The challenges with exercising good judgement

I have blogged earlier about decision-making as essentially an exercise in good judgement.

Such judgements face formidable challenges. For a start, most often agents face incomplete information and have inadequate experience. After all good judgement is about inductive reasoning - draw on data points or experiences and draw generalisable inferences. Second, agents are captives of their own prejudices and preferences, which brings in bias to their judgements. Such biases can offset even experience and information.

Finally, such judgements, even when processed based on comprehensive information and free of biases, are inherently inconsistent, with wide variances, something which Daniel Kahneman et al have described as noise. They write,

Professionals in many organizations are assigned arbitrarily to cases: appraisers in credit-rating agencies, physicians in emergency rooms, underwriters of loans and insurance, and others. Organizations expect consistency from these professionals: Identical cases should be treated similarly, if not identically. The problem is that humans are unreliable decision makers; their judgments are strongly influenced by irrelevant factors, such as their current mood, the time since their last meal, and the weather. We call the chance variability of judgments noise. It is an invisible tax on the bottom line of many companies.

Some jobs are noise-free. Clerks at a bank or a post office perform complex tasks, but they must follow strict rules that limit subjective judgment and guarantee, by design, that identical cases will be treated identically. In contrast, medical professionals, loan officers, project managers, judges, and executives all make judgment calls, which are guided by informal experience and general principles rather than by rigid rules. And if they don’t reach precisely the same answer that every other person in their role would, that’s acceptable; this is what we mean when we say that a decision is “a matter of judgment.” A firm whose employees exercise judgment does not expect decisions to be entirely free of noise. But often noise is far above the level that executives would consider tolerable—and they are completely unaware of it.

Their conclusion,

Where there is judgment, there is noise—and usually more of it than you think.

The 'missing middle' in public commentators!

What is the mark of truly independent public commentators? 

In my opinion he/she should be able to see-through the clouds of ideologies and narratives and engage with issues, their positive and negative aspects, on their substance. In other words, what matters should be the distinctive merits of each issue.

F Scott Fitzgerald, the American writer and novelist, famously said,

“The test of a first-rate intelligence is the ability to hold two opposed ideas in mind at the same time and still retain the ability to function.”

In other words, a true intellectual is one who can engage with contrasting views on an issue unbiased by ideology and narratives and with equal objectivity (to the extent possible for any human consciousness which is a captive of his/her socialisation), and form a considered view. How many such public intellectuals do we have?

Unfortunately, and especially so in such a deeply polarised world, this is far from the reality. 

Take the case of the polariser-in-chief, American President Donald Trump. It is nobody's case that he is not whimsical, flippant, and bigoted. While he stacks up a long list of failures, he also has some very notable successes. The reset of the relationship with China, which has bipartisan support in the US and likely a permanent geo-political shift, is, at least for many people, one of President Trump's signature achievements. It is a different matter that one can quibble about the way in which he has gone about this or how the cards will be played in the weeks and months ahead. But the mainstream media have been consistently critical of all Trump's actions on China.

A close analogy, as a friend says, is with our assessment of anything connected with Adolf Hitler (taking this provocative example precisely to illustrate the deep point). As a mass mobiliser and an agenda-based unifier, Hitler has very few parallels in history. None of this is to take away from the fact that he was also a butcher whose bigotry led to more than 6 million deaths. Now these are two distinct and conceptually unrelated facts. But the latter is so cognitively overwhelming and stigmatising that we are loath to even explore what made him such a successful mass mobiliser and assess any useful learnings. Even a honest discussion on the latter cannot happen without being branded as anti-semitic, just as any support of Trump risks being branded as bigoted and ostracised.

The liberals across the world have a similar scornful view of a certain kind of democratically elected leaders across the world. They are disparaged as "populists", as though reflecting the concerns of ordinary people is something to be loathed.

The liberal consensus goes something like this. You need to espouse liberal causes - American-style democracy and social mores, global citizenry, support free trade and globalisation, entrepreneurial capitalism, minimum government - and values and aspirations of the elites. A person who holds conservative or traditional personal values, or worse still any one not adhering to any of the liberal causes, is an outcaste and denigrated accordingly.   

Social psychologist Jonathan Haidt points to the six taste buds of the moral sense - care, fairness, liberty, loyalty, authority, and sanctity or purity. Modern day liberals tightly embrace the first two, and the third only so long as it concerns themselves. The last three are mostly viewed as conservative values, at the least secondary to care and fairness, with authority and sanctity being often viewed as degenerate. Accordingly, anyone who supports traditional values and practices is often disparaged and ostracised. 

What explains the inability of even intellectuals to break-free from ideologies and narratives and interrogate issues on their merits? Where have the truly independent commentators disappeared?

At what cost progress?

Excellent article in Quartz (HT: Ananth) about the decline of physical bank branches in remote areas of Scotland as banks embrace the digital revolution and the problems it creates for local communities. The article covers the spate of branch closures by RBS, a bank rescued with tax payer bailout during the financial crisis, and currently 60% owned by the UK government. 

It’s one more way this century has favored the young, urban, and digitally minded, concentrating power in capital cities (where nobody has heard of a bank van) and destabilizing communities that were already losing jobs and people. It draws a sharp contrast with a time, decades ago, when the bank branch manager mattered in Britain. “They had a place in society, they knew what was going on, they knew everybody who was worth knowing,” Derek French, who started working at National Westminster Bank in 1957 and spent the next 38 years there, said in an interview. But that system, where the local branch manager made key lending and banking decisions, began disintegrating before the internet came along, during the 1970s and 80s. As credit scoring and algorithms caught on, authority was sucked out of the provinces and concentrated in London. The business model proved cost efficient and unstoppable, not unlike the rise of internet banking now.

As their autonomy declined, branches were “de-skilled” and became little more than sales outposts. Their decor matched their new role. “The previously dark and imposing banking halls were redesigned into brightly coloured, open-plan retail spaces,” Pål Marthon Vik, a research fellow at the University of Salford, wrote in a PhD thesis titled The Demise of the Bank Branch Manager. The digital revolution supercharged the shift in algorithmic, automated authority that was already underway. To be sure, there are benefits beyond fattening banks’ bottom lines. The patriarchal branch manager was almost always a man and more likely to make loans to other men, with the same color of skin (that is, almost always white). Credit scoring makes lending more efficient, precise, and profitable.

But banking also became less personal, and the real-world connection of a banker who is embedded in a community was lost. While the old system could be unfair, people seem to trust the algorithms even less: The percentage of Brits who think banks are well run fell from 90% in 1983 to 19% in 2009. Another worry is that big data and better algorithms could exclude some parts of society from getting financing, according to Mick McAteer, co-director of the Financial Inclusion Centre, a not-for-profit policy group. Providers are getting better at identifying the most profitable and most risky consumers. While some, maybe even most, people will get better deals on loans and insurance than ever, those who can least afford it will pay more financial services, he said.

This about the changed role of banks is important,

The government’s role goes deeper when it comes to modern banking. Commercial banks operate in an unusual sphere, where they rely on government subsidies (access to central bank lending and money, and cheaper funding costs from the implicit expectation that big banks will get bailed out in emergency), yet exist as for-profit entities with an obligation to enrich shareholders... the underlying principle that banks are for-profit entities that derive much of their earnings from government support remains. (But) they’re designed to benefit private shareholders rather than society as a whole. 

The article is a great case study about the double-edged sword that is progress and development. Conventional wisdom would have it that digital banking and fintech constitutes technological advance and economic progress. It lowers transaction costs, access barriers, and expands opportunities, and thereby enhances economic efficiency. But amidst all these benefits, its social costs most often get glossed over.

The biggest problem with such progress is that while its benefits accrue disproportionately to the non-poor, its costs are borne disproportionately by the poor (and those living in backward areas). Like with other common signatures of progress - farm mechanisation, e-commerce, cross-border trade, immigration, globalisation, and so on - the asymmetry in their impact is the big challenge to overcome. Given that markets will not address these (at least in the short-run, and in the long-run, as Keyenes said, we are all dead), it is essential that public policy assume an important role.

While there may be specific public policy levers available for each problem, the political economy challenges associated with redistribution and other mitigating actions make them difficult to implement. In the circumstances, perhaps the only practical response may be to have a reasonable enough social safety net that cushions the vulnerable from the inevitable adverse effects of such trends that accompany development and progress. 

More on monetary policy transmission challenge in India

The RBI's rate cut has again raised the issue of monetary policy transmission in India. The same story will repeat - RBI will lower rate and hector banks to pass on by lower lending rates, and the latter will tend to look the other way. 

As Aparna Iyer has written in Livemint, "What good is lowering the price of money if its availability remains doubtful." In fact, far from declining, for a variety of reasons, GSec yields have been climbing unabated. While the central bank has been on the accommodation path, the transmission has remained muted.

Ananth writes about the cost of capital for banks. The prevailing Marginal Cost of Funding based Lending Rate (MCLR) regime is essentially an administered, cost-plus floor rate, and that too one which does not take into account the true cost of capital. As Ananth, says, several cost drivers do not change in the short-run. So he suggests dispensing with administered rates and allowing banks to set their own rates based on their true cost of capital, as reflected in their actual Net Interest Margins (NIMs). In simple terms, the MCLR is neither marginal nor a true cost, and is for all practical purposes a fully administered price. 

Indira Rajaraman writes about the centrality of the GSec yields and also makes the point about how administered savings schemes are a dampener on monetary transmission. She also points to the fact about state governments exiting borrowing from the higher cost National Small Savings Fund (NSSF) and shifting towards the regular financial markets, and the resultant crowding-out effect and the knock-on effect of keeping GSecs high. 

I had blogged earlier about how the nature of the liabilities of Indian banks comes in the way of monetary transmission. Even by the standards of other developing countries, Indian banks are excessively, almost completely dependent on deposits. As a share of liabilities, Indian banks' dependence on deposits is striking even when compared to other developing countries.

This reliance makes them vulnerable on both sides. On the one hand they need to keep deposit rates high to compete with administered savings rates schemes. And on the other hand, given that bulk of deposits are time deposits, this cost of capital is inherently inflexible to accommodate repo rate cuts. Exacerbating the problem, deposit growth rates have remained lower than credit growth rate. In simple terms, as long as Indian banks remain excessively reliant on deposits, their ability to transmit repo rate cuts down to borrowers will be constrained. 

Relaxing the constraints on monetary transmission is a major structural reform, one that involves pulling multiple levers. Dispensing off with administered prices and shifting to market determined rate is easier said than done. It requires dismantling the dual-price market for savings, diversifying the financing sources of banks and lowering the excessive dependence on deposits, and allowing banks to set rates at their true marginal cost of capital. But the short-term ripple effects of any action to address these can be significant and uncertain and can disrupt the market in unanticipated ways. Then there are the political economy constraints. 

A prudent compromise may be required in this regard. For example, on the dismantling of the dual-price market of administered small savings schemes, it will have to be done in a phased manner. A good place to start may be to target these savings schemes to only certain lower categories of employees and with lower upper savings limits. Another choice may be to ring-fence the incremental "subsidy" support from the budget itself, and have that amount transferred directly. 

As an aside, this is an example of how academic research on India, both by Indians and by outsiders, is so skewed towards marginal microeconomic issues. I am not aware of a rigorously researched empirical study that examines the different factors hindering monetary transmission in India. For example, what explains the excessive dependence on deposits for Indian banks, even compared to its developing country peers? An empirical study that examined the contributors and dynamics of monetary policy transmission (or its depression) would have been a good example of truly evidence-based policy making.

Promise of e-commerce in enabling access to markets for rural entrepreneurs

While the mainstream discussion on e-commerce had centred around the consumption side, its real development impact may lie on the production side. In particular, I am intrigued by the possibility of e-commerce platforms being a critical market access linkage provider for entrepreneurs and producers in the rural areas. 

In terms of the scale of rural penetration of e-commerce, there is no comparison to what is happening in China,

The number of Taobao Villages, defined by AliResearch as administrative village with e-tailers clustering and total annual e-commerce transaction volume of more than RMB 10 million (about US$1.5 million), as well as at least 10 percent of village households actively engaging in e-commerce or at least 100 active online shops operated by villagers, primarily with the use of Taobao.com and Tmall.com marketplace – increased from 212 in 2014 to 2,118 in 2017 and to 3,202 in 2018. Most Taobao Villages are in coastal areas and show significant trends of clustering... Development of Taobao Villages shows a trend of clustering, and new Taobao Villages tend to emerge next to existing Taobao Villages... According to Ali Research, the clustering trend results from a developed industrial base in old Taobao Villages, attractiveness from e-commerce success in existing nearby Taobao Villages, and the rapid development of an e-commerce service ecosystem. The number of active online shops in Taobao Villages increased from 70,000 in 2014 to 660,000 in 2018. Clothing, furniture, and shoes were the top three purchased products from Taobao Villages. Luggage and leather goods, auto accessories, toys, and bedding products are also very popular.

Its impact on the development side,

A fast-growing body of case studies on e-commerce in rural China focuses on Taobao Villages. Numerous cases show the prosperity of Taobao Villages and that people gain wealth and have better lives through participating in e-commerce. Case studies, such as Shaji in Jiangsu province and Caoxian in Shandong province, show many young and talented people, including women, have returned to their hometown in rural areas, earning income similar to or higher than they were as migrant workers in the cities and at the same time enjoying family life with their elderly and children. Many have become leaders of e-commerce in their home villages and are role models for their fellow villagers. Case studies in Mengjin in Henan province show people are enriched by access to new markets through online platforms for traditional cultural products such as peony painting and Tang tri-color ceramics. Case studies in Xifeng in Guizhou province show households received higher farmgate price for kiwis and therefore have an incentive to increase the production through online sales to domestic as well as European markets. Many cases, including poverty-stricken counties in remote and mountainous areas, show that access to an online market allows people in rural areas to enjoy the convenience, variety, and similarly low prices that are enjoyed by people in big cities... 

The Chinese Taobao Village development report (Nanjing University and AliResearch 2018) examines the patterns of formation of the Taobao Villages and their evolution, drawing from extensive fieldwork. The report shows that, while most Taobao Villages on the coast developed spontaneously, often led by a couple of first-mover return migrants and followed by fellow villages, government support, often through subsidized service provision by experienced e-commerce service firms, has become a major force for the incubation of inland Taobao Villages.

There is a stark contrast in the entrepreneurship and successes of the Chinese e-commerce providers with that of their Indian counterparts. Even government initiated large flagship initiatives like the agriculture national market place (e-NAM) have had limited success. In fact, I am not aware of even one Indian manufacturer of any kind who has become reasonably big by growing from their rural base, whereas such Township and Village Enterprises (TVEs) were the stars of China's growth.  

This is surprising since efforts to enable wider market access to local farm and non-farm produce has been a major part of rural development programs of the central and state governments for decades. It was then thought that there was a large urban market for, say, indigenous crafts and organic produce, and the major constraint was a means to connect rural sellers and urban buyers, or matching frictions. Now that the e-commerce marketplaces have solved that matching problem, one would have thought that the market ought to have taken off. So clearly there are other binding constraints on the flourishing of rural e-commerce. I am not aware of any research exploring these in the Indian context. 

One can understand the challenges associated with agriculture produce given the concerns of quality and standards, reliability of delivery, transportation times, settlement of payments etc. While the same applies just as well to non-farm production, they are perhaps less daunting. For example, unlike agriculture produce, perishability is not a problem, and even quality and standards can be defined reasonably well in case of many products.

There are also issues of broadband connectivity, use of smartphones, transportation logistics etc which are critical. We may also need to question whether there is availability of large enough concentrated demand pockets so as to leverage economies of scale with production, marketing, and logistics. Maybe there is a need for some form of receivables credit provision. Or maybe Indian consumers, socialised by distrust, prefer to buy such products through physical retail. 

I am not sure what public policy can do to promote this market. One may be to work with the industry and define clear and simple standards for the main manufacturing products in rural areas, facilitate access to standards accreditation, enable receivables credit financing, and encourage e-commerce players to venture into rural areas with targeted incentives in this regard. There may be a role for the National e-Government market, GEM, place to play some role in catalysing this - perhaps even promoting strategic public procurements from rural areas etc.

A more promising and perhaps effective and quicker approach would be the emergence of some  deep-pocketed and visionary market participant like Alibaba. It would need to make the large upfront fixed investments, including hiring local marketing personnel and enlisting rural partners, required to dramatically expand the market and capture it. None of the incumbents in India like Flipkart or Amazon are likely to take a leaf out of Alibaba's book play this game. Jio?

If a private provider like Jio is interested in pursuing this approach, I don't think it is a bad idea to support that player, even at at the cost of contributing to massive private benefits at tax payer expense. It is just that the total long-term social value creation and externalities would far-exceed the private benefits. 

Weekend reading links

1. Very good survey in The Economist on the future of the aviation industry. This in particular is a great summary of the science and the commerce behind the manufacturing process of a passenger aircraft.

Each of the finished planes sits at the apex of a system of supply chains which fans out across the world, bringing 3.5m components together into a single product. An A 350’s airframe is composed of seven sections. Three are assembled into the fuselage, two being made at another site in France and the third in Germany. The two wings are made in Britain, then transferred to Germany to be finished. The tail fin and the horizontal-stabiliser assembly are made in Spain. All of these pieces are flown to Toulouse in special transport aircraft called Belugas—after the whale, which they resemble, rather than the sturgeon, which they do not. They are made, mostly, of carbon-fibre-reinforced plastics (CFRPs). These are composite materials that cannot be riveted in the way metal is because of the damage this causes to the fibres. They are therefore held together by lock-bolts inserted through 10,000 specially drilled holes in the flanges where the sections overlap.

Connecting the sections involves fitting them together, drilling the holes (a process less damaging than riveting), unfitting them, cleaning the holes and surrounding areas of debris, applying a sealant to the flanges, fitting the pieces back together again and then inserting the lock-bolts. At this point the myriad cables which keep a modern aircraft flying, and which have been pre-fitted into the airframe sections, are linked up. Before their final bonding, however, the fuselage sections have had what are known as “monuments” installed. These are bits of equipment—galleys, crews’ quarters and so on—that would be too big to carry through the cabin doors later. Afterwards, the rest of the fitting-out is done, the plane is painted in the customer’s livery and the crucial finishing touches, a pair of engines, one under each wing, are added. The whole process takes about a month.

2. This summarises the giant leap made by the mobile phone market in India,

In 2014, the cost of one GB of mobile data was ₹270. Now, it is ₹10 per GB. As a result, mobile data consumption has soared. In late-2014, an average user on Airtel’s network (India’s largest telecom operator back then) used 622 megabytes (MB) of data in a month. By late-2018, the number of users had tripled, but, despite a broader base, average data usage stood at 10GB a month.

The Economist has a good briefing article on how entertainment is driving the penetration and use of internet, and India leads this trend - "internet is the leisure economy of the world's poor"or "timepass". 

“Timepass” is the essence of the internet. The vast majority of the top 25 apps by revenue in both Google’s and Apple’s app stores are games (and both companies announced new paid gaming services this year). Tencent became one of China’s internet giants because of games. Facebook grew into the world’s sixth-most valuable company by giving people a place to “do timepass”. YouTube is the gateway to several lifetimes’ worth of timepass. The fastest-growing new apps of recent years have all been aimed at timepass: Fortnite, WhatsApp, Instagram, Snapchat. TikTok, which consists of 15-second videos, is timepass in its essence, made by bored kids in mofussil towns who have found vast audiences by doing silly things.

In fact, India has the cheapest mobile internet in the world, nearly 48 times cheaper than the US to download a GB of mobile data!

3. IndiaSpend has a good series on informalisation of labour market in India. This and this covers the trend of contractualisation whereby firms prefer to contract than recruit workers, allowing them to skimp on benefits and statutory payments and keep wage costs down. This covers the fate of the 1.5 million people employed by ride-hailing providers,

Our interviews with workers... revealed that many of them were migrants to the city and spent long hours on the job to earn incentives to be able to send savings back home or make their existence in their adoptive city a bit more comfortable. They had little or no employment benefits such as insurance, and complained that their incomes were declining... All drivers for app-based cab companies complained about falling earnings due to increased competition--more and more cabs are plying every day.

4. The fastest growing retail activities in UK are the classic non-tradeables,

The article itself is a very good account of the disappearance of retail shops,

Technology will continue to transform shopping, and there are some good arguments for embracing this. Why shouldn’t people have easier lives, if the fridge tells you when it is on the last yoghurt and the supermarket delivers an hour after you’ve ordered on its website? The reason isn’t obvious: it will reveal itself only slowly, as the gift of sociability that shops give for free is withdrawn... The disappearance of shops, where the commercial exchange can be encased in a social one, will be something of a disaster if nothing of equal social use takes their place.

5. Jarrod Kimber revives the dying art of great cricket writing with this beautiful article.

6. Hubert Horan nails the Uber story, taking on the sustainability of its business model, cost structure, and commercials. This is a good summary,

An examination of Uber’s economics suggests that it has no hope of ever earning sustainable urban car service profits in competitive markets. Its costs are simply much higher than the market is willing to pay, as its nine years of massive losses indicate. Uber not only lacks powerful competitive advantages, but it is actually less efficient than the competitors it has been driving out of business... Uber pursued a “growth at all costs” strategy financed by a staggering $20 billion in investor funding. This funding subsidized fares and service levels that could not be matched by incumbents who had to cover costs out of actual passenger fares. Uber’s massive subsidies were explicitly anticompetitive—and are ultimately unsustainable—but they made the company enormously popular with passengers who enjoyed not having to pay the full cost of their service. The resulting rapid growth was also intended to make Uber highly attractive to those segments of the investment world that believed explosive top-line growth was the only important determinant of how start-up companies should be valued. Investors focused narrow­ly on Uber’s revenue growth and only rarely considered whether the company could ever produce the profits that might someday repay the multibillion dollar subsidies... Uber’s longer-term goal was to eliminate all meaningful competition and then profit from this quasi-monopoly power... Uber’s most important innovation has been to produce staggering levels of private wealth without creating any sustainable benefits for consumers, workers, the cities they serve, or anyone else...

This is not a case of a company with a reasonably sound operating business that has managed to inflate stock market expectations a bit. This is a case of a massive valuation that has no relationship to any economic fundamentals. Uber has no competitive efficiency advantages, operates in an industry with few barriers to entry, and has lost more than $14 billion in the previous four years. But its narratives convinced most people in the media, invest­ment, and tech worlds that it is the most valuable transportation company on the planet and the second most valuable start-up IPO in U.S. history (after Facebook). Uber is the breakthrough case where the public perception of a large new company was entirely created using the types of manufactured narratives typically employed in partisan political campaigns. Narrative construction is perhaps Uber’s greatest competitive strength. The company used these techniques to completely divert attention away from the massive subsidies that were the actual drivers of its popularity and growth. It successfully framed the entire public discussion around an emotive, “us-versus-them” battle between heroic innovators and corrupt regulators who were falsely blamed for all of the industry’s historic service problems. Uber’s desired framing—that it was fighting a moral battle on behalf of technological progress and economic freedom—was uncritically ac­cepted by the mainstream business and tech industry press, who then never bothered to analyze the firm’s actual economics or its anticompetitive behavior.

And even the less worse (albeit heavily red) bottomline comes from squeezing driver pay,

If Uber drivers still received their 2015 share of each passenger dollar, Uber’s negative margins would still be in the triple digits... Starting in 2015, Uber eliminated most of the incentives it had used to attract drivers and unilaterally raised its share of passenger fares from 20 percent to 25–30 percent. Almost all of Uber’s margin improvement since 2015 is explained by this reduction of driver compensation down to minimum wage levels, not by improved efficiency. These unilateral compensation cuts resulted in a direct wealth transfer from labor to capital of over $3 billion.

7. Pramit in Livemint has a graphical summary of the ongoing controversy over Indian GDP statistics.  

8. Fascinating article on kidnapping and ransom insurance, with its 20 odd firms operating out of Lloyds of London. This captures the essence of how it is able to get kidnappers and insurance providers to work together,

"It's a one-off transaction between the family and the kidnapper, but it's a repeated interaction for the insurance market."

9. Ananth points to a very good article by Andy Mukherjee with actionable proposals (a refreshing change from the general 36000 ft ideas that oped writers typically offer) to address the liquidity squeeze that is being faced in India's shadow banking sector. While the suggestions are all logical, the problem is with getting the government too deep into solving these problems. For example, the incentive distortions likely with a government sponsored refinance facility are numerous. 

Building infrastructure in NYC

From an article on the exorbitant cost of building infrastructure in New York City,

New York intends to spend $39 million per station for the next round of accessibility retrofits, which will add elevators and ramps to make stations accessible on a step-free basis. According to figures gathered by transit researcher Alon Levy, $39 million is double what it costs to make a station accessible in London, five times what it costs in Madrid, and 15 times what it costs in Berlin.

In terms of station costs, sample this, 

Madrid: $14 million per station. The three-mile, two-station Line 9 extension was completed in 2015 for a total cost of $140 million. Los Angeles: $120 million per station. L.A. spends as much as New York to dig tunnels, but “cut and cover” station construction has kept the cost of its four-mile, three-station Purple Line extension down to $2.8 billion. Paris: $168 million per station. Adding nine miles and seven stations to Line 14 will cost the city $4.4 billion in total — a lot for a city that usually spends between $90 million and $135 million per station. New York: $425 million per station. The world’s most expensive subway line per mile, the first phase of the Second Avenue Subway cost $4.5 billion after two miles and four stations.

Among the contributors to the cost problem, this is instructive,

The MTA tries to shift the risk of cost overruns onto outside companies it contracts with, even if those overruns are caused by factors outside their control; the companies are not stupid, and they respond to this by inflating their bids for work on MTA projects in what’s known as the “MTA premium.” New York has unusual laws about contractor liability that make insurance very expensive... City Council Speaker Corey Johnson (says)... “When the projects are being negotiated, many, many times, the MTA just signs off on what the contractors put in front of them. There’s no forensic auditing or effort to see if costs have been inflated in an unscrupulous way.”

And there is also the standard problem of lack of co-ordination and attendant delays and cost over-runs, arising from the Metropolitan Transit Authority's (MTA's) lack of institutional power. This would resonate with construction in India's metropolitan cities,

When the MTA wants to build a big new shiny thing, it’s at the mercy of a lot of people and entities it doesn’t control. The MTA often needs utilities moved in order to do construction below city streets, but it lacks the city’s power to boss the utilities around, so it pays more for the same sort of utility relocations that the city might demand. It can’t force the city to make zoning changes that would aid its capital projects or its finances... things go to hell when agencies have to work together... executives overseeing the MTA’s capital construction may lack the authority to boss around its operating agencies such as Long Island Railroad (LIRR) and NYC Transit. And even when the MTA’s internal units work constructively together, they are dependent on outside agencies like Amtrak to cooperate and make rights of way available for construction in a timely manner; when they don’t, which is frequent, expensive labor time is wasted. In the case of East Side Access project, the new LIRR tracks to Grand Central must connect with existing tracks at the Harold Interlocking in Queens, which is already the busiest rail junction in America. Coordinated labor is often required, with Amtrak and LIRR employees and workers for various private contracting firms needing to be present simultaneously. But sometimes Amtrak’s crews don’t show up and everyone else sort of stands around, getting paid but not building East Side Access... In an April 2018 letter, MTA chief development officer Janno Lieber alleged that Amtrak’s failures to cooperate on the East Side Access project — for example, by canceling planned service outages and not providing its workers as scheduled — had added $340 million to ESA’s costs over four years...

Megaprojects built in New York today are subject to extensive delays and costs imposed by community input, as with the requirement that all the excavation spoils from East Side Access be carted out via tunnel to Queens, avoiding disruption in midtown Manhattan but adding significant costs and delays to the project. Those costs and delays don’t just mean a longer wait for LIRR trains to roll into Grand Central; they impose opportunity costs that mean other projects don’t get built or proposed at all. A more empowered MTA would mean fewer costs like that, but it would also have to mean a reduction in the fetishization of community input. As a public that has chosen Jane Jacobs over Robert Moses, we would have to get more comfortable with the idea that sometimes things won’t be built exactly as we want them.

Clearly New York is illustrative of the problems that contribute to cost over-runs in many cities of India. And the counter-point to NYC and MTA is perhaps London and its Transport for London (TfL).