Revisiting the debate on PPPs

The Economist has this critique of the calls for re-nationalisation of public services in UK. In contrast, Nina Shapiro has this to say in Naked Capitalism on public private partnerships (PPPs),

... the greater efficiency of private enterprise and the consequent cost savings of public-private partnerships. Yet, studies of their use in infrastructure investment provide little support for this presumed cost saving, and, as the critics of these partnerships argue, there is little reason to believe that they would save money. Governments can finance investment at a lower cost than private companies can, and, unlike the investments of governments, those of private concerns must earn a profit. Their capital costs for the same investment will be higher than that of the government’s, and the government must cover all of these in its payments to them. And since private enterprises can fail, and some of those partnered by governments have failed, bail out money has to be factored in as well. The government... cannot allow the disruption of essential services (such as air traffic control), and thus must bail out the firms providing them.

Private enterprises may be more efficient than governments, but for a public-private partnership to save taxpayer money, the efficiency of the private firm has to be high enough to offset the higher capital costs of its investment, as well as the cost of enforcing its contract and risk of its failure. And, here, we must remember that the cost savings of private firms can come at the expense of employees and suppliers – governments generally pay more – and at the cost of the services provided as well... Costs can be reduced in ways other than through productivity increases. Their reduction may have more to do with the reduction of services or incomes than with efficiency gains, and the profit that drives the economizing of private firms is not received by the people that actually fulfill the government contracts either...

Public-private partnerships conflate public and private interests, and in conflicts between them, the private interests win out... Private interests dominate the affairs of nations, and public-private partnerships not only reflect the influence of these interests, they also increase it. Government is run more like a business. Investments become “assets”, and citizens “customers”, officials are elected (and appointed) because of their success in business, and services decided by profitability. And since the profitability of a service depends on the relation between benefits and costs, values have to be placed on benefits to the public. The worth of these, in monetary terms, has to be determined, and this requires valuing such invaluable products as clean air and water, healthy and cultured lives.

If we were to do an evidence-based assessment of PPPs in developed countries in recent times, there is now a rich body of examples that illustrate many of the concerns raised in the above three paragraphs.

Only recently I blogged about the increasingly negative experience of water sector privatisation in UK. In the latest comes news that Ofwat, the water regulator, has found Thames Water, the country's largest water company, guilty of "unacceptable failure" to control water leakages resulting in a fine of £8.55 m. This comes on top of the company being fined £20 m by Ofwat for releasing 1.4 bn litres of sewage water between 2012 and 2014 into London's main water supply source. This comes even as the company paid out £100 m in dividends in the year to March 2017. An FT report writes,

Thames Water investors include pension and sovereign wealth funds, such as the Abu Dhabi Investment Authority and the China Investment Corporation, and since May — when Australian infrastructure bank Macquarie sold its final stake in Thames Water for £1.35bn — the Canadian pension fund Omers and the Kuwait Investment Authority. In the decade between 2006-16, Macquarie paid itself and fellow investors £1.6bn in dividends, while Thames Water was loaded with £10.6bn of debt, ran up a £260m pension deficit and paid no UK corporation tax, according to research by the Financial Times. At the same time it has in effect forced the Treasury and taxpayers to contribute to the cost of London’s new super-sewer or Thames Tideway tunnel, which started construction last year.

This comes in the backdrop of new rounds of privatisations in these countries. In the US, President Trump has initiated a proposal to privatise air-traffic controllers in the US, currently with the Federal Aviation Authority (FAA). Prisons privatisation is being widened, with the latest being the 20 year £1.5 bn contract bagged by Serco to run the largest prison in Australia, the Grafton Correctional Centre in New South Wales from 2020. 

The case for prison privatisation runs against a growing body of global experience that caution against it. In 2015, Serco was stripped off running the Mt Eden prison in Auckland, the largest prison in New Zealand, on allegations of ill-treatment and allowing "organised fight clubs" between inmates. In the UK, both Serco and G4S, the two largest prison operators globally, were found guilty of overcharging the government for electronic tagging of offenders.  

In a famous paper, Oliver Hart, Andrei Shleifer, and Robert Vishny examined the trade-off in delivering public services between improving service quality and cost reduction. Examining specifically the case of prisons, they wrote,

If contracts are incomplete, the private provider has a stronger incentive to engage in both quality improvement and cost reduction than a government employee. However, the private contractor's incentive to engage in cost reduction is typically too strong because he ignores the adverse effect on non-contractible quality... We have examined the conditions that determine the relative efficiency of in-house provision versus outside contracting of government services. Our theoretical arguments suggest that the case for in-house provision is generally stronger when non- contractible cost reductions have large deleterious effects on quality, when quality innovations are unimportant, and when corruption in government procurement is a severe problem. In contrast, the case for privatization is stronger when quality- reducing cost reductions can be controlled through contract or competition, when quality innovations are important, and when patronage and powerful unions are a severe problem inside the government...

We concluded that the case for in-house provision is very strong in such services as the conduct of foreign policy and maintenance of police and armed forces, but can also be made reasonably persuasively for prisons. In contrast, the case for pri- vatization is strong in such activities as garbage collection and weapons production, but can also be made reasonably persuasively for schools. In some other services, such as provision of health care, an analysis of the efficiency of alternative arrange- ments is a great deal more complicated. 

Eduardo Porter has the best summary of the incentive mis-alignment in the case of prisons,

Private prison operators who bid for government contracts by offering the lowest cost per inmate will most likely focus on cutting costs rather than tightening security. 

Alex Tabarrok has this critique of the Hart et al paper. His primary arguments being that cost reduction can come not just from cutting something but also from quality improvements themselves, and weak incentives (as with public bureaucracies) do not encourage improving quality. Both these arguments appear far less compelling when seen in light of experience. In developed economies, the cost reduction gains from incremental quality improvements, even in the long-run, is likely to be marginal. And, nor do stronger incentives, when coupled with incomplete contracts (on non-contractible quality) and far larger cost-reduction opportunities (by lay-offs and skimping on investments), necessarily encourage improving quality in all its dimensions.

Another less appreciated problem with such contracts that differentiate them from any other competitively procured services comes from the nature of these services. These are essential public services and it is very difficult to cancel a contract, force losses and bankruptcy of the provider, and transfer the same to another operator without unacceptable levels of disruption. The operators know this leverage and position themselves to extract the maximum leverage from this reality. In fact, even the technical challenge associated with shifting operators is non-trivial. The monopolistic dimension to such contracts come more from these considerations than the pure monopoly arguments of Econ 101.

So does this mean we should step away from PPPs? Far from it! My assessment of PPPs hinge on "the efficiency of the private firm has to be high enough to offset the higher capital costs of its investment, as well as the cost of enforcing its contract and risk of its failure". As I have written earlier, while there is scope for PPPs, they have to be chosen for the right reasons - efficiency improvements and not fiscal considerations, and for O&M and not construction-cum-management.

But I see at least three dimensions to this challenge - which sector is amenable to private service delivery, state capacity, and the nature of the private investor. Let's examine each.

The identification of activities appropriate for privatisation has to revolve around Hart and Co's analysis of private contractors incentive to ignore non-contractible quality. In case of activities like prisons and air traffic controllers, there are either too many dimensions of quality or externalities generated that, their capture in a contract poses prohibitive transaction costs and detracts from the provider's operational control. But without their inclusion, the private operators will doubtless be incentivised to skimp on investments or ignore those dimensions so as to reduce costs.

I am inclined to argue that the costs-benefits assessment makes many current PPPs less favourable in developed countries. In these countries, state capacity is strong and local governments  are reasonably capable of managing large utility systems - after all UK is the only country with a fully private water and sewerage utility. The efficiency gains from private transfers is likely to be small enough to off-set the countervailing costs. At best management contracts may be the more prudent option for these countries.

For example, in case of the air-traffic control privatisation, of which Serco is incidentally one of the big operators, there may be reasons to question the rationale. US air-traffic controllers, though with an excellent safety record, have often been blamed for persisting with antiquated technologies and not shifting to the satellite GPS-based NextGen system. Lack of adequate funding and slow-moving bureaucracy have been blamed for this, though the persistence with old technologies for this long period has to be attributed more to the former than the latter. In that case, and especially since FAA is even today funded with passenger fees (by way of a taxes on tickets) and recent safety record has been excellent, it is not clear where private providers will be able to find the resources and have the incentive to make significant investments in air-traffic control systems. In the circumstances, lay-offs will be inevitable and efficiency gains questionable. And, if experience elsewhere and across sectors is a marker, then price increases for travellers is also inevitable. After all, now while the passenger fees is collected as a tax on tickets, once privatised, it will be levied as a user fee by private negotiations with Airlines. 

In contrast, in developing countries state capacity is extremely weak. Public service delivery is, in most cases, of very poor quality - leakages, contamination, interruptions, unreliable, poor safety and so on. Most egregiously, public systems are over-staffed and with limited accountability to deliver on outcomes. The scope for efficiency improvements from private operation and management can be very high, though it often comes at a steep and unaffordable price for consumers. But this has to be traded-off with the state's weak capacity to manage even simple contracts, with the attendant risk of having lopsided contracts or its management which ends up causing private benefit at public cost. 

Finally, as to the capital, the failings with private equity investments in public services in the US are well-documented. Unlike a specialised private provider, a PE firm has less incentive to protect its industry reputation, an important ingredient to aligning incentives to ensuring quality and limiting skimping. To that extent, President Trump's proposal which mandates that the private entity should be a non-profit appears a good decision.

In other words, any decision on privatisation should preferably be done on efficiency considerations, and an examination of, among other things, the type of activity, state capacity, and nature of the capital. Unfortunately, this is easier said than done, and demands both vast experience (of actual implementation, and not mere theoretical knowledge) and exercise of very sound judgement. 

Notes on the GST roll-out

Two articles about India's impending Goods and Services Tax (GST) caught my attention.

1. Livemint explores the potential impact of GST on the country's massive informal sector. It points to the likely shift in activity from the informal to formal firms in sectors with high share of informality. 

It alludes to possible adverse impact on informal livelihoods, even as they are recovering from the effects of demonetisation. It writes, 

Many of the firms operating in this part of the economy make profits largely due to tax evasion and non-compliance with regulatory norms, which allows them to offer products at comparatively lower prices. However, in the GST-era, it will be a struggle for survival for such firms because they will be faced with taxes, lower margins and a sharp spike in the cost of compliance. Some firms in the unorganized sector may go under, while others could find their profits curtailed. To be sure, in some instances the two sets of companies cater to different customers, but there is always some overlap. And it is not just the manufacturers in the informal economy who will suffer but also the smaller dealers and wholesalers... The move to the new tax regime has the potential to cause immense disruption to the shadow economy that is the source of livelihood for many, although it is nobody’s case that firms that survive by flouting regulations and evade taxes continue to do so.

This theory of change of informality shrinking in any significant manner in response to GST goes contrary to historically observed trajectories of informality across countries. I have blogged on several occasions, in the context of the demonetisation, about the futility of trying to shrink the informal economy. See this, this, this, and this. 

Andrei Shleifer and Rafael La Porta have the most authoritative work on this,

Informality declines, although very slowly, with development. This is not to say that we oppose the structural policies such as simplification of registration or equalizing labor tax burdens across formal and informal sectors... These policies surely have desirable effects, but our reading of the evidence is that it is modernization, rather than structural policies, that shrinks informality... although avoidance of taxes and regulations is an important reason for informality, informal firms are too unproductive to thrive in the formal sector. Lowering registration costs neither brings many informal firms into the formal sector, nor unleashes economic growth... the informal economy is largely disconnected from the formal economy. Informal firms rarely transition to formality, and continue their existence, often for decades, without much growth or improvement... as countries grow and develop, the informal economy eventually shrinks and disappears. The formal economy comes to dominate economic life.

In simple terms, the policy focus should be on growing the formal sector than shrinking or nudging the informal sector. 

2. The GST will have a five-member National Anti-Profiteering Authority (NAPA) to look into cases of profiteering by firms not passing lower taxes to consumers, including power to deregister violating firms. Firstpost has a nice article which writes, 

Many other countries had... just one rate of GST across all goods and services. This naturally makes it easier to calculate the cost of inputs and taxes on them, set-offs, etc. Instead, what do we have in India? Multiple slabs with even the same product sometimes falling in different slabs. Not all the inputs going into a television set will be taxed at a same rate, the manufacturer will have to deal with multiple rates. The only person who will have a field day is the tax official and the worthies on the NAPA. Even in Malaysia, the anti-profiteering mechanism didn’t work too well and led to a large number of disputes and enormous litigation. Later the rules were eased and given up altogether after a year. The government then limited itself to appealing to businesses not to increase prices.

The point about easing regulations in response to emergent situations is instructive. It resonates with the adoption of minimum viable product approach in such complex roll-outs. Starting out with strict regulation is a strong signal of intent and can be valuable in shaping expectations and disciplining potential errant behaviours. The challenge though will be the ability to respond swiftly to the emerging scenarios. 

Unfortunately, it is here that Indian bureaucracy's decision paralysis is likely to hurt. Easing regulations will most certainly invite populist backlash with critics presenting it as favouring those profiteering. No logical analysis of the costs and benefits can take place in likely high-pitch media-mediated debates on such issues. Therefore, when faced with the choice of what is the right thing to do and the most politically correct thing to do, governments are most likely to incline to the latter. This deprives the government off one of the important levers to manage complex initiatives like  the GST. 

Weekend visualisations - urban edition!

Very good video that explains the historical reasons for the high density urban cores in European cities as against the low density cores and sprawls in American cities. Despite its skyscrapers, the urban sprawl makes New York only half as dense as Paris! 

And this is a fascinating article on jobs accessibility maps, developed by the University of Minnesota's Accessibility Observatory, which chart how far you can travel on a transportation network in different times of the day from different locations in a city and the numbers of jobs within that travel zone. The map is constructed with three different datasets - public transit maps and timings, OpenStreetMap data on pedestrian routes and walking times, and census data on job counts in areas. 

The map below captures the change in accessibility to jobs, during the 7-9 AM peak morning commute window, between 2010 and 2013. The dark green areas have access to 100% more jobs in 2013 than 2010. The yellow arrow points to a region where bus frequency was significantly increased along a particular line. The red arrow highlights a corridor with a new BRT line, with major improvements to job access right around a single station.

This is an extremely powerful tool for decision-support on planning public transit routes and urban transport infrastructure. Unfortunately, for cities in developing countries, the challenge is with the availability of good underlying data. In this case, very few cities have information about area-wise jobs data or its changes. 

But the emergence of innovative business models that incentivise private entities to collect such data does not look a very unrealistic prospect. 

The "bucketing" problem in financial markets

In recent weeks rating agencies have come for stinging rebukes from India to China and Russia. Livemint has this story examining whether rating agencies are biased against India and developing countries. 

In this context, I had blogged earlier pointing to this article by Nandini Vijayaraghavan which gets to the heart of the problem with any ratings approach - it buckets the target populations, and does ratings based on assessments across and within buckets, 

The impediment to the rating agencies upgrading India’s sovereign ratings lies in their methodologies. The sovereign rating methodology factors economic strength, institutional strength, fiscal performance, and susceptibility to event risk to assign ratings. The rating agencies categorise countries into buckets based on their size (GDP), growth, volatility of growth and per capita income, among other factors. The emphasis on per capita income is because countries with higher per capita incomes are better equipped to withstand cyclical volatility and are endowed with higher debt servicing ability. India’s low per capita income has resulted in its sovereign rating being lower than countries with higher deficits and indebtedness and lower growth prospects.

For example, in India till recently, rating agencies would not let even highly prudent, historically well managed, and resource rich sub-state entities, like say a state distribution company or a municipality, pierce the ratings accorded to their respective state governments. This deterred the best run and most credit-worthy urban local bodies and distribution utilities from accessing bond markets, with their higher cost of capital (due to the lower forced ratings), and preferred banks instead. This bears at least some part of the blame for the still-born municipal bond markets in India.

But rating agencies are not the only financial market agents to use the bucketing approach. Such bucketing produces distortions all round financial markets with profound impacts. Consider the interest surrounding MSCI's decision to include mainland Chinese equities, known as A-shares, in its benchmark emerging market equity index. This means that the $1.6 trillions of global funds that track the index will now have to buy into mainland Chinese stock market. 

Incidentally, despite mainland China being the second biggest global equity market, MSCI has limited the exposure of A-shares to just 0.73% of the index. It has clarified that any increase in weight will be dependent on corporate governance reforms in the mainland's equity markets, which had been subjected to very heavy regulatory intervention last year in response to downward volatility.  However, this small percentage is deceptive since the actual exposure of the index to Chinese companies, through those like Alibaba and Tencent floated elsewhere, is 25.3%. 

The MSCI is only the largest and most high-profile of indices that bucket emerging markets and force the bundling of these countries as asset categories. This is no small contributor to the observation that cross-border capital flows does not discriminate among emerging market economies and treat them all as a single asset class. Accordingly, even countries with strong economic fundamentals cannot escape the tyranny of asset buckets when sudden stops leads to capital flow reversals. 

In the context of the MSCI decision and the intense lobbying surrounding it, John Authers writes about the large power wielded by such indices,

Indices are not impartial or abstract constructs; they are an expression of someone’s opinion, and this should not be forgotten. And perhaps most importantly, there is something ungainly in the way such power has been outsourced to MSCI, a relatively small for-profit organisation based in New York. Such a momentous matter as the terms of trade in which capital flows between China and the rest of the world might seem more naturally to belong to democratic or governmental institutions. Either that, or this should be an issue for the market to decide, without intervention by governments, or heavy guidance from MSCI. As it is, the big multilateral organisations do not seem to be providing the leadership provided... something is not right with the way capital markets have come to operate when a small company like MSCI can tell the world’s most populous country what to do.

Indices, like rating agencies, distort the global equity markets in different ways. These distortions are best captured by the rise of passive exchange traded funds (ETFs), which form an increasingly large share of all equity assets. ETFs, which were introduced 25 years back, now track more than $3 trillion worth assets and are dominant share of the market in countries like Japan. 

For a start, such index tracking funds contribute to a Mathew Effect, making larger company scrips which are likely to be part of many indices, rise even more. Authers explains with the example of Amazon, 

But how much of the positive performance for Amazon is down to the momentum created for it by indexers? As investors switch to passive, or to the numerous different factor funds which currently hold a lot of Amazon, so the stock appreciates. The problem is that most stocks, including Amazon, are now judged as a series of factors with a series of properties, rather than as companies. Mark Lapolla of Sixth Man Research expressed this well, saying that the initial reaction to the Amazon purchase was not to be trusted and that

“there is a higher order context in play that is a critical point of understanding for fund managers. Namely, it is taking place at a time when passive and quantitative investors account for 60 per cent of equity assets under management and are estimated to drive about 90 per cent of daily trading volumes.”

On this basis, he says that Amazon

“is no longer a dynamic business, but a complex, inconstant security ‘type’ that proliferates data points used to predict the direction of its stock. In this new paradigm, cost basis and expected return have no meaning; the direction of price is all that matters.”

Finally, after looking through the latest holdings statements, he makes the devastating point that of late Amazon has been sold by the biggest active investors, whose stakes have ended up with big passive players:

— It will take some time for this acquisition to prove out and, in the meantime, unless the positive, fundamental buzz turns negative, AMZN will continue trading as quantitative “type” and not a business.

— A final note of interest: As of the Q1 13Fs, the largest, active investors have been selling their Amazon holdings to passive investors, quants, and the national banks of Norway and Switzerland.

More generally, index tracking funds, as Authers writes elsewhere, contribute to the inflation of bubbles and amplification of market volatility,

According to George Cooper, a fund manager and author of The Origin of Financial Crises, “the big beef” involved in tracking an index is that “you mech­anically lend most to the biggest borrowers, and buy the most overinflated stocks”. He draws an analogy with road safety: “Imagine a motorway where cars all benchmark their speed to everyone else. Then imagine what happens if everyone is trying to be a little faster than everyone else. They end up crashing.”

Paul Woolley, head of the London School of Economics’ Centre for the Study of Capital Market Dysfunctionality, suggests that market benchmarks inflate bubbles and should be abandoned altogether, in favour of comparing fund managers to rises or falls in gross domestic product... Most bond indices are weighted according to how much debt a company or country has issued. This means that the more indebted an issuer becomes, the bigger share it will take in the index, and the more of its debt passive funds will be required to buy. This is why many funds were led to load up on Argentine debt before its default crisis... 

But there are also concerns about the use of indexing in equities. Most indices are weighted according to market capitalisation. That means the more a company’s price grows, the more index-trackers will be required to buy of it, open­ing them up to accusations that they help to inflate bubbles. A second charge is that indexing at­tacks market efficiency. The more money passively tracking indices, the less devoted to seeking out underpriced stocks. If all money were managed passively, markets would cease to function.

In fact, it gets even worse,

Not only are the indexers powerful, but that power is concentrated in a few hands. Consolidation has left three big companies — S&P Dow Jones, FTSE Russell and MSCI — jointly providing the benchmarks for 73 per cent of US mut­ual fund assets, worth some $9.4tn. In bonds, the indices overseen by Barclays are dominant. Its bond aggregates (formerly known as the Lehman Aggregates) account for more than half of all ETF assets held in fixed income.

At one level, bucketing is a lazy approach to assessing and pricing risk. Instead of developing country-specific metrics and surveillance systems to keep track of them, rating agencies and indices initially preferred to take the easy way out by bucketing and reducing the monitoring data points. But once established, vested interests and market inertia have taken over and now come in the way of any change.

In the context of rating agencies, there is some evidence that Credit Default Swaps may be a better measures of credit-worthiness of the entity. Yang Liu and Bruce Morely used panel data from the main EU countries, US and Japan and found, 

The results indicate there is little evidence to show any relationship between the credit ratings and the sovereign CDS spreads,and the main drivers of sovereign CDS spreads are macroeconomic fundamentals which reflect the ‘health’ of the economy.

The need for caution with infrastructure costs-benefits assessments

Atif Ansar and colleagues at Oxford Business School examined infrastructure projects in China and claims to puncture the twin myths that infrastructure creates economic value and that China has a distinct advantage in its delivery. They scrutinised 95 road and rail projects worth $65 bn (2015 dollars) across China in the 1984-2008 period and have very negative conclusions,

Far from being an engine of economic growth, the typical infrastructure investment fails to deliver a positive risk-adjusted return. Moreover, China’s track record in delivering infrastructure is no better than that of rich democracies. Investing in unproductive projects results initially in a boom, as long as construction is ongoing, followed by a bust, when forecasted benefits fail to materialize and projects therefore become a drag on the economy. Where investments are debt-financed, overinvesting in unproductive projects results in the build-up of debt, monetary expansion, instability in financial markets, and economic fragility, exactly as we see in China today... China’s infrastructure investment model is not one to follow for other countries but one to avoid...

In line with global trends, in China actual infrastructure construction costs are on average 30.6 per cent higher than estimated costs, in real terms, measured from the final business case. The evidence is overwhelming that costs are systematically biased towards underestimation... China has built infrastructure at impressive speed in the past but, it appears, by trading off due consideration for quality, safety, social equity, and the environment... We estimate that cost overruns have equalled approximately one-third of China’s US$28.2 trillion debt pile... China’s M2 broad money grew by US$12.9 trillion in 2007–13, greater than the rest of the world combined... We conclude that, contrary to the conventional wisdom, infrastructure investments do not typically lead to economic growth.

The cost over-runs were as follows

Time over-runs...

And costs-benefits excess...

A few observations

1. All infrastructure projects, public or privately financed, will have cost over-runs. The more complex and long-drawn the projects, as the vast majority of them are, more likely the cost and time over-run. These cost over-runs are a combination of factors that cannot be controlled, under-estimation of the likely challenges, and "strategic misrepresentation" or "deception" associated with the interest to get the project approved in the first place. Further, the higher cost over-run with private projects may be an over-estimate. In case of private or PPP projects, the financial closure is an important milestone and is invariably done with great rigour. To this extent, the estimate that is used as a benchmark to calculate over-runs is a good measure of the project cost.

In contrast, in public financed projects, though it too has a financial closure, the original cost estimates are likely to be less reliable. For a start, since budget support is involved (both in grant allocation and loan repayments), Departments are likely to lowball estimates so as to get approvals of the Finance Ministry. Most commonly, some important last-mile components of the project are excluded to keep the estimate low enough to get approval of the Finance Ministry. Lenders too are likely to go along with less diligence given the budget guarantee. Finally, public projects are far more likely to start construction later than private ones.

2. It may be wrong to make costs-benefits assessments based on the traffic realisations and other benefits identifiable immediately after the project is commissioned. Apart from the existing traffic, diverted and induced traffic takes time to materialise. In fact, large increases in traffic, induced traffic, is likely to emerge from unpredictable developments in the vicinity of the road/rail alignments. There is a counterfactual problem in the analysis of such investments.

Consider an investor or entrepreneur deliberating a decision on where to invest on a new manufacturing unit. Transport connectivity would be among his primary considerations. Villages and towns along the alignment, which in the absence of transport connectivity, would not have been even considered as an investment decision, now become attractive location for investors. If the investment happens, the spill-overs from that decision can, over time, lead to very large induced traffic in those locations. And this would be apart from the overall economic development benefits from it. Given that there are likely to be several such investment decisions that different investors contemplate over a very long period of time, the likelihood of actual investments happening due, in large part, to the connectivity is very high.

To this extent, the authors claims of having refuted the new economic geography theorists' arguments about the importance of transport costs in investment decisions appear questionable.

3. On a purely quantitative basis, from the 95 projects analysed, the excess in costs-benefits ratio in just over a quarter of the projects increase due to an over-estimation of benefits. In any case, as mentioned earlier, comparing the traffic realisation immediately after the project completion with the traffic forecasts in the loan documents is almost a meaningless exercise if we are to credibly assess the long-term costs-benefits ratio of such large projects. Further, the major share of these projects predate this millennium and cover a period when Chinese contractors were building their capacity. To this extent, the analysis does not reflect the phenomenal infrastructure construction capacity that Chinese contractors have developed over the past decade or so. 

Update 1 (10.02.2020)

The Economist writes about the problem with BC analysis, in the context of UK's HS2 project,

The benefit-to-cost ratio (BCR) calculated for hs2, at around one, is hardly a ringing endorsement. But just as the costs of big transport projects are often underestimated, so are their long-term benefits. The extension to London’s Jubilee tube line, for instance, was approved with a bcr of less than one, but recent analysis suggests that it has been more like 1.75. And that includes only the revenues that go directly to the railway, not the economic consequences of the revival of London’s Docklands area, which the tube line made possible. The main point of hs2, similarly, is its impact on the cities and towns along its route and beyond. Boris Johnson, the prime minister, is on a mission to boost growth in northern and western areas left behind by the country’s lopsided, London-centred pattern of growth. On its own HS2 won’t make that happen, but doing so without a new railway would be tough.

Update 2 (25.10.2020)

Good summary of the pervasive nature of cost over-runs with transportation projects.

The importance of a minimum viable product in development interventions

Tim Harford's last book Messy, has this to say about Amazon's turbulent initial years,

Bezos combined a grandiose vision with the sketchiest understanding of how the vision could be achieved... Bezos was making big claims to his customers... but he didn't know how those promises were going to be kept. He trusted that they would figure something out. On eight have thoughts that these early weeks were a good time to pause and regroup, to concentrated on making sure Amazon was able to deliver on its early promises. But Bezos believed in seizing opportunities rather than pausing for breath. In Amazon's second week of business, he received an email from David Filo and Jerry Yang, the founders of Yahoo. Filo and Yang wanted to list Amazon on the Yahoo home page - would that be okay? Bezos's computer guru warned him that it would be like trying to sip from a fire hose. Bezos ignored him and accepted the offer from Yahoo... The workload was inhuman... Accepting the Yahoo offer within a fortnight of launch was characteristic of Bezos... 

In 1999, Bezos decided to start stocking kitchen equipment. In Amazon warehouses that had been designed to store, sort and dispatch books, naked carving knives were suddenly scything down the chutes and into the sorting machines. Meanwhile the company's database would be asking whether the knife was a hardback or a paperback. 1999 was also the year in which Amazon started stocking toys... When Christmas came... Amazon employees across the US bought Costco and Toys'R'Us inventory in bulk and drove it to the Amazon warehouses... The scramble was too much for Amazon systems... Products lost somewhere in the vast distribution centres would send Amazon's databases haywire. Unshipped orders clogged the chutes in the sorting facility, each blockage spawning half a dozen further delays. Internally, the company was on its knees: it launched a 'Save Santa' campaign and Amazon staff members were booked into hotels near warehouses (two to a room) and didn't go home for a fortnight. As Christmas passed, 40% of the toy inventory was unsold and probably unsellable... In the summer of 2000 came the dot-com bust... There was a real possibility that Amazon wouldn't see Christmas 2001... As Brad Stone writes, "Amazon survived through a combination of conviction, improvisation, and luck".  

Harford points to the concept of OODA, coined by US Airforce colonel named John Boyd, 

OODA stands for 'Observe-Orient-Decide-Act' - or, in plain English, working out what's happening, then responding... If you could make quick decisions, that was good. If you had a strong sense of what was going on around you, that was good too. 

I am inclined to argue that this applies with even greater relevance for public policy implementation. The conventional wisdom on the implementation of new program and projects is that the leadership plans every implementation step to the last detail, allocating responsibilities and putting in place monitoring mechanisms, and then goes ahead and the executes them to perfection. Unfortunately,  this approach is unlikely to work with any complex development interventions.

All such implementations require an OODA-type approach. Figure out a reasonably good initial implementation plan, a minimum viable product (MVP). Put in place a very good war room with competent people. Respond very quickly to emergent problems and refine the implementation plan. Iterate extensively with short feedback loops and improve the implementation plan as much as possible over as short a period of time as possible. 

I have written about why this is perhaps the most effective approach in the roll-out of large infrastructure projects like a new airport or a new urban transportation project. It applies just as well to new projects in most development sectors.

The GST roll-out is a very good example of a situation where the MVP approach can be adopted. Given the complexity of responses of various stakeholders and resultant uncertainty, the best that Governments at Centre and States can do is to roll-out the platform and keep the powder dry so as to respond to emergent problems. The success or otherwise of the roll-out will critically depend on the swiftness and competence of the response. 

Weekend reading links

1. Is telecoms the latest millstone around the neck of India's battered banks? A significant share of the Rs 4.85 trillion debt owed by telecom firms is showing strains on the back of a bruising battle following the splash made by new entrant Reliance Jio with its more than six months of promotion free data and voice plan that captured 70 million consumers in no time. The interest cover of telecoms sector firms came down from 1.1 times to an alarming 0.3 times by end-December 2016.

The telcos, excluding Jio, have been lobbying the government for a bailout package that involves demands for major cuts in various statutory dues like spectrum charges and interconnect charges, as well as extending the payment schedule for spectrum from 10 to 20 years. A Group of Ministers of the government has been constituted and it met recently. Consider this,

The total revenue of Indian telcos declined for the first time since 2008-09 in the year ended 31 March 2017, thanks to Jio, according to brokerage CLSA. It fell to Rs1.88 trillion from Rs1.93 trillion the previous year... Cellular Operators Association of India's (COAI's) recommendations include reduction of the spectrum usage charges, licence fee, a five-year moratorium on deferred spectrum payment against two years at present and a 12% goods and services tax rate against the proposed 18%... In 2017-18, telcos will have to pay Rs53,000 crore as interest and Rs28,000 crore to the government for spectrum already bought. The industry’s aggregate earnings before interest, tax, depreciation and amortization (Ebitda) will come to only about Rs50,000 crore. Telecom companies paid Rs78,000 crore to the telecom department in 2016-17 alone by way of recurring licence fees and other charges, according to the Union budget of 2017-18.

And this,

Jio claims that it has been funded by 6 times as much equity as debt. By contrast, Bharti Airtel has raised 33 times more debt than equity since 2010. Not just that. Almost 70% of the incumbent trio’s gross debt is on account of what they owe the government. Thanks to these deferred payment liabilities, which are for purchase of spectrum at irrationally high prices, technology upgrades have taken a hit. Still, with Rs1.4 trillion to recover from the top seven mobile operators (excluding Jio) for airwaves, New Delhi has a stake in making sure nobody goes under.

And a measure of Reliance Jio's aggression,

Based on Trai’s data for the March quarter, analysts at Kotak Institutional Equities estimate that Reliance Jio may have paid other services providers as much as Rs1,500 crore for calls terminating in their networks (caveats/assumptions below). For perspective, Reliance Communications Ltd, with a somewhat similar subscriber base, reported gross revenues of Rs1,653.7 crore for the March quarter to Trai, which is nearly as much as what Reliance Jio spent on only one of its many expense items. Reliance Jio is estimated to have had an average active subscriber base of around 70 million in the quarter, which translates to an outgo of more than Rs70 per subscriber per month on just interconnection usage charges. Note that even after the company announced that it will start charging its subscribers, its revenue per subscriber per month is only around Rs100 for its flagship plan. So while Reliance Jio may have started charging for its services, its tariffs still appear to be way below levels where they make economic sense. Analysts at CLSA Research said in a recent note to clients that the company’s current tariffs are at a 66% discount to full tariffs. 

Two observations. One, doesn't all this irrationality - excessive spectrum bids, competitive tariff lowering, and Jio's scorched earth policy - raise questions about the effectiveness of markets? It does appear that not only too little but also too much competition can result in market failures. Two, various numbers ranging from $20-30 bn have been bandied around as the investment made by Jio, and that too in 4-5 years. These are seriously large numbers by standards of India economy. Has someone probed the veracity of these numbers?

2. Politico pokes holes at China's assumption of the "leadership" of the global climate change agenda by taking over from the US. It claims that China is the largest polluter and continues to build thermal plants both at home and in much greater numbers outside.

Three observations. One, China's biggest contribution to the global climate change agenda may have been an unwitting one - its massive industrial policy to promote renewables has dramatically driven down costs and has single-handedly made renewables achieve grid parity with thermal. In simple terms, the renewables boom across the world has been largely subsidised by China! Two, China cannot be faulted for selling thermal plants abroad since that is exactly what every other country in its place would have done, and what the US has done for decades by exporting polluting industries, selling weapons etc. Three, no country, including the US, has ever exercised genuine self-less leadership in climate change. The self-declared US "leadership" of climate change has largely been driven by selfish national interest.

3. FT has the latest in the Chinese economic engagement with Africa,

In 2000, China-Africa trade was a mere $10bn. By 2014, that had risen more than 20-fold to $220bn... Over that period, China’s foreign direct investment stocks have risen from just 2 per cent of US levels to 55 per cent, with billions of dollars of new investments being made each year. China contributes about one-sixth of all lending to Africa, according to a study by the John L Thornton China Center at the Brookings Institution. Certainly, China has been attracted by Africa’s abundant resources: oil from Angola, Nigeria and Sudan, copper from Zambia and the Democratic Republic of Congo, and uranium from Namibia. In recent months, Chinese companies appear to have made an effort to corner the market for cobalt, crucial for the production of electric car batteries, with multibillion-dollar purchases of stakes in mines in Congo, the world’s biggest producer... the emerging China-Africa relationship goes well beyond commodities.

Chinese development assistance to African countries come with one crucial difference - eschew political interference. Whereas western donors and multilaterals have sought to assist African countries with anti-poverty program assistance, the Chinese have preferred to assist with investments in infrastructure and resource extraction, which have the potential to alleviate constraints to economic growth.

4. India may be adding solar generation capacity at a rapid clip, but the capacity is not translating into actual generation. While renewables capacity rose 134% over past five years, actual generation increased by just 60%. In fact 33% fossil-fuel free capacity contributed just 20% of actual energy generated.

Inadequate evacuation capacity and reluctance of state distribution companies to buy the renewable power has been the major reasons for the lag in actual generation.

5. Harish Damodaran and R Jagannathan attribute the farm distress and the eruption of farm loan waivers across Indian states to the demonetisation. Harish points to a generalised fall in farm produce prices and explains how demonetisation hurt the farmers,

Much of the produce trading in India is cash-based and financed through a chain of mandi intermediaries, processors, input dealers and retailers. While difficult to establish, anecdotal reports suggest that this traditional agro-commercial capital was dealt a body blow by demonetisation. The collateral damage from it has been a haemorrhaging of liquidity from the markets. With the trade, which used to previously buy and stock up whenever prices fell, no longer active — it neither has the cash, nor the confidence now — the produce markets are suddenly without an important source of liquidity... One does not know how long it would take for formal finance, banks, commodity trading houses or organised retail, to fill the void left by traditional agro-commercial capital whose transactions were largely in cash. Till that happens and liquidity truly returns, the ultimate sufferer is the farmer, evidence of which is visible in mandi prices and restive hinterlands.

He points to the sharp fall in prices of onions, tomatoes, potatoes, garlic, fenugreek, and grapes. Interestingly all of these are horticulture crops which do not enjoy the protections like insurance and support prices. They also form an increasing share of the total crop production and have been found to exhibit a much higher price volatility.

Mitigating this price volatility risk falls on the government. In the absence of insurance or minimum support price, the government needs to figure out a way of addressing the challenge. 

6. As Janet Yellen begins the roll back of the Fed balance sheet, which rose from $700 bn in 2008 to $4.5 bn today, uncertainty looms large. Consider these - Fed owns 15% of US government debt, 25% of mortgages not owned by agencies like Fannie Mae, and a third of the long-term government bonds. The Fed's current plan is to gradually let these securities expire and shrink the balance sheet by half over the next 3-5 years. 

The uncertainty though is about how the markets will react. Will the gradual withdrawal of the mortgage bonds spook the markets and drive up yields, thereby hurting mortgage holders? Or will the early expiry of short-term bonds ahead of long-term bonds flatten the yield curve? No amount of theory can help answer these questions. It is critically dependent on the uncertain emergent dynamics of market reaction to the Fed actions.

7. Amazon springs a surprise by gobbling up upscale grocer Whole Foods for $13.4 bn, marking its entry into brick-and-mortar retail to complement its e-commerce business. Farhad Manjoo has a nice take on what Amazon will do with Whole Foods,

Shopping for food is broken. Both the in-store experience and the many attempts at online delivery — from Webvan to Instacart to Amazon’s own service, Amazon Fresh — have failed to create the sort of seamless buying experience we enjoy with nonfood e-commerce... Mr. Bezos is a committed experimentalist. His main way of deciding what Amazon should do next is to try stuff out, see what works, and do more of that. 

So the best way to think of this deal is to look at Whole Foods as a kind of guinea pig for Amazon — a pricey, organically sourced one, perhaps, but a guinea pig all the same. Amazon almost certainly doesn’t know yet how exactly Whole Foods will fit into its long-term plans. You can expect it to make few dramatic changes to Whole Foods in the near future. Instead, Mr. Bezos and his team will most likely spend years meticulously analyzing and tinkering with how Whole Foods works. They will begin lots of experiments. When something works, they will do more of that, then more, and then even more. They may take over the world all the same — and, in the process, probably usher in big changes to large swaths of the economy, affecting everything from labor to urban planning — but they’ll do it in ways we won’t be able to predict now.

8. Nice interview of Dani Rodrik by John Judis. I agree with this characterisation of Trump,

Like most everything with Trump, I think there is a significant element of truth in the causes that he picks up. He is addressing some real grievances. But then the manner in which he addresses them is completely bonkers. So in the case of Germany, I do think Germany is the world’s greatest mercantilist power right now. It used to be China. China’s surplus has gone down in recent years, but Germany’s trade surplus is almost 9 percent of GDP. And they are essentially exporting deflation and unemployment to the rest of the world... it is not a trade problem. It is a macro-economic problem. The solution is to get German consumers to spend more and save less and the German state to spend more and to increase German wages. It is not the trade policies of the US or any other country that is going to be able to address this issue. It is similar to the way Trump has picked up grievances about how trade agreements have operated in the United States. These agreements have created loses, and grievances that have not been addressed, and I think there is a lot of truth to those kind of things, but I don’t think he has any realistic way of dealing with those things.

He outlines three rebalancing requirements for the world economy,

One is moving from benefiting capital to benefiting labor. I think our current system disproportionately benefits capital and our mobile professional class, and labor disproportionately has to bear the cost. And there are all sets of implications as to who sits at the bargaining table when treaties are negotiated and signed, who bears the risk of financial crises, who has to bear tax increases, and who gets subsidies. There are all kinds of distributional costs that are created because of this bias toward capital..

The second area of rebalancing is from an excessive focus on global governance to a focus on national governance. Our intellectual and policy elites believe that our global problems originate for a lack of global agreements and that we need more global agreements. But most of our economic problems originate from the problems in local and national governance. If national economies were run properly, they could generate full employment, they could generate satisfactory social bargains and good distributive outcomes; and they could generate an open and healthy world economy as well... the problem now is not that we are insufficiently globally minded, but that we are insufficiently inclined to pursue the national interest in any broad, inclusive sense... if you have well functioning markets, you need to embed them in institutions of governance. Markets aren’t just created on their own... We have national political systems that provide stabilization, regulation, and legitimation. Now what happens when your markets are global? Who is going to provide those supporting institutions?... Nationally we have democratic institutions for deciding who benefits from markets and how resources and income are to be distributed. Internationally, all we have are tool shops and arrangements whereby trade lawyers and technocrats decide on a global agenda without any of the legitimacy or authority that you have at the national level... In fact, there is no other single global reform that would produce larger overall economic benefits than having more workers from poorer nations come and work, for a temporary period, in rich country markets.

The third area for rebalancing is that in negotiating trade agreements, we should focus on areas that have first order economic benefits rather than second or third order. When tariffs are already very small, you do not generate a lot of economic benefits by bringing them down further. When you restrict governments’ ability to regulate capital flows and patent/copyright rules, or when you create special legal regimes for investors, you do not necessarily improve the functioning of our economies. In all these areas, global agreements generate large distributional effects — large gains for exporters, banks or investors, but also large losses to rest of society – and small net benefits, if any at all. In other words, past agreements addressing trade and financial globalization have already eked out most of the big efficiency gains. Pushing trade and financial globalization further produces tiny, if not negative, net gains. One major unexplored area of globalization where barriers are still very large is labor mobility.

The importance of a generous social safety net as a pre-requisite for trade and financial liberalisation is underlined,

At some point, the United States could have done what Europe did in an earlier stage in this history when Europe became an open economy. That is to erect very generous social insurance and safety nets. The kind of insecurities and anxiety that openness to trade creates can be compensated or neutralized by having extensive social policies, and that’s what Europe managed to do. Europe is much more open to trade in the United States. Yet to this day, trade remains uncontroversial in Europe. When you look at populism in Europe, it’s not about trade at all. It is about other things, it is about immigrants going to reduce the welfare state... In Britain, the Brexiteers wanted to leave Europe in part so that they could pursue free trade policies unencumbered by Brussels. The issue of trade and import competition was largely neutralized as a political issue in Europe by the tradeoff of a generous welfare state. When the United States became an open economy in the ‘80s and ‘90s, it largely went the other way. We didn’t try to erect a stronger safety net. If anything, the safety net was allowed to erode. That I think was a major wrong turn and we are paying the price for that now.

And about saving capitalism itself,

I think the change comes because the mainstream panics, and they come to feel that something has to be done. That’s how capitalism has changed throughout its history. If you want to be optimistic, the good news is that capitalism has always reinvented itself. Look at the New Deal, look at the rise of the welfare state. These were things that were done to stave off panic or revolution or political upheaval. I don’t want to overdramatize but I think in some ways we are at the cusp of a similar kind of process. You have the populists at the gate, and the centrist political figures and the powers behind them are looking for ways of maintaining the system, and I think they realize they need to make adjustments.

8. Finally on the farm loan waivers. Business Standard estimates the total farm loan waiver  bill till 2019 elections could be Rs 3.1 trillion ($49 bn) or 2.6% of GDP! The report puts this in perspective,

A waiver of this scale could pay for the 2017 rural roads budget 16 times over or pay for 443,000 warehouses or increase India’s irrigation potential by 55% more than the achievements of the last 60 years.

The loan waivers in UP and Maharashtra are confined to small and marginal farmers, those with less than 2 hectares. But of the 32.8 million such farmers in the eight states where there is demand for such waivers, only 10.6 m have access to institutional credit. This makes the efficacy of loan waiver in alleviating farm distress questionable.

Cities facts of the day

FT has a series on the future of cities. On the economic importance of cities,

The Chicago Council on Global Affairs places 42 global cities among the world’s 100 largest economies... The independent city state is extremely rare: Singapore is today’s most significant example. Even so, some city-regions dominate the economies of their countries: Seoul and Incheon, together, generate 47 per cent of South Korea’s gross domestic product; Rotterdam and Amsterdam, together, generate 40 per cent of the GDP of the Netherlands; Tokyo generates 34 per cent of Japan’s GDP; and London produces 32 per cent of the UK’s.

On the manufacturing squeeze faced by cities,

A 2015 report for the Greater London Authority found that 600 hectares of industrial land was lost in London in the previous seven years... The Greater London Authority says there were 7,420 manufacturing businesses with employees in 2016, or 24,820 including sole traders. Of these, 155 employ more than 100 people... only Ford employs more than 1,000 people in London... New York City had 1m manufacturing jobs in 1950, down to 76,000 by 2010, according to the New York City Economic Development Corporation (NYCEDC). From the 1990s, the city lost 5,000 manufacturing jobs a year but this levelled off and after 2010 rose slightly.

Gentrification and the hollowing out of the urban cores off the less well off should constitute the biggest threat to sustaining the urban growth,

In his latest book, The New Urban Crisis, Richard Florida bemoans the divides within the “winner takes all” super-cities of the 21st century... Soaring property values are turning the west’s largest metropolises into walled-off playgrounds of cosmopolitan elites. Mr Florida once celebrated the rise of the creative classes. Now he worries about the backlash of the uncreatives... Perhaps the starkest divide in the Brexit referendum and the US presidential poll of 2016 was between big-city voters and those in the suburbs, smaller towns and countryside. Just as London voted to remain in the EU, so New York opted overwhelmingly for Hillary Clinton, the defeated Democrat... More than half of Moscow’s voters rejected Vladimir Putin in 2012. The Russian president was still re-elected by a landslide. Fewer than one in 10 Parisians voted for Marine Le Pen in the second round of the French presidential election last month, against a third of the nation. Similar gaps exist between Istanbul and the rest of Recep Tayyip Erdogan’s Turkey — and so on.

And even Chinese cities are facing the gentrification problem, with first time buyers being priced out of the big cities.

China's bridges to growth?

From a NYT essay on China's bridge building frenzy,

“The amount of high bridge construction in China is just insane,” said Eric Sakowski, an American bridge enthusiast who runs a website on the world’s highest bridges. “China’s opening, say, 50 high bridges a year, and the whole of the rest of the world combined might be opening 10.” Of the world’s 100 highest bridges, 81 are in China, including some unfinished ones, according to Mr. Sakowski’s data... In 2016 alone, China added 26,100 bridges on roads, including 363 “extra large” ones with an average length of about a mile, government figures show... 

The vertiginous Duge Beipan River Bridge, the world’s highest, vaults a 1,853-foot-deep chasm in southwest China. On the Aizhai Bridge, drivers shoot out of a tunnel to cross a 1,165-foot-deep gorge and then whiz straight into another tunnel. The Qinglong railway bridge carries high-speed trains over a graceful arch 968 feet above the Beipan River in Guizhou Province... China also has the world’s longest bridge, the 102-mile Danyang-Kunshan Grand Bridge, a high-speed rail viaduct running parallel to the Yangtze River, and is nearing completion of the world’s longest sea bridge, a 14-mile cable-stay bridge skimming across the Pearl River Delta, part of a 22-mile bridge and tunnel crossing that connects Hong Kong and Macau with mainland China. 

These projects face apparently insurmountable financing challenges,

The projects are often financed by loans from state-owned banks to companies owned by local governments, which collect tolls to repay the loans. But on many routes in less populous inland regions, tolls are not keeping pace with the costs, setting off a spiral of mounting debt and rising expenses. The Chinese government estimated that expressways nationwide lost $47 billion in 2015, more than double the loss in 2014.

These are clearly massive challenges and for sure there has been a bridge building frenzy with all its excesses and corruption. Further, the costs-benefits assessment for many of these bridges may not be favourable. 

But I am not too concerned by these. In such infrastructure projects, especially in large countries with huge economic growth potential, it is not at all a bad idea to back the "build and they will come" approach. As the article itself acknowledges, the high-speed rail and Pudong skyscrapers were built on that and their benefits are now being fully realised. And the same has been happening to the many much derided ghost towns.

We live in a world of compressed memories, with even academic accounts of costs-benefits assessments being squeezed into time. But these costs-benefits assessments may have to be done with a far longer time horizons, over several decades. It may not be possible to anticipate many of the emergent developments that add to the value of these investments. For example, a new factory comes to a town adjacent to the bridge/highway, which in turn triggers the large-scale transformation of the region. Also, the vast majority of these investments may have to be funded by governments. For many of them, given that these are completely virgin routes, there is likely to be limited diverted and induced traffic for years after construction. It may be possible only after several years, maybe decades, to realise a part of the capital expenditures by monetising them. Some of them may never recover the investments. 

The interstate highway system in the US is the best counterfactual. If the standard costs-benefits analysis and tolling considerations were taken into account, the vast majority of the stretches in the system would have been considered unviable and there would have been no interstate system. Fortunately, the leadership at the time was far-sighted and not hamstrung by such narrow considerations and the entire system was constructed with public finance. For decades after construction, vast swathes had low traffic density. Today, it is acknowledged as an important contributor to US economic development. What's more, parts of it are being tolled or monetised.

It is also for this reason that the debt overhang of Chinese public sector banks should be viewed differently from conventional bad loan problems. To the extent that a major share of these loans finance infrastructure projects which ought to have been financed by the government, the public sector bank loans should be considered as a proxy for public finance. In the worst case scenarios, the low government debt to GDP gives the government in Beijing sufficient space to accommodate these loans. None of this is to condone the undoubted inefficiency and corruption associated with this approach to building public infrastructure. 

Business concentration and labour share of incomes

Tim Harford points to the work of David Autor, David Dorn, Lawrence Katz, Christina Patterson, and John Van Reenen which finds evidence that the declining share of labour in national output has been caused by business concentration across sectors. They analyse micro-panel data on 676 industries from the US Economic census since 1982 and international sources and find the causal chain running into the rise of "superstar firms",

If globalization or technological changes advantage the most productive firms in each industry, product market concentration will rise as industries become increasingly dominated by superstar firms with high profits and a low share of labor in firm value-added and sales. As the importance of superstar firms increases, the aggregate labor share will tend to fall. Our hypothesis offers several testable predictions: industry sales will increasingly concentrate in a small number of firms; industries where concentration rises most will have the largest declines in the labor share; the fall in the labor share will be driven largely by between-firm reallocation rather than (primarily) a fall in the unweighted mean labor share within firms; the between-firm reallocation component of the fall in the labor share will be greatest in the sectors with the largest increases in market concentration; and finally, such patterns will be observed not only in U.S. firms, but also internationally. We find support for all of these predictions.

They found that while in the early 1980s, the largest four players in any given US manufacturing industry averaged 38% of sales, it had risen to 43% three decades later. In utilities and transportation, the share rose from 29% to 37%, while in retail, it rose form 14% to 30%. In the same time workers share of the economic value added declined from 66% to 60%.

About the reasons for such business concentration, they posit a few possible contributors,

One source for the change in the environment could be technological: high tech sectors and parts of retail and transportation as well have an increasingly “winner takes all” aspect. But an alternative story is that leading firms are now able to lobby better and create barriers to entry, making it more difficult for smaller firms to grow or for new firms to enter. In its pure form, this “rigged economy” view seems unlikely as a complete explanation. The industries where concentration has grown are those that have been increasing their innovation most rapidly as indicated by patents. One might be concerned that these patents are designed to thwart innovation and enshrine monopolies... A more subtle story, however, is that firms initially gain high market shares by legitimately competing on the merits of their innovations or superior efficiency. Once they have gained a commanding position, however, they use their market power to erect various barriers to entry to protect their position...

The rise of superstar firms and decline in the labor share also appears to be related to changes in the boundaries of large dominant employers with such firms increasingly using domestic outsourcing to contracting firms, temporary help agencies, and independent contractors and freelancers for a wider range of activities previously done in-house, including janitorial work, food services, logistics, and clerical work. This fissuring of the workplace can directly reduce the labor share by saving on the wage premia (firm effects) typically paid by large high-wage employers to ordinary workers and by reducing the bargaining power of both in-house and outsourced workers in occupations subject to outsourcing threats and increased labor market competition. 

Business concentration has other implications. An OECD study found that the "productivity gap between the most productive firms and the rest is growing". It also squares up with the work of Jason Furman and Peter Orzag who find that the widening inequality in the US is "driven more by a widening gap in the average earnings of workers in different companies than by a widening gap between pay checks inside individual businesses", and that too driven by top-tier firms in healthcare, finance, and information technology.