Saturday, November 18, 2023

Weekend reading links

1. John Burn-Murdoch points to two studies about the likely impact of AI on occupational groups. Here's about the first one (paper here)
In an ingenious study published this summer, US researchers showed that within a few months of the launch of ChatGPT, copywriters and graphic designers on major online freelancing platforms saw a significant drop in the number of jobs they got, and even steeper declines in earnings. This suggested not only that generative AI was taking their work, but also that it devalues the work they do still carry out. Most strikingly, the study found that freelancers who previously had the highest earnings and completed the most jobs were no less likely to see their employment and earnings decline than other workers. If anything, they had worse outcomes. In other words, being more skilled was no shield against loss of work or earnings.
And this about the second study (paper here) where the consultancy BCG randomly gave GPT-4 to its employees and monitored its impact,

BCG staff randomly assigned to use GPT-4 when carrying out a set of consulting tasks were far more productive than their colleagues who could not access the tool. Not only did AI-assisted consultants carry out tasks 25 per cent faster and complete 12 per cent more tasks overall, their work was assessed to be 40 per cent higher in quality than their unassisted peers. Employees right across the skills distribution benefited, but in a pattern now common in generative AI studies, the biggest performance gains came among the less highly skilled in their workforce. This makes intuitive sense: large language models are best understood as excellent regurgitators and summarisers of existing, public-domain human knowledge. The closer one’s own knowledge already is to that limit, the smaller the benefit from using them.
There was one catch: on a more nuanced task, which involved analysing quantitative evidence only after a careful reading of qualitative materials, AI-assisted consultants fared worse: GPT missed the subtleties. But two groups of participants bucked that trend. The first — termed “cyborgs” by the authors — intertwined with the AI, constantly moulding, checking and refining its responses, while the second — “centaurs” — divided labour, handing off more AI-suited subtasks while focusing on their own areas of expertise.

His conclusions

First, regulation will be key. Online freelancing is about as unregulated a labour market as you will find. Without protections, even knowledge workers are in trouble. Second, the more multi-faceted the role, the less risk of complete automation. The gig-worker model of performing one task for multiple clients — copywriting or logo design, for example — is especially exposed. And third, getting the most out of these tools, while avoiding their pitfalls, requires treating them as an extension of ourselves, checking their outputs as we would our own. They are not separate, infallible assistants to whom we can defer or hand over responsibility.

He has a tweet thread here.  

2. Interest rates on 10 year Treasury bonds touched 5% in the US for the first time since 2007. Investment grade corporate bonds now yield 6%, and junk bonds demand nearly 10%. Small businesses are paying almost 10% for short-term loans, up from 4.1% in mid-2020.

The rising rates has also stressed developing countries,

That has left the proportion of emerging and developing countries whose borrowing costs are more than 10 percentage points above those of the US at 23 per cent. That compares with less than 5 per cent in 2019, the World Bank calculates, in an indication of the stress those economies are now under. As a result, debt interest payments as a share of government revenues were at their highest level since at least 2010, according to the bank... Emerging market and middle-income countries’ average gross government debt burden is heading above 78 per cent of GDP by 2028, according to IMF forecasts, compared with just over 53 per cent a decade earlier...

The volume of foreign currency debt issued in emerging markets has slumped dramatically over the past two years as the cost of borrowing has soared. Emerging markets have issued about $360bn of foreign currency debt this year, according to Dealogic, following total issuance of $380bn in 2022. That follows issuance of between $700bn-$800bn in each of the previous three years. Issuance has been hit by a lack of demand, as investors favoured issuers with high credit ratings, and waning supply as many sovereigns with low credit ratings lost market access during the rapid increase in US rates of the past 18 months.

3. Extremely detailed map of NYC neighborhoods here

4. The Belt and Road Initiative (BRI) has transitioned from being the largest sovereign debt provider to being the largest sovereign debt collector

After lending $1.3 trillion to developing countries, mainly for big-ticket infrastructure projects, China has shifted its focus to bailing out many of those same countries from piles of debt... now the two main Chinese state banks that provided most of the infrastructure loans have reduced their new lending. Rescue loans climbed to 58 percent of China’s lending to low- and middle-income countries in 2021 from 5 percent in 2013, according to a new report from AidData, a research institute at William and Mary, a university in Williamsburg, Va., that compiles comprehensive information about Chinese development financing. “Beijing is navigating an unfamiliar and uncomfortable role — as the world’s largest official debt collector,” the institute wrote...

Much of the work for the Belt and Road Initiative has been done by Chinese construction and engineering companies, which sent thousands of engineers, heavy equipment operators and other specialists across Asia, Africa, Latin America, Eastern Europe and the Pacific... China provided the money almost entirely as loans, not grants, and the loans tended to be at adjustable interest rates. As global interest rates have soared for the past two years, poor countries have found themselves owing far higher payments to Beijing than they expected. Chinese lenders and contractors were able to build projects rapidly because the Chinese government seldom required extensive environmental studies, financial viability reviews or checks on the displacement of local populations forced to give up land. National governments of developing countries were required to guarantee repayment of loans made to their local and provincial governments. In the early years, 65 percent of the loans were made by China’s state-owned policy banks, notably the China Development Bank and the Export-Import Bank of China, AidData found.

5. Interesting trends on post-pandemic work commutes. Post-pandemic commutes have become shorter across cities

But public transport usage has shrunk faster than personal vehicle use, and the decline has been very steep.
And these trends carry further manifestations of widening inequality. Their nature of work means more educated workers can at least partially work from home, which in turn lowers their commute times. And this is reflected in the sharp reduction in commute times of the more educated workers. 
6. The much welcome trend of broadening of India's equity base which is attenuating the impact of foreign capital exits,

October witnessed the highest ever monthly SIP inflow of Rs 16,928 crore. While that takes the seven-month aggregate in the current fiscal to Rs 107,240 crore, it follows an inflow of Rs 155,972 crore in FY’23 and Rs 124,566 crore in FY’22... In October, for example, the FPIs pulled out a net of Rs 28,891 crore from Indian equities between October 1 and November 9. But its impact on the benchmark Sensex has been just 1.5 per cent. Contrastingly, an FPI outflow of Rs 10,836 crore in August 2011 led to the Sensex fall of over 8 per cent. Even in October 2018, an FPI outflow of Rs 28,921 crore led to a Sensex fall of nearly 5 per cent. Why this difference? While the domestic institutional flow led by mutual funds exceeded the FPI outflow by 21 per cent between October 1 and November 9 this year, the DII flow in August 2011 was lower than the FPI outflow by 23 per cent. In October 2018 too, the DII inflow was nearly 10 per cent lower than the FPI outflow. The DII strength has come on account of the SIP inflows. If the SIP inflow in October 2018 amounted to Rs 7,985 crore, it more than doubled to Rs 16,928 crore in October 2023. The big change has come not in large cities but in those that are ranked below 110. Their share in industry AUM (assets under management) in September 2023 — is 18.33 per cent. Ten years ago, the share of these cities in the MF industry AUM on a much smaller base was just 2.57 per cent and five years back, it was 9 per cent.

7. GMR has lined up a Rs 4000 Cr loan with a 14 year tenor at an interest cost of around 10% with a consortium of banks led by IIFCL for the construction of its 2200 Acre Bhogapuram greenfield airport at Visakhapatnam. The consortium also includes REC and PFC.

This fund raising highlights the importance of bank loans in infrastructure financing. Contrary to the conventional wisdom that bond markets should form the major share of infrastructure financing, experience from across the world show that except for US and China, bank loans are the biggest source of infrastructure financing. 

8. Top Four banking sector dominance in the US,

The four biggest US lenders grabbed almost half of all banking profits in the third quarter, highlighting their growing advantage in the new era of higher-for-longer interest rates. Earnings at JPMorgan Chase, Bank of America, Wells Fargo and Citigroup were up 23 per cent according to BankRegData, which collates quarterly reports from lenders to the Federal Deposit Insurance Corporation. Of the nation’s almost 4,400 banks, the big four made 45 per cent of the industry’s overall profits in the third quarter. That was up from 35 per cent a year ago, and well above the 10-year average of 39 per cent. By contrast profits at all other institutions dropped by an average 19 per cent in the quarter, their largest fall since the early months of the coronavirus pandemic.

And there's an implicit too big to fail subsidy enjoyed merely by being big.

But the biggest reason for the divide is the fact that the big banks, perhaps because of technological advantages or perceived safety due to their size, have not had to pay up as much to keep depositors. The big four were paying less than 2 per cent a year on accounts that paid interest in the third quarter. That compares to nearly 3 per cent average for regional banks. What is more, more than 40 per cent of the deposit accounts at the nation’s four largest banks pay no interest at all. That compares to 30 per cent for the industry overall.

9. A stunning expose of Ray Dalio and Bridgewater Associates, the world's largest hedge fund, by Rob Copeland in the NYT. Far from being a meritocracy where investment decisions were made by a team after rigorous research and vigorous debate, Bridgewater appears to be like an institution where the founder, Dalio here, took all the decisions. 

Mr. Dalio was Bridgewater and Mr. Dalio decided Bridgewater’s investments. True, there was the so-called Circle of Trust. But though more than one person may have weighed in, functionally only one investment opinion mattered at the firm’s flagship fund, employees said. There was no grand system, no artificial intelligence of any substance, no holy grail. There was just Mr. Dalio, in person, over the phone, from his yacht, or for a few weeks many summers from his villa in Spain, calling the shots...
Mr. Dalio largely oversaw Pure Alpha, the main fund, with a series of if-then rules. If one thing happened, then another would follow. For Pure Alpha, one such if-then rule was that if interest rates declined in a country, then the currency of that country would depreciate, so Pure Alpha would bet against the currencies of countries with falling interest rates. Many of the rules dealt simply with trends. They suggested that short-term moves were likely to be indicative of long-term ones and dictated following the momentum in various markets. Bridgewater’s rules gave it an unquestionable edge in the wildly successful early days, in the late 1980s and 1990s, when most people on Wall Street, from junior traders to billionaires, still believed in the value of their intuition. 

As the years passed, however, Mr. Dalio’s advantage softened, then seemingly ceased by the 2010s and into this decade. The rise of powerful computers made it easy for any trader to program rules and to trade by them. Rivals quickly matched Mr. Dalio’s discoveries, then blew past them into areas such as high-frequency trading. Mr. Dalio stuck to his historic rules... And if Bridgewater’s main hedge fund had for years fallen behind the pace of global markets, it still mostly avoided negative results, and so could fairly say it was making clients money on an absolute basis. Its growth was a testament to the firm’s marketing prowess, which had cultivated a mystique around Pure Alpha’s hands-off, rules-based approach... A newcomer to the investment team as recently as 2018 was gobsmacked to learn that the world’s biggest hedge fund’s trading was still reliant on Microsoft Excel, a decades-old software... Mr. Dalio’s grand automated system — his investment engine — wasn’t nearly as automated or mechanized as was promoted. If he wanted Bridgewater to short the U.S. dollar (as he did, unsuccessfully, for roughly a decade after the 2008 financial crisis), the trade went in. There was not a rule more important than what Mr. Dalio wanted, Mr. Dalio got...
With the hope of turning around the firm’s investment performance, members of the Circle of Trust put together a study of Mr. Dalio’s trades. They trawled deep into the Bridgewater archives for a history of Mr. Dalio’s individual investment ideas. The team ran the numbers once, then again, and again. The data had to be perfect. Then they sat down with Mr. Dalio, according to current and former employees who were present... One young employee, hands shaking, handed over the results: The study showed that Mr. Dalio had been wrong as much as he had been right. Trading on his ideas lately was often akin to a coin flip.

The group sat quietly, nervously waiting for the Bridgewater founder’s response.

Mr. Dalio picked up the piece of paper, crumpled it into a ball and tossed it.

This about how Dalio courted the Kazakh government is typical of the way Wall Street snares gullible sovereigns

In 2013, Kazakhstan began developing what was then the most expensive oil project — a giant field in the Caspian Sea — helping it grow a $77 billion sovereign wealth fund. That money would have to be invested somewhere, and Bridgewater’s client services squad put a meeting on Mr. Dalio’s calendar with Berik Otemurat, the fund’s chief, a bureaucrat who had begun his career barely 10 years earlier.

Mr. Dalio showed interest in the delegation. “What are they doing beforehand?” he asked Bridgewater’s marketing team.

His underlings answered that Mr. Otemurat would be in New York a few hours before he was due in Westport. 

“How are they getting here?” Mr. Dalio then asked.

Bridgewater had arranged for a chauffeur in a Mercedes.

“Get ’em a helicopter.”

The dramatic entrance preceded an unconventional presentation, at least compared with what Mr. Otemurat had experienced in New York. There, titans of industry, such as KKR’s co-founder Henry Kravis and Blackstone’s Stephen Schwarzman, wooed him over sea bass, caviar and an orange hazelnut Napoleon dessert loosely based on the Kazakh flag.

Mr. Dalio drew an indecipherable chart on a dry-erase board and rambled on about the nature of markets. He barely mentioned the specifics of Bridgewater’s approach, according to a person present. There was an undeniable charm — and confidence — to it all. Bridgewater’s marketing team had seen this move before. The end goal would be something other than money. So when Mr. Otemurat raised the prospect of investing $15 million in Bridgewater’s main hedge fund, the fund’s representatives shooed away the suggestion. “We don’t want a relationship with you right now,” one marketing executive said. “We’re in it for the long game.”

Inside Bridgewater, a relationship meant access. The country’s new oil field had taken more than a decade to develop, with near-constant delays. Anyone who knew how the project was proceeding could adjust bets on oil accordingly. Bridgewater’s representatives told the delegation that their firm would be happy to offer free investing advice, and Bridgewater’s team would likewise appreciate the opportunity to ask questions about industries of local expertise. Mr. Otemurat and others in his delegation seemed eager to chat. Soon enough, Bridgewater got it both ways. A few months after Mr. Otemurat’s Westport visit, the Kazakh fund asked again if it could invest in Bridgewater’s funds. This time, it dangled a sum far larger than $15 million, and Bridgewater assented, former employees said.  

Dalio comes across as a super-sized version of the local broker/fixer who has his ears on the ground and an excellent network which he uses to drive his business. 

10. Noah Smith has a simple explainer of Singapore's fabled housing policy,

The government of Singapore owns most of the land, and has a government agency called the Housing Development Board (HDB) that builds a ton of housing. It then sells this housing, mostly to first-time homebuyers (i.e. young people), at a cheap price. who are then free to resell it. The combination of the discount for first-time buyers and the government’s ability to make prices appreciate slowly and steadily over time means that HDB apartments function not just as a cheap place to live, but also as a sort of wealth-building pension system. In other words, in Singapore there’s no contradiction between using your home as a place to live in, and using it to build wealth... The government doesn’t actually sell HDB apartments to people; it sells them 99-year leases.

11. In a big boost to the bankruptcy resolution process, the Supreme Court has ruled in favour of the constitutionality of the IBC provisions on personal guarantees issued by promoters of companies. 

As of September, 2,289 insolvency applications have been filed against personal guarantors involving corporate debt of more than ₹1.64 trillion. Some of these cases involve high-profile names. Guarantors had been seeking legal protection against the automatic application of insolvency proceedings against them, claiming that it amounted to a violation of natural justice, and that resolution professionals could not play an adjudicatory role... So far, as many as 150 of the applications filed against personal guarantors have reportedly been rejected, while 282 were admitted. After the top court’s order, this ratio will likely see a far sharper skew in favour of case admissions.

Meanwhile the performance of IBC has so far fallen short of expectations, though its incentive shaping impacts may be much bigger than these numbers convey

In the first five years of its operation, the value of capital realized from all cases admitted under IBC was only 20%, implying a rather dismal haircut rate of 80% for creditors. And nearly 30% of all cases were landing up in liquidation, which was not the main aim of the law... In its latest quarterly report, the Insolvency and Bankruptcy Board of India (IBBI) reports that more than 65% of the cases under resolution have exceeded 270 days... The IBBI also noted that till June 2023, creditors have realized ₹2.92 trillion of value under various resolution plans of the initial total claims of ₹9.23 trillion. Thus, the realization ratio had improved marginally since 2021, from 20% to 32%.

12. Long read in Mint on the K-shaped economic growth.

13. iPhones are programmed to disallow an increasing share of parts being replaced or repaired.

For a decade, it was easy to get help repairing an iPhone. Cracked screens could be replaced in minutes, and broken cameras could be exchanged without a hitch. But since 2017, iPhone repairs have been a minefield. New batteries can trigger warning messages, replacement screens can disable a phone’s brightness settings, and substitute selfie cameras can malfunction. The breakdowns are an outgrowth of Apple’s practice of writing software that gives it control over iPhones even after someone has bought one. Unlike cars, which can be repaired with generic parts by auto shops and do-it-yourself mechanics, new iPhones are coded to recognize the serial numbers for original components and may malfunction if the parts are changed.

This year, seven iPhone parts can trigger issues during repairs, up from three in 2017, when the company introduced a facial recognition system to unlock the device, according to iFixit, a company that analyzes iPhone components and sells parts for do-it-yourself repairs. The rate at which parts can cause breakdowns has been rising about 20 percent a year since 2016, when only one repair caused a problem.

iFixit has analysed iPhone parts and have this excellent graphic on iPhone parts repairing issues,

This trend is part of the Apple business model, which not only leaves nothing on the table for others but also maximises its own revenues from post-sale possibilities.

The software phenomenon, which is known as parts pairing, has encouraged Apple customers to turn to its stores or authorized repair centers, which charge higher prices for parts and labor. In recent years, only approved parts and sanctioned repairs have avoided the problems. Replacing a shattered screen typically costs nearly $300, about $100 more than work done by an independent shop using a third-party screen.

To put it another way: The cost of replacing a cracked screen on a year-old iPhone 14 is nearly equivalent to the phone’s value, which Apple appraises at $430 in trade-in credit. Apple’s grip on the repairs creates an incentive for customers to spend up to $200 on device insurance, known as AppleCare, which provides free battery replacements and screen repairs. Apple collects an estimated $9 billion annually for the service.

And this practice of controlling parts repair has now become a feature in the industry, with Apple showing the way.

Using software to control repairs has become commonplace across electronics, appliances and heavy machinery as faster chips and cheaper memory turn everyday products into miniature computers. HP has used a similar practice to encourage people to buy its ink cartridges over lower-priced alternatives. Tesla has incorporated it into many cars. And John Deere has put it in farm equipment, disabling machines that aren’t fixed by company repair workers.

14. Times reports about the problem of fake reviews in the internet commerce sites. The problem is likely to get acute in the times ahead as the advances in the likes of generative AI make fake reviews easier to write and more difficult to detect. 

Fake reviews are a good example of the negative externalities created by e-commerce, which is not even anywhere near sufficiently internalised by the private e-commerce firm. There's a market failure. 

For the individuals shopping on an e-commerce site, reviews are often the most important (even the only) signal of credibility and genuineness. Purchase decisions that compare across multiple choices on the same good/service are usually made purely on the ratings score. But the platform operator's incentive is limited to deploy just enough resources to ensure that fake reviews don't swamp the platform. And this incentive too is inversely related to the size of the platform. 

This is a very good example of one more regulatory arbitrage enjoyed by internet companies - they are allowed to socialise costs while appropriating all profits.  

15. Even after all that has happened on the US-China relationship front, interest among corporate bigwigs in the US in courting China does not appear to have dimmed. FT reports that Xi Jinping was given a standing ovation at a meeting at San Francisco. 

Xi himself appeared in a conciliatory tone, reminiscent of the better days of the US-China relationship, as he courted US investments in China. He spoke of an enduring friendship between the two countries that could not be impacted by the recent turmoil. This change of tone is perhaps an indicator of the desperation in China as the economy gets squeezed by the restrictions imposed by the US. 

16. Finally, as the World Cup Final beckons, a striking feature of the Australian and English media coverage of the event has been its gracelessness, hypocrisy, dishonesty, and small-mindedness. This article in Daily Telegraph stoops as low as it can get. There are several other pieces in Daily Mail, Sydney Morning Herald etc, which reek of yellow journalism. Mark Ramprakash puts the issue in perspective, describing such attitudes as arising from jealousy and unconscious bias. 

For the large part, this is a combination of racism, colonial hangover, and inability to stomach the sudden and complete reversal of history (both in terms of controlling global cricket administration and dominating in the field). For a few years now, both off the field, and on the field and across formats, India have become dominant. In this World Cup they have raced so far ahead that there's daylight between them and the rest. 

On India's dominance in this World Cup, I think it can be traced almost completely to the rare and fortuitous confluence of all players getting into their peak form at the same time. Every one of its batsmen and bowlers have hit their respective peak form. I cannot recollect any instance in test or ODI or T20 when an Indian team have had its entire team being in form at the same time. Even having the majority of players in the peak form is rare. As an example, even as India has been a dominant test nation over the last 4-5 years, it has struggled with the form of all its batsmen and have had to rely on Rishabh Pant with the support of an ever  changing cast of one or two among the rest and lower order batsmen. 

And when a team consisting of as talented individuals as India have all their members in peak form, there's little that even the strongest opponents can do to beat them. They can only hope that they play at their best and some among the Indians have an off-day. 

No comments:

Post a Comment