Paul Walker

Dr Paul Walker is an economist at University of Canterbury. He has expertise in microeconomics, institutional economics and industrial Organization. He blogs for The Dismal Science.

Is competition policy outdated? - The Dismal Science

May 03, 2015

As the economy changes so should competition policy. But does it? The "new economy" or the "information economy" or the "knowledge economy" or whatever you want to call it has altered the way the economy works. Changes in technology, in particular information and communication technology (ICT), have become the major drivers of change in the economy and of economic growth. But has competition policy kept up with this change? May be not.

The European Commission’s antitrust case against Google is the latest in a series of attempts to prevent tech giants from "monopolising" EU markets, or so we are told. But it can be argued that past cases against Intel and Microsoft demonstrate the need review what may be an outdated competition policy in the EU. And not only in the EU.

The European Commission has formally charged Google with anti-competitive practices, the latest twist in a case that was first launched way back in 2010. The EU’s competition watchdog accuses the US tech giant of systematically favouring results from its own specialist search engine, Google Shopping, over competitors like Amazon and eBay. And this, it is alleged, has had an adverse impact on competition and consumer well-being.

Diego Zuluaga has been looking at the case and argues, in an article at EurActiv.com, that,

Superficially, it would indeed seem that Google holds a dominant position online, with a 92 per cent share of the EU market for general (known as ‘organic’ or ‘horizontal’) search. Google has also been expanding its offering of specialist (a.k.a. ‘vertical’) search engines for items like flights and consumer products. Does this mean that competition online is being undermined, and that regulatory authorities should intervene to put it right?
It could of course be that they are dominant simply because they are better than the competition.

Zuluaga goes on to say that the answer to his question is, Not necessarily.
The test for any antitrust investigation must be whether competition, not individual competitors, are being harmed. And by any available measure, competition and specialisation in online search services is thriving. New players focusing on specific market niches, from SkyScanner for flights to DuckDuckGo for greater privacy, have emerged in recent years. In the specific case of comparison shopping which the Commission is worried about, it does not look like Google Shopping is catching on, despite the tech giant’s best efforts: In three key EU markets – Germany, France and the UK – Google’s own product search engine is a marginal player, with Amazon, eBay and local competitors (Idealo in Germany, Fnac in France) boasting multiple times the number of user visits of Google Shopping. What is more, the gap between Google’s own service and its leading competitors is growing, if anything.

Intuitively, this makes sense. If I want to purchase Malcolm Gladwell’s latest bestseller, I am much more likely to browse for it on Amazon, as the latter is reputed for its excellent catalogue, user reviews and related recommendations. Rather than search for it on Google and then look for the best result, I will go to the Amazon website directly. The same goes for flights, hotels, restaurants and any other topic where there is a wealth of specialist search services. Even for those who tend to go through Google, competing options are still there – one only need scroll down to see them. This makes it hard for Google to divert large amounts of traffic to its own services – and it helps explain why Google Shopping has not taken off, as we might have expected it to.
A question one could ask is, Have past actions by competition authorities in previous digital cases been appropriate? May be not, just think of the Intel and Microsoft cases.
In 2009, the Commission fined chip-maker Intel more than €1bn for offering ‘predatory discounts’ to computer manufacturers, which allegedly harmed Intel’s competitors. Yet, by all measures, competition in the chip sector was fierce during the period of Intel’s anti-competitive behaviour. Chip prices declined by up to 75 per cent, while performance grew tenfold. Far from increasing at the expense of competitors, Intel’s market share remained stuck at 80 per cent, and the fluctuations in its share are strongly correlated with new product launches, both by itself and by rivals like AMD.

How about the other previous high profile digital probe, the 2004 ruling against Microsoft? Commission officials worried at the time that the company founded by Bill Gates was strengthening its grip on all PC-related products and services, thanks to its dominance of computer software. Barely a decade later, it is astonishing how things have changed: Microsoft still provides software for a lot of the world’s PCs, but the rise of smartphones – where Google’s Android and Apple’s iOS prevail – has made its share of the overall software market (for smartphones, tablets as well as PCs) shrink to as low as 20 per cent, according to Goldman Sachs research from 2012. Innovation killed the software star.

Both the Intel and Microsoft cases illustrate the shortcomings of EU competition policy when it comes to the digital sector: Despite Intel’s large market share in the chip market, competition was no less aggressive, and consumers still benefited from steadily dropping prices and ever better performance. And even though Windows was the dominant player in software in 2004, innovation outside the PC market – which no one, least of all Microsoft, foresaw – has turned it into one among several competitors in a much larger market.
Aa obvious point is that competition policy rulings should be grounded in sound economic analysis. It is not clear that they have been.
DG Competition has enormous powers to intervene in the internal market, acting as judge, jury and enforcer of antitrust proceedings in the EU. This makes it imperative that its rulings be rooted in sound, convincing economic analysis. Such analysis seems to be lacking in the case of Google Shopping, as it was in the Intel and Microsoft rulings. The digital economy lies at the heart of economic growth in the 21st century, so getting antitrust wrong in this sector will have a longstanding negative impact on innovation and growth in Europe.
A lesson from this is that the new technology underlying the new economy has changed the way companies do business in all sorts of sectors, and competition policy, in all countries, must evolve with it. Is not clear that it has in many countries, including New Zealand.

Should cities pay for sports facilities? - The Dismal Science

May 01, 2015

Over at the Offsetting Behaviour blog Eric Crampton comments on how bad the business plan for a new $156 million cycleway for Christchurch is. But here in the People's Republic of Christchurch there are plans afoot not just for cycleways but also for other big ticket items such as a big new, super-sized stadium. Sports stadiums are a favourite of cities all over the world, not just in New Zealand. And you can be sure that they all come with (bad) business plans or economic assessment reports to show how great they are.

As Adam M. Zaretsky says when writing about the building of stadiums in the U. S.,

[...] cities with home teams are often willing to go to great lengths to ensure they stay home. And cities without home teams are often willing to dangle many carrots to entice teams to move. In either case, the most visible way cities do this is by building new stadiums and arenas.
But should cities pay for sports facilities? This is the question Zaretsky looks at in an article in the The Regional Economist, a publication of the Federal Reserve Bank of St. Louis,.

Zaretsky writes,
Between 1987 and 1999, 55 stadiums and arenas were refurbished or built in the United States at a cost of more than $8.7 billion. This figure, however, includes only the direct costs involved in the construction or refurbishment of the facilities, not the indirect costs—such as money cities might spend on improving or adding to the infrastructure needed to support the facilities. Of the $8.7 billion in direct costs, about 57 percent—around $5 billion—was financed with taxpayer money. Since 1999, other stadiums have been constructed or are in the pipeline [...], much of the cost of which will also be supported with tax dollars. Between $14 billion and $16 billion is expected to be spent on these post-'99 stadiums and arenas, with somewhere between $9 billion and $11 billion of this amount coming from public coffers. The use of public funds to lure or keep teams begs several questions, the foremost of which is, "Are these good investments for cities?"
The short answer to this question is "No." If you are a 'just read the executive summary' kind of guy then you can stop reading here since what follows just makes the case for the answer already given.

Zaretsky goes on to say,
When studying this issue, almost all economists and development specialists (at least those who work independently and not for a chamber of commerce or similar organization) conclude that the rate of return a city or metropolitan area receives for its investment is generally below that of alternative projects. In addition, evidence suggests that cities and metro areas that have invested heavily in sports stadiums and arenas have, on average, experienced slower income growth than those that have not.
When stadiums are build with ratepayers money, Are they worth it?
The dollars being invested in sports facilities are quite substantial considering the overall contribution the industry makes to the economy. In testimony before the U.S. Congress, economist Robert Baade said that Chicago's professional sports industry—which includes five teams—accounted for less than one-tenth of 1 percent of Chicago's 1995 personal income. Baade further commented that even when compared with the revenue of other industries, professional sports teams contribute small amounts to the economy. He noted, for example, that "the sales revenue of Fruit of the Loom exceed[ed] that for all of Major League Baseball (MLB), while the sales revenue of Sears [was] about thirty times larger than that of all MLB revenues."

Still, cities are driven by the idea that playing host to professional sports teams builds civic pride and increases local tax receipts from the team-related sales and salaries. When it comes to salaries, however, economist Mark Rosentraub noted in a 1997 article that there is no U.S. county where professional sports accounts for more than 1 percent of the county's private-sector payroll.

Although sports facilities certainly generate tax revenues from their sales, the pertinent question is whether these revenues are above and beyond what would have occurred in the region anyway. To address this question, city proposals to use taxpayer money to finance sports facilities are routinely accompanied by "economic impact studies." These studies, which are often commissioned by franchise owners and conducted by an accounting firm or local chamber of commerce, generally use spurious economic techniques to demonstrate the number of new jobs and additional tax revenues that will be generated by the project. The assumptions that are made in these studies—such as how much of the newly generated income will stay in the region and how many "secondary" jobs will be created—often cannot be substantiated by economic theory.

Estimates of income that will be generated and, hence, spent in the region are often overstated. Most of the "big" money in sports goes to the owners and players, who may or may not spend the money in the hometown since many live in other cities. And because athletic careers are usually short-lived, much of the players' income is invested. Moreover, league rules often require ticket revenues be shared with franchise owners in other cities as a way to subsidize teams in smaller markets. In the case of the National Football League, every visiting team leaves town with 34 percent of the gate receipts from each game.

On top of all this, the value of the subsidy a team receives when a city foots the bill for a new stadium or arena often shows up as a higher team resale price, which then ends up in the owner's pocket. For example, Eli Jacobs bought the Baltimore Orioles for $70 million in 1989, just after the team had convinced the state of Maryland to build it a new $200 million ballpark from lottery revenues. The enormously popular Oriole Park at Camden Yards opened in 1992. The following year, Jacobs sold the Orioles for $173 million. The sale netted Jacobs an almost 150 percent return, with no money out-of-pocket for the new ballpark.
But what about the economic impact studies that show tax revenue increases from a stadium.
Economic impact studies also tend to focus on the increased tax revenues cities expect to receive in return for their investments. The studies, however, often gloss over, or outright ignore, that these facilities usually do not bring new revenues into a city or metropolitan area. Instead, the revenues raised are usually just substitutes for those that would have been raised by other activities. Any student of economics knows that households have budget constraints that are binding, which means that families have only so much money to spend, particularly on entertainment. If the family chooses to spend the money at the ballpark, for example, then those funds cannot be spent on other activities. Thus, no new revenues are actually being generated.

Public funds used for a stadium or arena can generate new revenues for a city only if one of the following situations occurs: 1) the funds generate new spending by people from outside the area who otherwise would not have come to town; 2) the funds cause area residents to spend money locally that would not have been spent there otherwise; or 3) the funds keep turning over locally, thereby "creating" new spending.

Very little evidence exists to suggest that sporting events are better at attracting tourism dollars to a city than other activities. More often than not, tourists who attend a baseball or hockey game, for example, are in town on business or are visiting family and would have spent the money on another activity if the sports outlet were not available.

Economists Roger Noll and Andrew Zimbalist have examined the issue in depth and argued that, as a general rule, sports facilities attract neither tourists nor new industry. A good example, once again, is Oriole Park at Camden Yards. This ballpark is probably the most successful at attracting outsiders since it is only 40 miles from the nation's capital, where there is no major league baseball team. About a third of the crowd at every game comes from outside the Baltimore area. Noll and Zimbalist point out that, "Even so, the net gain to Baltimore's economy in terms of new jobs and incremental tax revenues is only about $3 million a year—not much of a return on a $200 million investment."

The claim that sporting facilities cause residents to spend more money in town than they would otherwise is harder to substantiate. To prove such a claim, the agency performing the analysis would need for its report both detailed information about the spending patterns of households and the ability to ferret out the information about their spending in other regions, which, at best, is extremely difficult and may even be impossible. Without such information, the report's authors could back into this claim only with some fancy footwork and shaky assertions. That is, they would have to contend that residents are spending more in town because of higher incomes that enable households to devote more of their entertainment budgets toward local sporting events. Then, the authors would have to demonstrate that incomes are up because money was spent on the stadium. If they can't, the argument falls apart since the only conclusion is that incomes rose for unrelated reasons; throwing tax dollars at the stadium did not affect households' spending patterns.
The most spurious economic concept applied to stadiums is the "multiplier".
Of the three circumstances described that purportedly generate new revenues, the third—funds keep turning over locally, thereby "creating" new spending—is probably the most spurious from an economist's viewpoint. Such a claim relies on what are called multipliers. Multipliers are factors that are used as a way of predicting the "total" effect the creation of an additional job or the spending of an additional dollar will have on a community's economy. It works something like this: A stadium is built, which creates new jobs in the region. Because more people are working, they spend money in the area (for lunch, parking, etc.), which in turn requires local businesses to hire additional workers to support the increased demand. These extra workers further increase demand for goods and services in the area, requiring more new jobs...and so on. That is, the dollars keep turning over locally. The story is the same for fans spending money at the arena, which provides income for arena workers, who then spend the money, generating incomes for other workers...and so on.

On their faces, these are compelling arguments. Some researchers have even attempted to quantify these effects, developing precise multipliers that tell analysts how much the new spending or job creation should be "multiplied" by to arrive at the "total effect" the spending or job creation will have on the local economy. These multipliers are often specific enough to distinguish between various industries, occupations and locations. Thus, economic development specialists and planners will generally latch onto multipliers and confidently proclaim that the 1,000 new jobs created by this industry will actually create 4,355 new jobs and generate $5.5 million in new revenue in the community when all is said and done. Makes for great headlines, but are such outcomes believable?

Probably not. As Mark Twain once said: "It's not what we don't know that hurts. It's what we know that just ain't true." For one thing, these new jobs most likely just lure workers away from other jobs in town and do not actually lead to a net change in jobs in the area. For another, many of the jobs are low-paying, part-time and needed only on game days. Moreover, authors of these economic impact studies often choose multipliers arbitrarily or with clients' wishes in mind to get the desired outcome. As economist William Hunter has pointed out, multiplier analysis can be used to justify any public works project because "even the smallest multiplier will guarantee community income growth in excess of public expenditures."

Even if economic impact studies are taken at face value, however, the cost of creating these jobs is usually out of the ballpark. In Cincinnati, for example, when two new stadiums were proposed to keep the NFL Bengals and the MLB Reds in town, the economic impact study claimed that 7,645 jobs would be created or saved because of the stadium investment. Since the project was estimated at $520 million, each new or saved job was reported to cost about $68,000.

When economists John Blair and David Swindell examined the $68,000 figure a bit closer, though, they discovered it was too low because the study's estimate of 7,645 new or saved jobs was too high. Blair and Swindell then re-evaluated the report, corrected for double-counting and other problems, and concluded that only 3,530 jobs would be created or saved if the stadium proposal passed. Thus, the cost per job was actually going to run more than $147,000. In contrast, state economic development programs spend about $6,250 per job to create new jobs.
And let us not forget "opportunity cost" .... but most impact studies do in fact forget it.
Another glaring omission from these economic impact studies is the value of the next-best investment alternative—what economists call the opportunity cost. "There's no such thing as a free lunch" is a favorite economist expression because it sums up exactly what opportunity cost means: When making a choice, something always has to be given up. The value of the "losing" choice must be considered when making the decision and when calculating the value, or return, of the "winning" choice. In other words, when a city chooses to use taxpayer dollars to finance a sports stadium, the city's leaders must consider not only what the alternative uses of those funds could be—such as schools, police, roads, etc.—but they must also figure what return the city would receive from these other ventures. Then, the return from the city's next-best alternative (for example, schools) must be subtracted from the total return of the "winning" choice to arrive at the "actual" return of the stadium investment. This adjusted calculation, though, is almost always missing from sports stadium impact studies. Why? Because in just about every case, the adjusted calculation would show that the next-best alternative was actually the better alternative.

Has financing sports stadiums ever been the best alternative? Research shows "No." In their book, Noll and Zimbalist—along with 15 other collaborators—examined the local economic development argument from a wide variety of angles. In every case, the conclusions were the same. "A new sports facility had an extremely small (perhaps even negative) effect on overall economic activity and employment. No recent facility appears to have earned anything approaching a reasonable rate of return on investment. No recent facility has been self-financing in terms of its impact on net tax revenues. Regardless of whether the unit of analysis is a local neighborhood, a city, or an entire metropolitan area, the economic benefits of sports facilities are de minimus.

In fact, research has shown that subsidizing sports facilities usually does not affect a city's growth and, in some cases, may even hurt growth since funds are being diverted from alternatives with higher returns. In a 1994 study that examined economic growth over a 30-year period in 48 metropolitan areas, Robert Baade found that of the 32 metro areas that had a change in the number of sports teams, only two showed a significant relationship between the presence of a sports team and real per-capita personal income growth. These cities were Indianapolis, which saw a positive relationship, and Baltimore, which had a negative relationship.

Moreover, Baade found that of the 30 metro areas where the stadium or arena was built or refurbished in the previous 10 years, only three areas showed a significant relationship between the presence of a stadium and real per-capita personal income growth. And in all three cases—St. Louis, San Francisco/Oakland and Washington, D.C.—the relationship was negative.
So what is the overall conclusion? The weight of economic evidence shows us that ratepayers end up spending a lot of money and ultimately don't get much back for their forced investment. And when this paltry return is compared with other potential uses of the funds, the investment, almost always, seems uneconomic.

Interestingly, a pamphlet from The Greens turned up at home last week arguing that Christchurch should reassess its spending on "big ticket" items, such as the super-sized stadium. A good proposal as far as it goes. What would be better is to say we should abandon the stadium idea (and a number of other projects) altogether.

Real wage inequality - The Dismal Science

May 01, 2015

Inequality is the trendy topic of the moment. Wage inequality being part of the story. We hear about differences in wage between males and females, skilled and unskilled workers, different racial groups, age groups etc. One point to keep in mind when l...

Does vertical integration decrease prices? - The Dismal Science

May 01, 2015

The double marginalization problem is a classic problem with applications in industrial organization, innovation policy and development. The problem is what happens to social welfare, prices, and profits when one monopoly sells to another monopoly. Below is a video discussion of the double marginalization problem by Alex Tabarrok of George Mason University from MRUniveristy,


The double marginalization problem makes the (monopoly) producers worse off as well as making consumers worse off.

Now as noted in the video combining the seller monopoly with the buyer monopoly into one will, in theory, make both the producers and consumers better off. But does it do so in practise? At least as far as consumers are concerned.

This is where a new paper Does Vertical Integration Decrease Prices? Evidence from the Paramount Antitrust Case of 1948 by Ricard Gil in the American Economic Journal: Economic Policy, 2015, 7(2): 162–191, helps us. Gil finds, utilising data from the Paramount antitrust case of 1948, that vertical integration, i.e. combining the two monopolies, does in fact decrease prices, and thus increases consumer welfare, in a manner consistent with the elimination of the double marginalization problem.

The paper's abstract reads:

I empirically examine the impact of the 1948 Paramount antitrust case on ticket prices using a unique dataset collected from Variety magazine issues between 1945 and 1955. With information on prices, revenues, and theater ownership for an unbalanced panel of 393 theaters in 26 cities, I find that vertically integrated theaters charged lower prices and sold more admission tickets than nonintegrated theaters. I also find that the rate at which prices increased in theaters was slower while integrated than after vertical divestiture. These findings together with institutional details are consistent with the prediction that vertical integration lowers prices through the elimination of double marginalization.

Britain’s housing crisis - The Dismal Science

May 01, 2015

This short video comes from The Economist magazine:Rocketing prices and limited construction are putting home ownership out of reach for many young Britons. As the election nears, few politicians have a realistic solution.Obviously the particulars in t...

To eat the rich, first they must stay still - The Dismal Science

Apr 30, 2015

Taxation is always a problematic issue but the taxation of very high income earners is becoming an even more controversial subject in a number of countries. One problem with high tax rates is that it could lead high earners to move abroad. A new column at VoxEU.org suggests that top-tier inventors are significantly affected by top tax rates when deciding where to live. It is argued that the loss of such highly skilled agents could entail significant economic costs in terms of lost tax revenues and less overall innovation.

There is much debate, but little evidence, about the effects of high tax rates on high earners. The anti-tax side argue that higher top tax rates will cause an exodus of valuable, high income and highly skilled economic agents. They claim that high tax rates will unavoidably lead to a brain drain and an exodus of the most qualified people, especially as barriers to labour mobility between developed countries are reduced. The pro-tax side maintain that migration decisions are driven by other (possibly non-economic) considerations and would not respond very much to higher taxes.

It is generally acknowledged that non-human capital is highly mobile in a globalised world. This fact is used to justify lower taxation on capital. Much less is known about the mobility of human capital in response to taxation.

In their article, The effects of top tax rates on superstar inventors, Ufuk Akcigit, Salome Baslandze and Stefanie Stantcheva argue that inventors are highly valuable economic agents as creators of innovations and potential drivers of technological progress.

A group of highly valuable economic agents that policymakers perhaps might worry about is inventors, the creators of innovations and potential drivers of technological progress. Inventors may well be important factors for a country’s development and competitiveness – highly skilled migration has been shown to be both beneficial for a receiving country’s economy and to disproportionately contribute to innovation [...].

Consider Alexander G Bell, the inventor of the telephone; James L Kraft, who patented a pasteurisation technique and founded Kraft Foods; Ralph Baer, the inventor of the first home video gaming console that contributed to the expansion of the video gaming industry; or Charles Simonyi, a successful product developer at Microsoft. In addition to being very prolific inventors, they had something else in common: they were all immigrants. This is not very surprising given that migration rates increase in skill [...] and inventors are ranked very highly in the skill distribution.
It’s true, however, that inventors vary vastly in their quality and innovativeness. The big question is to do with the behaviour of the 'superstars': Do 'superstar' inventors respond to top tax rates?
In recent research (Akcigit, Baslandze, and Stantcheva 2015) we study the international migration responses of superstar inventors to top income tax rates for the period 1977-2003 using data from the European and US Patent offices, as well as from the Patent Cooperation Treaty [...]. Our focus is on migration across eight technologically advanced economies: Canada, France, Germany, Great Britain, Italy, Japan, Switzerland, and the US. To abstract from capital and corporate taxes as much as possible, we restrict our attention to inventors who are company employees and are not the owners (‘assignees’) of their patents.

Superstar inventors are those in the top 1% of the distribution of citations-weighted patents in a given year and ‘stars’ are inventors who are just below superstars in terms of quality and are in the top 1-5% of the citations-weighted patent distribution.

From outside survey evidence, we know that superstar inventors are highly likely to be in the top tax bracket and, hence, directly subject to top tax rates. Stars or inventors of lower quality are much less likely to be in the top bracket. The top tax rate, which can also be viewed as a ‘success tax’ can also have either an indirect motivating or discouraging effect on inventors in general, even on those who are not yet in the top bracket.

There has is a strong and significant correlation between top tax rates and those inventors who remain in their home countries. The relation is strongest for superstar inventors. [Results] show that superstar inventors are highly sensitive to top tax rates. The elasticities imply that for a ten percentage point reduction of top tax rates from 50% to 40%, a country would be able to retain on average 3.3% more of its top 1% superstar inventors. This relation weakens as one moves down the quality distribution of inventors – the top 25-50% or the bottom 50% of inventors are no longer sensitive to top tax rates.

[...]

At the individual inventor level, we have developed a detailed model for location choice. This wasn’t easy for two reasons. First, location choices are clearly also driven by factors other than taxes – such as language, distance to one’s home country, and career concerns – for which we include controls. Second, inventors may earn different pre-tax wages in different countries. This is a counterfactual we cannot observe and have to control for through a detailed set of proxy measures.

The results highlight that superstar top 1% inventors are significantly affected by top tax rates when deciding where to live. For instance, our results suggest that, given a ten percentage point decrease in top tax rates, the average country would be able to retain 1% more domestic superstar inventors and attract 38% more foreign superstar inventors.

[...]

We also consider long-term mobility, defined as a one-way move abroad. It turns out that long-term mobility is still affected by taxes, but to a lesser extent. This seems to imply that there are some adjustment costs to moving that might prevent inventors from moving back once they leave due to higher taxes.
What is the influence of companies on inventors’ migration responses to taxes?
One would expect companies to have an important influence on the inventor’s decision to move abroad. For instance, working for a multinational company might facilitate an international move, both directly within the company and indirectly by providing international exposure. Depending on the bargaining power between employer and employee, the relocation decision might well be driven by the former rather than by the latter. In that case, and if the employer has other considerations than personal income tax, we would observe a dampened migration effect of taxes in the data. Note, nevertheless, that employers should take personal income taxes into account to some extent, if competition for superstar inventors forces them to pay higher wages as a compensation for higher taxes.

We find that inventors who have worked for a multinational company are more sensitive to tax differentials in their location choice. On the other hand, inventors whose company has a significant share of its innovative activity in a given country are less sensitive to the tax rate in that country. This seems to suggest that career concerns can outweigh tax considerations. It could also signal that companies with very geographically localised research and development activities will strongly prefer to keep their superstar inventors at the main research location and dissuade them from moving to lower tax countries.
The upshot of all of this is that labour, like capital, might be internationally mobile and respond to tax incentives. The loss of highly skilled agents such as inventors might entail significant economic costs, not just in terms of tax revenues lost but also in terms of reduced positive spillovers from inventors and, ultimately, less innovation in a country.

Ref.:
  • Akcigit, U, S Baslandze and S Stantcheva (2015), “Taxation and the International Mobility of Inventors”, Working Paper 21024, National Bureau of Economic Research.