Wednesday, 21 November 2018

Amazon, Google & Big Tech’s Productivity Paradox


Whatever you may think of the multi-billionaire founders of Amazon and Alphabet-Google, [1] there would seem to be one undeniable fact about their companies: they have massively improved productivity. Amazon has an e-commerce system that delivers very efficiently; Google has revolutionised Internet search. Yes, there is quite a list of undeniable negative facts too – poor working conditions in Amazon ‘fulfilment’ centres, the hoovering up of personal data by Google, how each company’s rise to power has upended the economics of other businesses, and much else besides. But the productivity benefits of their services seem unimpeachable. It lies behind Amazon’s increasing share of the consumer market and Google is now used in around 90% of Internet searches. Nevertheless, a closer look at these tech giants shows that all is not what it seems.

Productivity and the market

Being more productive is a good thing, or at least it should be. It means producing with fewer resources – less time spent working, travelling or waiting to get the same output, using raw materials and other inputs more efficiently, and so on. In the framework of the capitalist market, however, this can have all kinds of bad repercussions. Rather than society being able to be better fed, in better health, with more leisure and time to enjoy life, the burden of work for some increases while others are made unemployed, lives are disrupted and the benefits of productivity go to a few.
Apologists for capitalism may accept this point, but would argue that the market system encourages all kinds of innovation and, while there are some unfortunate side effects, in reality it is only this kind of system upon which progress for society as a whole can be built. That perspective leaves out many things, not least capitalism’s propensity for wars and destruction, the monopolisation of the world’s resources by a few powerful companies and governments, and the oppression of hundreds of millions of people for whom being part of the capitalist world economy more often means a ticket for the treadmill rather than a path to progress.
But I will leave such damning truths for now. I will also give insufficient attention to how invention is most commonly a social phenomenon, even one backed by state resources, not a bright spark from a lone genius. Or how innovation is ever more dominated by rich capitalist companies that buy into ideas to help them build or sustain a monopolistic position in the supposedly competitive market. I won’t even discuss how the slump in Alphabet-Google’s and Amazon’s share prices since the summer will give them problems with investors, since, like Facebook, they have a policy of paying no dividends. Instead, I want to uncover the peculiar features of productivity at two of the Big Tech giants.

Google’s advert stream

Despite Alphabet’s forays into robotics, artificial intelligence and ride-hailing, the company’s business still very much depends on advertising revenues from Google.[2] The non-Google business, termed ‘Other Bets’ in its accounts, generated barely 1% of sales revenues and made a loss of $3.4bn in 2017. By stark contrast, the Google ‘segment’, including YouTube and other items such as cloud computing, registered an operating profit of $32.9bn on revenues of $109.7bn in 2017.
Within the Google operation, the importance of advertising revenues has fallen from around 99% of the total up to 2007 to around 85% now. This reflects both the company’s increased difficulties in boosting such revenues and how cloud computing, money from selling apps and Google Play have become more significant. Even a monopolist has to diversify! Google’s advertising numbers are nevertheless key for the conglomerate’s business and are likely to remain so for years to come. Examining these and other related data brings out important aspects of corporate productivity – with all the caveats attached to that term indicated in the previous section.
It is instructive to look at Alphabet as principally a company that sells adverts. Its ‘output’ is then how many adverts it sells.
This output can be assessed in two ways: by simply looking at the advertising revenues and by counting the adverts themselves. However, there are no figures available for the latter. Instead, one has to adopt the philosophical view that something only exists when it is experienced and, in this case, more specifically, when someone ‘clicks’ on the advert. Alphabet goes further than this and, as befits a capitalist corporation, reports changes in the total number of clicks on the adverts that a company must pay for. Annoyingly, it does not give totals, only percentage changes, but it does also note the change in the ‘average cost per click’, in other words, the average price it receives from the clicks that have taken place.

Google's Clicks

I have trawled through Alphabet-Google’s accounts and noted the annual changes in the total paid clicks and the average cost per click since 2005. It was not necessary to go back any further to bring out the main features, leaving aside the serious threat to my sense of well-being that this would have entailed. The picture is clear: a very big rise in paid clicks and a fall in the average cost per click. So, a huge increase in ‘productivity’, as the volume of output has risen while the unit cost has dropped.
Of course, selling more adverts at a lower cost is not what a normal person would call being more productive. But this is the capitalist market view, not an assessment of what may be good for society. That view means discarding social values in favour of market values, where the magic of the market is also responsible for downward pressure on the prices charged for the adverts. Fortunately for Alphabet-Google, the result of the more adverts/lower costs trend has been a strong growth in total revenues from adverts, reaching nearly $111bn in the year to end-September 2018, compared to around $50bn five years before. Unfortunately for the company, the pace of this revenue growth has faltered in recent years and there are other metrics more commonly followed by capitalist markets that do not augur so well for it.
Take another type of productivity measure, one that compares revenues or operating profits with the property and equipment assets the company holds. Such assets are the fixed capital needed to generate the returns, and one would expect that a larger volume of investment would lead to increased revenues and profits. They do, but at a slowing pace. This can be seen in how the ratios of revenues and profits to fixed assets have fallen, as illustrated in the next chart.
At the end of 2006, Alphabet-Google held $2.4bn in property and equipment assets. Ten years later, at end-2016, that had multiplied to $34.2bn. By the end of September 2018, less than two years later, it was more than 60% higher again at $55.3bn. Total revenues and operating income grew much more slowly, so the lines in the chart show a fall. A similar picture holds for revenues and income compared to the company’s expenditure on research and development. Alphabet-Google’s R&D is very high, at close to $20bn in the year to end-Q3 2018, taking up 15% of revenues and measured at 76% of operating income. But the R&D expenditures have also risen much faster than either revenues or income.

Google's Operating Income & Sales Revenues versus Fixed Assets

Amazon’s anomalies

My previous review of Amazon’s business (here) noted that it had surprisingly little profit for a burgeoning tech giant, and that much of its growth in sales revenues and profits had come from its cloud computing arm, Amazon Web Services (AWS), not from its widely known e-commerce operation. I must admit to wondering why Amazon had continued to expand outside North America, given persistent and widening losses. However, the latest data show that these International losses have finally begun to narrow. They were a little over $2.4bn in the year to the end of the third-quarter 2018, down from a $3bn loss in 2017. That may be enough to keep in place Amazon’s ambition to take over the e-commerce world, although the parcel delivery operation outside the US has so far looked more like an expensive branding exercise for AWS’s dramatic growth.
Amazon’s operating income (loss), 2015-2018 ($ million)[3]

2015
2016
2017
Year to Q3 2018
North America
1,425
2,361
2,837
6,708
International
(699)
(1,283)
(3,062)
(2,419)
AWS (Amazon Web Services)
1,507
3,108
4,331
6,473
Total
2,233
4,186
4,106
10,762
Amazon’s latest figures show a sharp rise in sales revenues and operating income. These are mainly from North America, and were boosted by Amazon’s purchase of Whole Foods Market in August 2017. At the same time, revenues and income from AWS are continuing to grow very rapidly. It could also be that the international e-commerce business will finally benefit from Amazon’s big investments. It will need to in order to turn around what are surprising trends in operations from a company that would otherwise appear to be the epitome of cost cutting.
Take Amazon’s ‘fulfilment centres’, for example. These are the enormous warehouses of goods, not only staffed by low paid workers, but also full of amazing technology and robotics to optimise selection, packaging and delivery, together with algorithms to minimise the paths taken. One would expect that the costs of running these would increase as the sales business expands. But, at the same time, these costs should not rise as quickly as sales, since economies of scale would kick in. Nevertheless, Amazon’s accounts show that the costs of these fulfilment centres have risen faster than Amazon’s net sales. In 2009, fulfilment centre costs were $2bn, which was 8.4% of net sales that year. The proportion had risen to 14.2% by 2017, when such costs were $25bn, and it was higher still in 2018.
It would be difficult for a new entrant into this market to outcompete the Amazon machine. Yet the rising ratio of costs to net sales revenues raises questions about how productive these centres really are. Is their efficiency exaggerated in the minds of those who only see the robots, the disciplined workforce and the smooth running system, and who ignore what all this costs?

Where some of it happens ...

One reason for the increase in costs compared to sales revenues is due to Amazon opening lots of extra fulfilment centres worldwide – see the information on these centres here. This will incur costs before they are fully operational and generating extra sales. Another reason lies in the distinction that must be made between a physical, productive efficiency and the value of the goods delivered. Unfortunately, Amazon provides only sporadic details on this topic, as on others, so a view cannot be properly verified from data in their accounts. Nevertheless, the information available suggests both that the volume of throughput at fulfilment centres – the number of items, parcels, etc, per day – has risen sharply, and that the cost per item or parcel has fallen steadily. So, physical productivity has increased, as one would have expected.
That has not been translated into higher net sales revenues when measured against the costs of these centres, not only because of the rapid expansion in the number of these centres, but also because Amazon has reduced the average prices of the goods to its customers. This might seem a strange thing for a capitalist company to do, but it is an explicit part of Amazon’s strategy of building volume and increasing its share of the market.
An example to support the view that average prices have fallen is that in 2017 the value of consolidated net sales rose by just over 30%, to $178bn. In the same year, shipping costs rose by 34% in 2017. The volume of deliveries likely rose by even more than 34%, given that Amazon continues to pressure delivery companies to cut their fees.

Amazon passes on the costs

This price cutting strategy is a risk for Amazon’s profits, but that risk is reduced if it can pass on the pressure to its suppliers! This is something it has been very effective at doing.

 Amazon's Costs as a Percentage of Net Sales
Most of the expenses that can be measured against net sales – such as the investment in technology, fixed assets or administration – have been rising, just like those for the fulfilment centres. One item has not: the cost of sales. This number is principally made up from what Amazon pays the suppliers of the consumer goods it sells and is the largest expense item. The cost of sales was $112bn in 2017, 63% of the total value of consolidated net sales of $178bn. That percentage has fallen steadily from nearly 78% in 2010, dropping to just 58% in the third quarter of 2018.
This rapid fall in the cost of sales relative to total sales has kept Amazon’s e-commerce show on the road. The gap between the two numbers represents Amazon’s gross profit, from which it can fund other things. It reflects Amazon’s power as a platform for selling consumer products, a power that has grown dramatically in recent years and which allows it to force other companies to deliver its goods more cheaply.

Productivity and profitability

Investing more and undertaking R&D is what one would expect a productive capitalist company to do. What Alphabet-Google and Amazon may not have expected is that this would go alongside an increasing difficulty in producing extra revenue and profits. The two companies work in different ways, but the trend for each is similar. Their ‘output’ costs per item (of adverts or shipped goods) have fallen, but the scale of necessary investment to bring this about has risen faster than sales revenues and profits. It is perhaps stretching the interpretation of company accounts a little too far to see in this a tendency for their rate of profit to fall as the company multiplies up the scale of investment. Nevertheless, the figures for Alphabet-Google do show a distinct drop in both operating income and total revenues compared to its fixed assets.

Amazon's Sales and Operating Income vs Fixed Assets


A similar picture is true for Amazon (see the previous chart). Between 2004 and 2008, its net sales were more than 20 times the value of its fixed assets. By 2015, the ratio had fallen to just five times, hitting a low of 3.6 in 2017. There has been only a minimal recovery since as Amazon’s sales revenues have jumped by just a little more than the jump in fixed assets held. In the case of operating income versus fixed assets, there has also been a trend decline for Amazon. The ratio was 1.8 in 2004, falling to 0.6 in 2010 and 0.1 in 2015. That ratio has also recovered a bit in the most recent period as profitability improved, but was still only around 0.2.
Each company is a Big Tech giant, though Alphabet-Google’s business machine sells advertising slots while Amazon’s started out only selling more efficiently what others had produced. Each has tried to diversify operations, using the vast resources made available to them by their respective monopolistic positions. Each brings out the peculiar manner in which capitalist corporations boost ‘productivity’, one that is anti-social, given the effects on society at large. They are part of the imperialist world market and play a role in its domination of society, but, unfortunately for them, they cannot escape the constraints on profitability.

Tony Norfield, 21 November 2018


[1] The original Google company was reorganised, and from October 2015 it became part of a conglomerate, Alphabet Inc. In September 2017, a shell company was set up, XXVI Holdings Inc. The Google segment of the business remains by far the largest component of the overall operation, and that will be the main subject of this article.
[2] For a fuller account of Alphabet-Google, see my previous blog review here.
[3] Business accounting definitions can be tricky to follow, but note that Operating Income is defined as Net Sales minus Operating Expenses minus Stock-based compensation and other items. Also, Net Income is defined as Operating Income minus Non-operating income/expense, Provision for income taxes and Equity-method investment activity, net of tax.

Monday, 12 November 2018

Finance, Imperialism and Profits

Last Friday I took part in a panel to launch a new book, World in Crisis: A Global Analysis of Marx’s Law of Profitability, published by Haymarket Books, edited by Michael Roberts and Guglielmo Carchedi. The presentation was at this year’s Historical Materialism conference in London.

My presentation was on ‘Finance, Imperialism and Profits’, in which I stressed the need to develop Marx’s theory in order to explain the world today. I argued that an accurate measure of a rate of profit (in Marx’s sense) could not be gleaned from official statistics and that, among other things, this was because of the nature of the imperialist world economy. Also I noted that for some adherents of Marx’s ‘falling rate of profit’ theory, this theory was somehow consistent with their calls to nationalise banks and regulate finance. This expunges the revolutionary content of Marxist theory, shows a naïve faith in the capitalist state and makes concessions to nationalism.

This was a lot to cover in the twenty minutes available, so could only be done in summary form in the presentation (see below), but there was more time in the Q&A.


Tony Norfield, 12 November 2018

-----------------------------------------------------------






































Monday, 5 November 2018

Japan’s SoftBank: Tech Parasitism


The two Sons shake on $45bn
Masayoshi Son faced a dilemma in October: should the Japanese businessman go to an investment conference in Riyadh, Saudi Arabia? The guy running that conference had promised Son’s Vision Fund $45bn – that’s not a misprint, that’s forty-five thousand million US dollars – so not showing up would look more than a little ungrateful. He was also the Crown Prince of Saudi Arabia, next in line for the Saudi throne, and a person not known for taking lightly any lack of due respect. Yet the same guy had just been implicated in the murder and dismemberment of a journalist he did not like. While you and I would let this go as being just one of those things, the media and the political class of some powerful countries had shown themselves to be unhappy with the event. If Son attended the conference it could put his investment company SoftBank, and its Vision Fund, in an unfavourable light.
It was all so unfair. Nobody of any importance had complained about Crown Prince Mohammad bin Salman’s exploits in Yemen that were killing off a whole population! Why make a fuss about a minor journalist being disappeared in the Saudi Consulate in Istanbul? It was evidently all a matter of big power politics, and who was allowed to do what to whom and when. But the dilemma was nonetheless real for Masayoshi Son, given the risk of negative publicity for his investments in projects for the tech-wonderland future. After much consideration, he decided on a diplomatic compromise. He travelled to Riyadh to talk to the oil oligarch, but did not attend the conference itself.
Most people will not have heard of the Vision Fund, or of SoftBank, not least because both names sound like they were suggested by a bored publicist suffering business cliché indigestion on a dull afternoon. But it is worth paying them attention for the light they throw upon today’s imperialist world economy and how innovation becomes entrapped by a parasitic machine. SoftBank itself does not rank highly in the list of global corporations, with a stockmarket capitalisation of just $82bn at end-October. Yet its Vision Fund is the world’s largest ‘venture capitalist’. It specialises in investments in the technology sector and is reported to have investment funds available of nearly $100bn – of which more below.



The rising Son

In 1981, Masayoshi Son founded SoftBank in Japan, but for many years the company was almost unknown outside the country. It began as a distributor of packaged software, also getting into computer magazine publishing and running business events. By 1998, it had become big enough to have its shares listed on the First Section of the Tokyo Stock Exchange, and in 1999 it became a pure holding company aiming to expand its presence in other areas of the Internet and mobile technology sector. From the mid-1990s, SoftBank did a number of very profitable deals in Japan with US web services provider Yahoo, including one with Yahoo Japan of around $9bn which gave SoftBank 43% of the company. It also bought Vodafone’s Japanese mobile operation for $15bn in 2006 and, from the late 1990s, it began to make its first significant deals outside Japan.
SoftBank’s most successful investment has been in Alibaba of China. In 2000, SoftBank advanced a mere $20m for a 29% stake in Jack Ma’s fledgling company, plus a modest later investment. The value of this holding soared to $60bn when Alibaba went public in 2014, and is now valued at around $100bn. Other major SoftBank investments have been in 2012, when it invested $23bn in Sprint, the fourth largest mobile network operator in the US, in 2016 with the $31bn takeover of ARM Holdings, a UK-based chip designer for smartphones, and in 2017 with the $9bn or so put into the US ride-hailing company Uber for a 15% stake.
None of these have gone anywhere near as well for SoftBank as Alibaba. For example, Sprint, 83% owned by SoftBank, after losing market share and subscribers is now in the process of being rescued by a merger with T-Mobile US, owned by Deutsche Telekom. If that goes ahead, SoftBank will own 27% of the new business.
There have been many reorganisations and name changes of companies in the SoftBank group. Its portfolio of holdings has also increased dramatically in recent years, with investments ranging from a complete or near-complete takeover of another company to deals that involve SoftBank owning perhaps only 5-10% of its shares. The prices SoftBank paid for these have not always been clear, since it has often been part of a consortium of other funds that have bid for a stake in the particular venture.
Given SoftBank’s promoted image, a natural assumption is that all of its investments are in the ‘technology’ sphere. This would suggest e-commerce, mobile communications, online services and so forth. But often the investments extend into other areas that have little or no connection with these and may be just an online application to contact a service. Although the latter is a pervasive feature of the economy today, it is not so far removed from telephoning a company to make a booking, rather than being a sign of ‘hi tech’. Notable in this respect is WeWork, a US company leasing out office space in which SoftBank (and its Vision Fund arm) has already invested more than $4bn, and the $300m invested in the US-based Wag, a dog walking service! I will not mention SoftBank’s investment in a Japanese baseball team, the Fukuoka SoftBank Hawks.
Elements of this that remind me of the dotcom equity market bubble of the late 1990s. One anecdote from that time was that a laundry company saw its share price soar once it had changed its name to laundry.com or something similar. SoftBank is not the laundry company, but its share price had also boomed in that market bubble, to around ¥19,000 in early 2000, but by November 2002 it had slumped to just ¥300. Happily for Mr Son and his shareholders, SoftBank’s equity price has since risen and was at around ¥9,000 by end-October. But the vulnerability of the company to changing fashions is seen in the 20% drop through that month, partly prompted by the declining fortunes of another ‘son’, Mohammad bin Salman. Such volatility is not uncommon in the markets for financial securities, but an examination of SoftBank’s accounts, and the new Vision Fund, shows that there is a lot going on behind the headlines.

No, not this robot dog, a real dog!
---

Assets, debt liabilities, income

SoftBank’s annual report for the year to end-March 2018 gives the basic picture for its assets, liabilities and income that still holds today. Two features stand out. Firstly, the company’s cash revenues have derived mainly from its telecom operations in Japan and the US; secondly, it has a lot of debt.
The telecom operations have the advantage of generating an inflow of cash, with regular subscriber payments and sales of mobile phones, and in the annual 2018 report these accounted for just over 70% of net sales and over 90% of adjusted earnings before interest, tax payments, etc, for the group as a whole. This cash comes in handy for SoftBank’s appetite to invest in other companies, but most of SoftBank’s requirements are instead met by its loans from banks or its issue of bonds. This has led Softbank to accumulate an unusually high level of debt, amounting to $160.4bn by the end of March 2018.
Financial markets focus on a measure of how much interest-bearing debt that the company has outstanding and compare that to the equity investment of the company’s owners in the company itself. This ‘debt-equity ratio’ is one indicator of a company’s ability to pay back its debt liabilities if its operations get into trouble. Outstanding debt levels and also the debt-equity ratio will be different for different kinds of company, but industrial and commercial companies rarely have a debt-equity ratio above 1 or 100%. In other words, their outstanding debt is not greater than how much equity the owners have invested in the company.
The debt-equity ratio is not necessarily high for companies in the tech sector. Even startups usually get funds from equity investors, rather than depending much, if at all, on long-term bank loans and issues of bonds. For example, in 2017 Alphabet-Google’s debt-equity ratio was less than 3% while Amazon’s, although higher, was still below the 100% level at 89%. In 2015, the year before SoftBank took it over, ARM Holdings had no outstanding debt at all. By stark contrast, SoftBank’s own debt-equity ratio in March 2018 was 271%, and a still high 220% counting only the long-term debt of $130.1bn.
This level of debt is a problem for SoftBank because the funds have been used to invest in a wide range of tech (and not so tech) companies, as already noted. As their market value changes, so will the value of these assets on SoftBank’s books, which makes the company very vulnerable to a change in financial market sentiment on the outlook for these ventures. Meanwhile, the debt remains until it is paid off, and until then it has to be serviced. In the year to March 2018, SoftBank’s net income from continuing operations was $11.7bn, but this figure had been reduced by the interest paid on its debt of $5.1bn.
How could SoftBank continue to expand its investment in tech companies when it already had high levels of debt? One way was to sell off some existing assets as a means to raise cash. An example earlier this year was the $4bn sale of its holding in Flipkart of India to the US giant retailer Wal-Mart, registering a gain of some $1.5bn. Back in April, SoftBank also used its stake in Alibaba as collateral for a bank loan of $8bn. Furthermore, there is a plan for the public sale of some shares in SoftBank’s mobile business in Japan, hoping to get as much as $30bn, although that hope is undermined by Japan’s regulator forcing mobile companies to cut their charges by as much as 40%. But the real scope for expansion lies with the venture noted at the beginning of this article: the Vision Fund.

Double Vision: $28bn becomes $72bn

SoftBank’s Vision Fund was set up in 2017 after being announced the previous year. It is included in SoftBank’s reports as a division that aims to target ‘long-term investments in companies and foundational platform businesses that seek to enable the next age of innovation’. While there are many other hyperbolic statements with which the Fund describes itself, and details of its structure can be confusing, I would recommend keeping the following points in mind to clarify what is going on.
The logic behind the fund’s existence is the limit on expansion that SoftBank faced with its high level of debt. Otherwise there would have been little reason for SoftBank to make big efforts to attract outside investors. Related to this, an important aspect of the Fund is that it has now given Masayoshi Son huge resources from these outside investors over which he has complete control. Meanwhile, SoftBank has not limited itself from undertaking any investments it likes outside of the Vision Fund set up.
SoftBank’s investment in the Vision Fund is reported as $28bn, with the other, external investors providing $72bn, to make up the $100bn when all funds are committed. That makes a good headline, but all is not what it seems. Not simply because most funds are committed rather than having yet been allocated, and the number does not yet quite add up to $100bn anyway. Let us assume that all the commitments will turn up. Instead, the main issue to puncture the headline bubble is that more than $15bn of the capital, and perhaps as much as $25bn or so, is not a pile of new cash waiting to be invested. It simply represents the value of existing investments held by SoftBank that the company has transferred from its main accounts to sit now under the Vision Fund heading.
At end-September 2018, the value of Vision Fund investments was $35.8bn, with an acquisition cost recorded at $28.1bn. A big chunk of this, represented by acquisition cost, consists of previous investments made by SoftBank. For example, a little over $8bn for 25% of SoftBank’s ownership of ARM Holdings, $5bn from its stake in Nvidia, a couple of billion from its stake in WeWork and some smaller investments, including in Wag. SoftBank’s $9bn holding of Uber will also be transferred to the Vision Fund, but this had not happened by end-September.[1]
These SoftBank ‘investments’ in the Vision Fund are not new cash that it can use to invest in other things. So its firepower is significantly less than the $100bn number promoted in the headlines, although it is still clearly a big number. The key point, however, is that by establishing the Vision Fund, SoftBank can get control of up to around $70bn more from the funds committed by other investors.
Under refurbishment: Vision Fund London office

Vision investors, debt and equity

SoftBank’s 28% of the Vision Fund would appear to give a higher weight to the external investors, who have 72%. But there is another complication: whether the investors have an equity stake in the fund or whether they buy the ‘preferred’ units of the Vision Fund that will pay them an annual coupon, as if they owned a debt security. According to a Financial Times report in June, the Vision Fund set up is where the external investors have 62% of debt and 38% of an equity stake in the Fund for every billion they put in. SoftBank therefore has a majority equity stake in the Vision Fund, given that all its 28% investment is for equity.
External investors in the Vision Fund are of two kinds, and each has a different motivation that I will give myself the freedom to speculate upon.
The first kind is the Gulf investors with $60bn of commitments: $45bn from the politically-devalued Crown Prince, allocated from Saudi Arabia’s Public Investment Fund, and another $15bn from Abu Dhabi’s Mubadala Investment Company. These are funds that aim to boost the wealth of the already rich Gulf states by investing in something other than the low-yielding government bonds issued by the major powers.
It is not difficult for the Gulf investors in this venture to feel they are smart money capitalists when all they have to do is get a better return than on US Treasuries. The Vision Fund will have looked an attractive option, one full of a high tech optimism that helps obscure the reactionary reality back home, and doing so with a promised high return – for details of which see the next section.
The second kind of external investor is a group of four companies not new to the world of tech exploitation – Apple, Qualcomm, Foxconn and Sharp . They will offer $5bn in total to the Fund. For them, the amount is trivial, but it may give a reasonable return and it will also give them a valuable overview and early insight into developments that could impact their businesses.

‘Eat yourself’ returns and SoftBank upside

So what is the return for investors in the Vision Fund? These investors, and SoftBank itself, get paid in different ways, and this highlights that it is called the Vision Fund for a good reason.
Those who have equity stakes in the fund get the relevant portion of the returns from the portfolio of investments made, but that is after money has been deducted to pay for the annual 7% coupon on the Vision Fund debt securities purchased by external investors. While this 7% coupon looks attractive compared to other debt securities in the financial markets today, it may have escaped the external investors’ attention that this coupon payment will also reduce the return they will get from their equity stake. If the Vision Fund debt component amounts to $44.6bn (62% of the external $72bn), then around $3bn per annum will be deducted from the profits made on Vision Fund assets to deliver the external investors their coupon payments. They look to be protected from any downside in the equity and revenue performance by their fixed 7% coupon, but that leaves the tricky question of who will pay them the coupon money if the Fund’s return is insufficient.
The external investors will have noticed that they are paying SoftBank a management fee of around 1% for the privilege of running the Vision Fund, which could be up to $720m per annum. SoftBank will also cream off 20% of any return on investment over 8%. In the world of ‘venture capital’ investment funds, however, these conditions are, if anything, low cost.
Overall, the Vision Fund gives SoftBank a vast amount to finance future tech investments, and it gets around some of the constraints posed by SoftBank’s high debt levels. If there are difficulties paying the fixed 7% coupon, then that may be a Vision Fund problem with its investors, not a SoftBank problem of default on its bond liabilities.
Another important point is that the Vision Fund’s investments have delivered it very little in operating profit. Its recorded ‘income’ from its assets is overwhelmingly made up from capital gains on their market value, including unrealised gains. In the six months to end-September 2018, the operating income from the Vision Fund was around $5.5bn, but $1.5bn was from the gain on the sale of Flipkart and another $4bn or so was from increases in the value of Nvidia and some other assets. This points to problems that Mr Son’s venture will have in generating enough income when the market turns down.

Parasitic vision

In an interview with TechCrunch in September, a Vision Fund managing director set out the Fund’s investment policy. He explained that it was a ‘late stage growth fund’. It did not aim to give early advice to tech startups, but instead wanted to see how far they could become a key player in the market. If they were happy with a company’s plans, they would invest a minimum of $100m to finance its growth.
This reveals perhaps more than he realised. Yes, the Vision Fund provides a tech company with funds, but only after it has passed the difficult, uncertain, early stages of growth when survival is at risk, and when it now looks like the only barrier to dramatic expansion is a lack of funds. This is not so different from what a regular bank would do, except that the Vision Fund will make sure that it has an equity stake in what it hopes will be a rapidly growing business, rather than a bank that simply sees good market prospects as giving it confidence that a loan will be repaid. Far from being the daring investor backing ‘the next stage of innovation’, the Vision Fund is more like a money capitalist bean counter that will first ensure that all its boxes are ticked.
Another aspect of the Vision Fund shows that it understands the nature of the imperialist world market today, at least as it applies to the technology sector. The minimum $100m investment is to finance a big increase in the scale of which one of its chosen tech companies will operate, both within its national sphere and internationally. A key feature of businesses that have communications technology as a core element is economies of scale. Here, much the same cost infrastructure is needed to service tens of millions of customers as for tens of thousands, except perhaps the need for a bigger computer server and some better software. Costs per customer will tend to fall rapidly and net revenues can rise sharply.
This is also something that leads to monopolisation of markets. Companies that are backed with funds to invest and expand when they have no operating profit and, like Uber, may be running at a loss, can still invest to sideline competitors. SoftBank and the Vision Fund are involved in this process. One example is the likelihood that SoftBank will play a part in carving up the ride hailing market, given its stake in Uber and in a number of other companies in that area, notably DiDi of China, but also Ola in India and Grab in Singapore. Recent business media reports suggest that these companies, which are often rivals in the same markets as well as having stakes in each other, could decide to ‘cooperate’.

Tech in the machine

What we find today are many examples of technical inventions and innovation, but all of these get bound up in the monopoly machine of imperialist economics and finance. Rather than communications technology being developed to benefit humanity, any good outcomes that may result depend first upon whether the innovation can meet the machine’s demands.
Paradoxes also abound, highlighted especially by how some of the most reactionary regimes in the world put up many billions of dollars to fund ‘progress’. One acute observer of the tech world, Evgeny Morozov, speculated that the ‘disruptive innovation’ backed by Saudi Arabia would include killer robots and the ability to smoothly dispose of dissidents’ bodies. But one must not lose sight of how these regimes are also part of the imperial money-go-round, with full backing from the US and the UK.
The tens of billions of dollars allocated to SoftBank’s Vision Fund are only a small sample of the massive funds potentially available worldwide to address everything from debilitating diseases, to malnutrition and environmental destruction. Instead they are advanced with a beady-eyed parasitism to find the right profitable niche in the market and monopolise it. Even then, the decisions on how the world’s resources will be used rest with a small number of multi-billionaires and the states that back them.

Tony Norfield, 5 November 2018


[1] In 2016, Saudi Arabia’s Public Investment Fund had already invested $3.5bn in Uber, which faced strong competition from one of SoftBank’s other ride-hailing investments, in DiDi (which eventually took over Uber’s China operation, but also gave Uber a stake in the merged company). It has been reported that to avoid Saudi embarrassment of funding a competitor to Uber when it put money into the Vision Fund, SoftBank made sure that the DiDi holding was kept in a separate fund. This is shown in SoftBank accounts as the ‘Delta Fund’, but DiDi is its sole component as a $5bn investment.

Sunday, 4 November 2018

Artificial Intelligence & Images of the Real World



I have long had some scepticism about machines, having been frustrated many times when PCs, vending machines and mobile phones failed to work, so I was in the right frame of mind to see a recent BBC report on ‘machine learning’ and artificial intelligence (AI).[1] Touted as the next Big Thing in the hyped world of technology, the missing element of many reports on AI is that a fundamental building block is often taken from the bottom rungs of the world economy.
Especially in image-related applications, such as needed for driverless cars, the ‘machine learning’ can only work because a human teacher tells the machine what something is. For example, to learn that a set of pixels is a tree, or a road sign, or a pedestrian crossing the road, or a building or another car, the machine is fed many millions of tagged images of trees, signs, pedestrians, buildings and cars taken from the road environment. Now, where does this input come from? Surely the geniuses of Silicon Valley do not sit up all day and night staring at screens and tagging the images?
That’s right, they don’t. Instead, the countries and people that normally provide low cost raw material to the rich nations and their corporations also provide this low cost data raw material.

The BBC Click programme highlighted a case where a San Francisco-based NGO, Samasource, employs hundreds of people from a slum area of Nairobi, Kenya, to tag images for companies such as Google, Microsoft, Ebay, Walmart and Lyft. The head of Samasource, Leila Janah, said that they pay a ‘living wage’, $9 per day, and this means that on average their employees’ household income is increased to more than five times what it was before. Questioned about this still pitiful wage, she said ‘if we were to pay substantially more than that, it would throw everything off’.
Janah’s explanation of things going ‘off’ was that higher wages would have the effect of raising local living costs for housing and food. Although this was an evident implication, she did not mention in the broadcast interview that much higher wages would not fit the business model of those tech companies at the forefront of ‘innovation’ that were using the data from Kenya.
It is progress for impoverished slum dwellers to have a better living from being engaged in the world economy. But at the same time they are being fed into the machine that provides the main benefits for the rich corporations. This is how capitalism manages technical progress, and is another example of the way the imperialist world works.

Tony Norfield, 4 November 2018


[1] The points noted below are from a BBC Click programme broadcast in the UK on 3 November. The series is on Youtube, but I cannot find a link yet for this ‘Kenya’ episode.