection were:

Less attractive to work with entrepreneurship when you as an inventor only get a marginal portion of the ownership. The services that TTO provides does not justify their high portion of ownership.

I would never start up a company in the current system. In the current system, the TTO has a large ownership fraction and a dominating position from the start, and the entrepreneur has for example 33 percent. With venture financing, venture gets about 50 percent at every stage. It is common with 2–3 stages. Thus, the entrepreneur will have 16.6 percent after the first stage, 8 percent after two stages, and 4 percent after three stages . . . a low ownership share also means limited upside.

If you think I’m cherry-picking, you can see all of the comments in an online appendix.

You might think that this is just a quality issue with tech transfer offices. If TTOs have more or better staff, academics would be happier, and the relatively high equity shares to the university would be justified on the basis of the useful services they provided. But in Norway’s case, as time went on, the decline in entrepreneurship got even worse. While policies often get off to a rocky start, the evidence we have does not suggest TTOs have benefited from ‘learning by doing’.

Something that sticks out to me about Hvide and Jones’s survey data is that control doesn’t appear to be a primary issue. As part of the spinout terms, universities will often require that a representative from the institution sit on the board (usually as an observer; less often as a voting member who could fire the CEO). This involves giving up some control over the company, as did giving up the right to the patent after professor’s privilege was abolished. Yet, in their interviews, only one person mentioned the word ‘control’. Hvide and Jones identify that the primary frustration in Norway was simple: severely curtailing the possible financial upside makes it much less attractive to invent something or start a business.

Despite the stereotypes about not being very motivated by money, the main Hvide-Jones result shows that academics are highly sensitive to their income potential. It is probably also be that the academics most motivated by money are the ones most likely to start companies.

Cross-country evidence suggests that the supply of inventors is quite elastic with respect to the tax rates. My expectation is that, because Irish people are English-speaking and emigrate at the drop of a hat, Irish entrepreneurs move around significantly in response to taxes and policy. Naively, this makes spinout policy more important in Ireland than it would be in most places. If you mess it up, the relevant pool of talent will leave. This is related to why it is particularly important to encourage Irish startups in general. As Seán Keyes has written about in Progress Ireland’s founding essay and elsewhere, once you exclude the multinational sector, the value-add of Irish businesses is remarkably weak.

The trouble with measuring spinouts (and why they still matter)

Spinouts are particularly vulnerable to Goodhart’s law: ‘When a measure becomes a target, it ceases to become a good measure’.

Various countries have explicitly targeted raising the number of patents associated with research. Enterprise Ireland, to their credit, has recognised that this led to some perverse incentives, in which universities were encouraging academics to patent ideas, regardless of whether they intended to commercialise them or thought them to be useful.

The most natural metrics to focus on in spinout policy are the number of spinouts, and the valuation of those spinouts. But targeting an increase in the raw number of spinouts also seems fraught, not least because the ‘sneakout’ phenomenon discussed earlier makes it hard to measure how much entrepreneurship academics are really doing.

There are other ways that we could measure the success or failure of spinouts: the number of M&As, the number of companies to surpass certain revenue thresholds, or the quantity of outside investment. In the United States, the success of spinouts is sometimes implicitly evaluated on the basis of donations back to the home institution. But donating to one’s alma mater is a more niche pursuit in the rest of the world.

I remain confused as to why spinouts aren’t a bigger deal. In Hvide and Jones’s data, over the study time period (2000–2007), there were 48,844 startups, of which 128 were started by individuals at universities. That is just 0.3 percent. That is not to be entirely scoffed at. But in a previous blog post, I wrote about how, when basic research is considered as an investment, the standard estimates for the social annual rate of return on public R&D funding can be as high as 60–70 percent. The reason why these numbers can be so astronomically high is that research solves a market failure; since knowledge is the archetypal public good, firms systematically underinvest in knowledge production. We often emphasise ‘spillovers’ from research (another previous topic on the blog), but how is that compatible with there being so little supply or demand for companies forming in the research institutions themselves?

One good reason to care about spinouts, despite their small number, is that it’s a smart way to do regional development. The strategy taken by policy measures like the National Planning Framework has been to de facto limit the growth of Dublin, which we have elsewhere argued is a deeply mistaken approach. But top-tier research often takes place in universities that are not otherwise notably economically dynamic or productive. The National Development Plan profiles HookeBio, a microfluidics spinout from UL in Shannon, Co. Clare.

In a classic essay by Paul Graham, he writes about how Philadelphia came to be a biotech hub, with many of the world’s most important startups for cell- and gene-therapy, despite lacking (he says) a pre-existing entrepreneurial culture. Philadelphia has world-class universities, and a favourable set of commercialisation policies; it turns out that being a ‘hip’ place where the most talented young people wanted to move wasn’t necessary.

The entrepreneurial culture of Silicon Valley is not something you can take off the shelf. But the policies of a university which has not put up unnecessary obstacles to commercialisation are something you can, essentially, copy.

Where does Ireland stand?

When it comes to spinout policy, Ireland could have done a lot worse. We have never had professor’s privilege. The way our system works is that each higher education institution sets terms consistent with the National IP Protocol, published by Knowledge Transfer Ireland, an office within Enterprise Ireland. Irish universities will typically ask for 10–15 percent (dilutable) equity, in addition to a tiered revenue-sharing agreement for the licensing revenues, with a negotiable rate.

Ireland produces between 25 and 30 spinouts per year. Last I checked, there were 163 currently active spinouts in Ireland. To me, that seems low, but it is still higher than the UK, the US, and the EU in per capita terms.10 In some ways, this is an Irish success story, but absent detailed data on how much these companies are actually worth, it’s unclear how important this figure is. Sneakouts, and the possible effects of grants and subsidies sustaining businesses that would not survive on their own, mean that ‘spinouts per capita’, even if we could measure it accurately, would have a questionable relationship with human welfare.11

If you divide the figures above by the ~€1 billion public research spending, the government needs to spend ~€33 million of public money on research to create one additional company on average. You may or may not think that is a good rate of return. If research funding is subject to diminishing returns, then we should expect countries like Ireland that are outliers in how little they fund research to begin with, to perform better on this ‘cost per spinout’ measure.12

Although it came to be mocked because of its association with Reaganomics, Arthur Laffer’s observation that total revenue raised follows an inverse-U relationship with the rate of tax is entirely correct. In theory, there is a Laffer curve with respect to the fraction of spinout companies that universities intend to own.

My conjecture is that we are currently on the wrong side of this: Irish universities would make more money to fund their activities by asking for less equity, because the less onerous terms would encourage enough additional entrepreneurship to compensate. Top American colleges took the bet that with spinouts, it is better to own a small percentage of a bigger pie than a large percentage of a small pie. It seems to have worked out for them.

If Irish equity stakes are too high, what about the rate of the licensing fee? 3 percent of net sales is a common ballpark for Irish and American universities. It seems difficult to drive that number much lower, but it could do with being a lot more flexible. 3 percent of net sales would still be very steep for a low-margin software business.

There is no consensus about whether we should adopt the professor’s privilege wholesale. Many of us would welcome a switch to professor’s privilege in a place like Ireland; personally, I am convinced that there has been too little policy experimentation by smaller countries. The end of professor’s privilege substantially homogenised global spinout policy. Are we so sure that the system we arrived on is the best one?

Whenever Ireland ranks as ‘the best’ in the EU along some measure related to startups, my friends in tech sometimes say that we’re a giant among the pygmies. Things could always be better. Like many of Ireland’s policy problems, spinouts are afflicted by the malady of “restrict supply, subsidise demand”. We restrict supply with onerous equity and revenue-sharing – not to mention our expensive legal system – making it difficult to commercialise research. Simultaneously, we subsidise “innovation hubs”, coworking spaces, small business grants, startup accelerators, and other initiatives to encourage students and academics to start companies. The result? Expensive bloat and minimal returns.

Sam Enright is Innovation Policy Lead at Progress Ireland, and editor-in-chief of The Fitzwilliam. You can email him at sam@progressireland.org. He also runs a bounty system to find and catalogue mistakes he has made.