Is Amazon’s takeover of Whole Foods anti-competitive? Probably not.

This originally appeared as a guest post on the Adam Smith Institute’s blog last week.

A few days ago, Amazon announced its plans to purchase the predominantly USA-based grocery retail chain Whole Foods for almost $14bn. Although both companies operate in many countries, the main competition issues (if any) are likely to arise in the US, were both companies have a non-negligible presence.

Indeed, this announcement has resulted in a number of people claiming that the proposed merger will be anti-competitive. Specifically, there are some claims that the merger would result in 1) bundling and foreclosure of rivals; and/or 2) predatory pricing. In short, the first theory of harm posits that Amazon would force customers that wanted to purchase its distribution (or other) services to also purchase from Whole Foods (or vice versa), while the second theory of harm suggests that the merged entity would price below cost in order to drive out rival grocery firms before increasing prices once those rivals exited.

Importantly, both of these theories of harm require that the merged entity have some form of “market power” (i.e. the ability to charge a price above the competitive level and to act independently of its rivals). Typically, this is most likely to occur when a firm has a share of sales in a particular market of over 40%. However, as a general point, these theories of harm gloss over the fact that Amazon and Whole Foods’ shares in grocery sales are tiny – less than 5% combined in the US. As such, it is difficult to see how the combined entity can have any market power.  Clearly, the merged entity would not satisfy this for sales of groceries at the moment of the merger.


However, others might argue that Amazon does have a sufficiently high share of sales of “online retail” to be classed as dominant. As such, they argue that Amazon could “leverage” its power in that area to grocery retail by bundling some of its services with those of its groceries. However, as the merged entity will be active at the retail level of groceries, it is not obvious exactly what other services offered by Amazon could be bundled with them – for the bundling strategy to work, consumers would still have to want at least one of the items in the bundle, and could continue to purchase them separately from Amazon or elsewhere anyway. Hence, there does not appear to be a viable mechanism through which this bundling theory of harm could occur.

Predatory Pricing

Moreover, for the predatory pricing theory of harm to be valid, there must be strong evidence that 1) the merged entity would price its groceries below some measure of cost that represents the extra cost that would be incurred by supplying one extra unit of output (usually measured as average variable cost of long-run average incremental cost); and 2) it would have an incentive to do so.

The first condition is notoriously difficult to prove – one first has to decide which costs should be included / excluded in the measure (which really isn’t as easy as one would think – e.g. should advertising spend that applies to brand-related marketing, but isn’t specifically related to groceries, be included), as well as deciding the relevant time-frame over which costs are assessed.

The second condition requires proving that the merged entity would become dominant (and therefore be able to recoup the losses it had made in pricing below cost) in the future. This is where the theory of harm becomes incredibly speculative – it assumes that sufficient sales would switch to the merged entity from rival grocery firms that the merged entity would be dominant. In other words, it assumes that pricing below cost would be sufficient in and of itself to persuade consumers to switch (regardless of e.g. quality of service provided) and that rival grocery firms would not respond in any way to the merged entity’s actions. Clearly both of these assumptions are likely to be violated in practice and, as such, the predatory pricing theory of harm seems unlikely.


Given that the merged entity is unlikely to have the incentive or ability either to bundle its products together or recoup any losses made from pricing below costs, both of the theories of harm currently being bandied about are unlikely to be valid. As such, it is difficult to see how the cries that the proposed merger is anti-competitive are anything more than “a big firm is buying someone so they have to be stopped”. That should not be a basis on which a merger can be prevented.


Fact-checking a few claims about the NHS

What with the campaigning for the general election having gotten into full swing last week, many claims have been made regarding which Party would be better for which aspect of security, the economy, education etc. One particular video regarding the NHS started doing the rounds on Facebook a few days ago. This video makes a number of claims regarding the supposed impact that the recent Coalition and Conservative governments have had on the NHS, with the video then going on to suggest that a Conservative government would be bad for the NHS. For a bit of excitement, here is said video:



The claims made in that video are many. Some are valid, whereas others are not. Let’s take each of them in turn.

Claim 1: We are experiencing the largest sustained drop in NHS funding as a percentage of GDP since the NHS was founded.

Reality: This claim is false. As per the information shown in the graph below (from the Institute for Fiscal Studies) NHS spending as a proportion of GDP has been stable over the past couple of years, and the decrease between 2009 and 2012 was no larger or longer than decreases in the mid-to-late 1970s or mid-1990s.


Moreover, the more relevant metric of NHS spending per capita continues to increase – in other words, more is spent per person on the NHS than ever before, although the rate of that increase has slowed in recent years.


Claim 2: If the internal market was abolished we [i.e. the NHS] could save billions.

Reality: This claim is also false. The internal market actually creates savings and is not “wasteful” as is claimed in the video. On the contrary, it promotes competition and stimulates the NHS to provide better services – importantly, the benefits of competition in healthcare are well established. Furthermore, it is actually the refusal of many within the NHS to accept the proven benefits of competition that is causing some harm to the NHS – indeed one of NHS Improvement’s main aims is to promote and encourage “buy-in” of competition among those in the NHS. Hence, abolishing the internal market would actually cost billions rather than save them.

Claim 3: Health tourism costs the NHS £200 million per year, which is insignificant in terms of the overall cost of the NHS.

Reality: This is generally true – although the costs to the NHS associated with people who are not ordinarily resident in the UK are of the order of £2 billion per year, that includes many people who did not come to the UK specifically and solely to use the NHS (i.e. it includes people who are not “health tourists”. Instead, estimates put the upper bound of the costs associated with those who travel to the UK for the sole purpose of using the NHS at around £300 million per year. When compared to the total annual NHS budget of about £90 billion, the costs associated with health tourism are indeed a trivial amount.

Claim 4: Immigrants are not ruining the NHS, they’re running the NHS.

Reality: True. Immigrants from within the EU currently represent about 10% of doctors and 4% of nurses. If non-EU immigrants are included, therefore, the figures are likely to be slightly (although probably not a huge amount) higher. Given that there are already quite severe labour shortages within the NHS, it is clear that without the immigrants currently working within the NHS, the functioning of the NHS would be severely hampered. Moreover, immigrants are net contributors in terms of taxes vs benefits, so also contribute to the NHS in that way. Hence, the claim that immigrants are not ruining the NHS is clearly valid.

Claim 5: 1 in 10 nursing posts are vacant and the nursing bursary has been scrapped

Reality: True. The nursing bursary was indeed scrapped at the start of the year – this means that there is a much-reduced incentive for people to train to become nurses as they will now have to pay £9,000 in tuition fees per year in order to do so. This is likely to lead to problems recruiting sufficient nurses in future. Notwithstanding that, there are also problems recruiting nurses now – the Royal College of Nursing suggests that 1 in 9 nursing posts are now vacant. This figure is actually marginally worse than that claimed (11% vacancy rate vs the 10% claimed).

Claim 6: Tens of thousands of sick patients waited on A&E trolleys this past winter

Reality: Likely to be true. Using data from Quality Watch (and a bit of approximation / extrapolation), roughly 6 million people attended A&E last winter. Of these, around 15% were not seen within the government target of four hours – i.e. about 900,000 people waited more than four hours in A&E. Now, it seems unlikely that all of these people waited on trolleys specifically, but even if only 10% of these people (i.e. 1.5% of all admittances to A&E) did then the “tens of thousands” figure would be accurate. Hence, this claim seems plausible.

Conclusion: As with most of these election video type things, the video contains some claims that are true, some that are likely to be true, and some that are demonstrably false. Does this mean that the Conservatives are the worst Party for the NHS? Who knows?! That’s for you to decide and take into account (if you want to) when you vote. But at least when doing so, you’ll now have a more complete set of facts when you do.


Immigration and global output

A recently published IZA working paper by Clemens and Pritchett has provided an interesting development regarding the assessment of “optimal” rates of migration. Previous studies tended to focus on the impact of migration on income distribution in the countries that were being migrated to. The Clemens and Pritchett paper is part of a developing literature that, instead, examines the impact of migration on (global) efficiency).

Previous studies looking at the impact of migration on income distribution essentially assessed the extent to which migration affected wages in the countries/areas in which migrants were settling. These often produced mixed results – for example, Borjas found that increases in migration to a country were associated with decreases in wages in that country, whereas Ottaviano & Peri find that immigration actually increases wages in migrants’ destination countries. Just to confuse matters, Card finds that wages are completely uncorrelated with migration.

Hence, there is a need for an alternative way of looking at the impact of migration, which is where these recent developments in terms of the “global efficiency of migration” come in.

The basic idea is that the productivity of labour is low in the countries from which people migrate, but high in the countries to which migrates move. This means that moving people (i.e. labour) from a low productivity country to a high productivity country increases the mean global productivity of labour, such that global output increases.

Consider the stylised example set out in the table below, in which 50 people move from the low productivity country to the low productivity country – the rows indicate whether the situation is before or after this migration occurs. The second and third column of the table indicate the productivity of one unit of labour in, respectively, the migrants’ origin country (the low productivity country) and their destination country (the high productivity country). Columns four and five indicate the number of people in each country, while the fifth and sixth column indicate output in each country (simply each country’s labour productivity multiplied by the number of people in the relevant country).

The final column sums the output in each country to obtain total global output. Comparing this column before and after migration indicates that people moving from the low productivity country to the high productivity country can increase global efficiency. Empirical studies have found that, via this mechanism, global output could be increased by 50% – 150% if restrictions on migration were lifted.

Good case

However, a modification of this mechanism could mean that migration actually reduces global output. Specifically, it could be the case that people moving from low productivity (origin) countries to high productivity (destination) countries actually “bring” some of their low productivity with them, such that the productivity of all workers in the destination country is reduced. If the productivity of labour in the destination country is reduced by a sufficient amount, this could mean that migration reduces global output. In the previous example, it was assumed that the productivity of labour in each country (the second and third columns of the table above) remained unchanged after migration.Such “transference” of low productivity could occur via migrants bringing their cultural or institutional norms with them and potentially being slow to “assimilate” in their destination country.

The table below presents a revision of the previous example in which the only change is that labour productivity in the destination country is reduced by migration (note, however, that productivity in the destination country is still higher than that in the origin country). Even though everything else from the previous example is unchanged, if migration reduces productivity in the destination country, this could mean that migration actually reduces global output. This theory, called the “Epidemiological Model”, has been espoused by the likes of Borjas.

Bad case

Clemens and Pritchett’s working paper tries to bridge the gap between these two opposing mechanisms by modelling the impact of “transmission”, “assimilation” and “congestion” on the rate of migration that maximises global output while eqaulising labour productivity. In this context:

  • transmission refers to the extent to which migrants’ low productivity travels with (i.e. to what degree do migrants actually “bring” any cultural and institutional low productivity with them when they migrate);
  • assimilation is defined as the proportion of migrants that “convert” to being high productivity (i.e. of those that migrate, how many obtain the same high productivity as workers in the destination country); and
  • congestion refers to the impact of un-assimilated migrants on the overall productivity in the destination country.

As such, the model constructed by Clemens and Pritchett trades off the gains of moving labour from a low-productivity country to a higher-productivity country against the reduction in the productivity in the high-productivity country resulting from un-assimilated migrants. Hence, the model embodies the two opposing potential mechanisms by which migration can imapct global output as described above.

The model’s results indicate that optimal migration is higher when:

  • transmission is lower – i.e. if cultural and institutional low-productivity does not “travel” well;
  • assimilation is higher – i.e. if migrants easily and predominantly obtain the same high productivity as workers in the destination country; and
  • congestion is lower – i.e. un-assimilated migrants do not substantially reduce the productivity level in the destination country.

Although these results might seem relatively obvious given the description above, the paper then goes on to use estimates of  the rates of transmission, assimilation, and congestion to obtain an estimate of the “optimal” rate of migration from the perspective of maximising global output. The paper finds that this optimal rate is substantially higher than the actual rate of migration, with the implication that global output could be raised by reducing the current restrictions on migration.

However, there are some flaws with the paper. First, the model of global output that is used to determine the optimal rate of migration only includes labour as an input – i.e. it does not include capital (machinery, infrastructure etc.) as a determinant of output. This is despite the fact that most basic models of output do include capital. The absence of capital from this model is not a problem if migration does not affect incentives to invest in capital, but if migration does affect those incentives, then the results of the model are unlikely to hold in reality.

In particular, if migration increases investment (by reducing labour productivity, thereby making investment more attractive relative to labour), then increased migration increases output such that optimal migration would be higher. Alternatively, if migration reduces incentives to invest, then increases in migration could lead to reductions in capital, potentially decreasing global output. Although the paper tries to cover this off in a single paragraph towards the end of its results, this is far from sufficient (the paper only mentions the first potential impact of capital and sues that to claim that its results are conservative).

Second, the paper notes that the rates of assimilation, transmission, and congestion are relatively unknown yet it does not include a rigorous assessment/estimation of the true value of these parameters. Instead, in order to obtain empirical estimates of these rates, the paper relies on very simple regressions that appear far too basic to capture the various determinants of these rates. For example, the estimates of the rates of assimilation and transmission are based on regressions where the dependent variable is a person’s wage yet the paper only includes controls for age, education, and gender as well as the immigrant status of a person (the variable of interest), despite the fact that estimating the determinants of wages is a highly complex exercise.

Finally, the paper assumes that changes to productivity only flow in one way (i.e. that low productivity workers reduce the productivity in the destination countries but productivity in the origin country is unchanged despite the potential for technology transfers or stimulation of foreign direct investment) and claims that is conservative. In other words, the paper claims that ignoring this possible transfer means that their estimate of the optimal rate of migration is actually lower than the truly optimal rate.

However, this fails to take into account the fact that if such productivity changes flowed both ways, then the productivity in the low-productivity origin country would increase in future, thereby reducing the productivity difference between the high and low productivity countries (i.e. reducing the positive impact of labour moving from the origin to the destination country). This could have the effect of reducing the future optimal rate of migration, but is further complicated by the fact that raising productivity in the origin country might also mean that any reduction in productivity in the destination country through migration is ameliorated somewhat. However, the paper just glosses over this complex dynamic aspect.

Nonetheless, despite these flaws the paper does provide a useful framework and some novel insights regarding how the assessment of restrictions on migration can be developed in future.

Bad Pharma? Not quite!

Earlier this week there a Treasury Committee discussion regarding Intellectual Property took place, some of which was devoted to issues surrounding patent rights. Now, when people think of patents, they often think of pharmaceutical companies and how “evil” it is that such firms can patent and profit from products that are crucial to improving human health. Such protests can be found in publications as varied as Cracked, The Guardian, The Telegraph, The Washington Post, and The Independent, among others.

However, these protestations and complaints are unjustified. At the outset, it is important to note that patents for pharmaceutical products last, at most, twenty years, and that is to say nothing of the fact that many drugs are subject to off-patent, generic versions long before their patents expire. Hence, any potentially high profits earned by a pharmaceutical firm that owns a patent are only temporary.

Moreover, without the existence of said patent protection, a pharmaceutical firm likely would not invest in researching and developing any drugs in the first place. To see this, suppose that a pharmaceutical firm is considering investing in research and development for a new drug – its investment decision depends on whether the (expected future) revenues it can obtain from its investment exceed the (current) costs of the investment (for now let’s ignore the complications that future revenue streams are uncertain and need to be discounted). Simply put, if the revenues exceed the investment costs, then a pharmaceutical firm will invest in the R&D necessary for the new drug.

How much does it cost to research and develop a new drug for market? The BBC quotes an estimate from the Association of British Pharmaceutical Industry of £1.5bn per drug. Other, even higher, estimates also exist. Nonetheless, this means that for investment in a new drug to be worthwhile, the revenues from said drug must exceed at least £1.5bn.

If patents did not exist, then as soon as the new drug was marketed by the “originator” firm that developed it, rival manufacturers would be able to reverse-engineer their own version (usually within a few years, given the length of time it takes generic firms to produce their own version after patent expiry) without having to incur the large development costs to which the originator was subject. Hence, the rival firms would be able to undercut the price charged by the originator. In other words, without patent protection, an originator would only expect to receive minimal revenues for its investment. Under such a scenario, no rational private firm would incur the costs necessary to develop the new drug.

Under the patent system, however, rival firms are prevented from marketing their own versions for a longer time, thereby providing the originator with a longer amount of time over which it can make revenues. This means that the patent system provides the originator with a much stronger incentive to develop the drug in the first place. As such, it is clear that the existence of the patent system is essential to encouraging the development of new drugs.

However, that is not to say that the patent system is perfect. Indeed, it lacks the ability to incentivise pharmaceutical firms to develop treatments for diseases/conditions that only affect a small number of people. In such instances, the “market” is so small that the revenues obtainable under the existing patent system would not be sufficient to cover costs unless an exceedingly high price was charged (in which case health services probably would not purchase it). In these instances, drugs are not developed despite potentially being highly beneficial to those with the conditions/diseases concerned.

How can this be rectified so that drugs are developed for these “niche” conditions? One way would be for patent rights to be extended in length for drugs developed for these diseases. This would have the benefit of ensuring that the existing structure of research and development institutions is maintained, and that R&D occurs where it can be most efficiently undertaken. However, one downside would be that if a drug was to be developed, then health services would have to pay the “patented price” for that drug for a longer time (but that must be better than there not being a drug in the first place).

An alternative solution would be for a scheme that provides public funding or subsidies for the development of such drugs. If this were done via subsidies for existing pharmaceuticals, such a scheme would not have to provide the full amount of developing the drug from the public purse, but merely enough to make the investment by the private firm profitable. This seems much more feasible than a government taking on the role of a pharmaceutical firm itself – a government would have to incur substantial costs setting up and running its own pharmaceutical research and development unit, which seems needless and inefficient given the existence of such units within pharmaceutical firms already.

Obviously, a scheme to subsidise the development of treatments for niche conditions would give rise to some administrative issues. For example, would the subsidy be paid regardless of whether or not the drug being developed actually came to market? How would the amount of the subsidy be determined? What if the developed drug also had benefits for a more common condition? Nonetheless, these issues do not seem insurmountable, so there could very well be a way forward to developing drugs for niche conditions – either extend patent rights, or provide targeted subsidies to pharmaceutical firms.

Climate change and the discount rate: Part 2

As an (unplanned) follow-up to my previous post regarding the impact of the discount rate on analyses of the net benefits of mitigating climate change, let’s have a look into recent estimates of what the appropriate discount rate for assessing said mitigation benefits should be.

More specifically, the NBER recently linked to a paper written by Giglio, Maggiori, Stroebel, & Weber that purports to show that 1) the discount rate that should be used to assess the net benefits of mitigating climate change is at most 2.6%; and 2) therefore,  rates of return on some risky asset classes are not appropriate proxies for the discount rate.

It is the first claim that is demonstrably erroneous, and means that the discount rate that should be used is likely to be higher than 2.6% (and, hence, also implies that the paper’s second claim is not proven). As mentioned in my previous post, although using a discount rate as artificially as low as 2.6% would result in there being a positive net present value of mitigating climate change, using a (more appropriate) higher discount rate reverses that result.

(As an aside, the aforementioned NBER paper is riddled full of holes – for example, the authors claim that “there are no property taxes [that affect the return on property investments] to be considered in the UK”. This is patently false – the authors have ignored, or are simply unaware of various taxes that property developers and management companies have to pay. Either way, this simple error casts serious doubt on the accuracy and reliability of the rest of their results.)

The authors obtain the 2.6% estimate from a previous paper of theirs (rather shockingly, this flawed paper was published in the Quarterly Journal of Economics). In that paper, they looked at differences between the UK and Singapore prices of residential properties with time-limited leaseholds and those that are freehold. This difference reflects the value people place on holding a freehold relative to holding a leasehold with a certain number of years remaining and, hence, can be used to infer a discount rate for cash flows that occur in the very long run.

However, the results of that paper are driven by a fundamental misunderstanding of how long leaseholds and freeholds are treated by someone purchasing a house. The paper claims that leaseholds of more than, say, 500 years are different from freeholds because the leasehold can still expire and rely on this to validate their result. As such, the paper ignores the fact that, for all intents and purposes, a lease that long is treated in exactly the same way as a freehold (by buyers, sellers, estate agents etc.)

Moreover, the paper uses the result that rental values do not differ between freehold and leasehold properties (controlling for observable characteristics such as number of bedrooms etc) to claim that any systematic differences in unobserved characteristics between leaseholds and freeholds are unlikely to be prevalent. However, this ignores the fact that many properties (at least in the UK) are owner-occupied rather than rented. This means that the data they use misrepresent how people value residential properties and the factors that people take into account in their valuations. for example, it is plausible that someone purchasing a house to live in values it differently from someone purchasing the house as a buy-to-let – ignoring this means that the paper’s results are biased.

Furthermore, the paper relies on data covering only the period 2004-2013. In other words, a large part of the data used in the paper are affected by the impact of the global financial crisis, which is widely recognised to have depressed house prices. In addition, the paper fails to control for other relevant factors, such as changes in the housing stock in a particular area over time – as increases in the stock of housing in a particular area would affect the prices of existing houses, this is a grave omission.

Finally, the paper restricts their analysis to the prices of flats, and does not look at other forms of properties at all. This is a grave error since it biases their results towards being more leasehold-focused (flats are typically leasehold, whereas houses tend to be freehold). In other words, the paper fails to take into account a very large proportion of properties that are sold in the UK.

In other words, the result that the long-term discount rate is less than 2.6% is based on a flawed analysis. This means that the results of their latest paper (the one mentioned at the start of this post) are wrong. Indeed, the discount rate that should be used to obtain the net present value of the benefits from climate change is likely to be far in excess of 2.6%. Hence, the results shown in my previous post (that there is a net loss from mitigating climate change) still hold.

Milk prices in the UK: Subsidies and their impact

There has been more than a little recent commotion regarding the prices that supermarkets and milk processing firms pay dairy farmers for supplying their milk. Indeed, the BBC has reported that some farmers have paraded cows through aisles in certain supermarkets, block-bought all the milk no a supermarket’s shelves, or simply blockaded a supermarket (see here, here, and here).

The motivation underlying these protests appears to stem from small dairy farms (such as those operated by “family farmers”) being unable to make a profit at the prices that supermarkets and milk processing firms are willing to pay for their milk. The Guardian mentions farmers’ claims that their break-even level (i.e. the level above which they start to make a profit) is 30p per litre, but that farmers are only paid about 24p per litre, resulting in a loss of 6p per litre (see here).

This has lead to protests that farmers should be paid more for their milk. However, these protests do not appear to have considered the impact of this price increase on the overall efficiency of milk production. Presumably, the larger milk farms (i.e. those run on a large scale, rather than by individual households) are able to make a reasonable profit at the prevailing price – this profit can be considered a “reward” for those farms making the best use of their resources to produce at as low a cost as possible.

On the other hand, if small dairy farms cannot produce at a cost lower than the price they receive, ordinarily they would go out of business because they are too inefficient to compete with their rivals. This is a standard economic process of obtaining “productive efficiency” – i.e. that all goods are produced at the lowest possible price.

If small dairy farms are (effectively) subsidised via the supermarkets paying a higher price, that would keep them in business, but at what cost? First, the production of milk would not be using the most efficient resources available to it. Instead, inefficient farms would be encouraged to remain in business producing milk when those resources (the people working on the farm, the equipment used, and the land itself) could be put to a more productive use. Moreover, farms that would be able to become more efficient if they had the incentive to do so (i.e. were receiving a lower price) would no longer need to make the necessary investments that would make them more efficient. In other words, subsidising small dairy farmers has the potential to harm the overall efficiency and productivity of the economy, both in the short-term and in the long-run. In other words, the reduction in efficiency and productivity would result in the UK’s GDP being lower than otherwise had the resources tied-up in the inefficient milk farms been put to other, more productive use.

Second, the fact that supermarkets are paying a higher price for their milk means that their costs have increased – it is plausible that these costs could be passed on to consumers in the form of higher milk prices. As milk is a staple product, any increase in the price of milk is likely to affect a large number of consumers, so the effective subsidising of small dairy farms has the potential to cause wide-ranging effects across the whole economy. In particular, if consumers need to pay more for their milk, they would have less disposable income to spend on other products, resulting in a decrease in demand. However, given that any increase in the price of milk is likely to be small, the magnitude of this effect may well be negligible.

Third, any increase in the price of milk paid to farms (to the extent that the price of milk is not negotiated on a bilateral basis) means that those farms that are already efficient merely serves to increase the profits of these farms. It is somewhat surprising that this aspect of an increase in milk prices has not received more attention, particularly if one has ethical qualms regarding policies that increase the profits of international firms.

Obviously, weighed against these points are any potential benefits that arise from the continued existence of small dairy farms. Perhaps benefits are derived from their impact on the countryside and/or the local communities in which they operate. I am not in a position to elucidate on these potential benefits, but the “cons” of any increase in the price of milk via an effective subsidy certainly need to be taken into account.

A nice, non-controversial topic to start off!