Central Bank Communication and Financial Stability

Central Banks across the world are racing to go transparent. In United States, the Federal Reserve has raised transparency to new levels by publishing the long-term policy rate forecast and the next rate hike expectations of the individual members of the Federal Open Market Committee (FOMC). The Reserve Bank of India too has not remained unaffected by this trend as its minutes now reveal the individual opinions of its Monetary Policy Committee members.

The argument behind this transparency is that it minimizes information asymmetry and helps shape policy expectations among all market participants more effectively. However, it is far from certain that such transparency is beneficial in the long-term or any more efficient than the current strategy of relative opacity.

For example, in the instant case of the Fed, the longer-term forecasts of the individual members, albeit appropriately qualified, helps bake in expectations of longer-term policy direction among market participants. Though qualified with assumptions and conditions, this fine print is most likely to be overlooked, especially when there is a sustained period of upswing in various growth indicators. As the actions of Alan Greenspan and Co. over the first half of the past decade shows, the members are themselves likely to be blinded to the various cognitive biases that influence their decisions when the economy is either doing well or badly.

The resultant "irrational exuberance" is certain to bias the decisions of a large number of market participants. In these circumstances, the longer-term forecasts act as pro-cyclical amplifiers. In the absence of such forecasts, the market participants would atleast have been that less certain about the economic trends. And given the formidable reputation that central banks have assiduously built-up over the past few years, these forecasts are likely to carry much greater punch than other sources of information.

Therefore, such transparency enhancing forecast publications, tend to make the market participants lean more towards the wind than would have been the case in their absence. Perversely enough, a dose of uncertainty may have diminished the "irrational exuberance" and herd mentality and made market participants more guarded in their economic decisions. In simple terms, instead of improving market efficiency, the reduction of information asymmetry has the potential to lower market stability.

Update 1 (24/2/2012)

It appears that the Fed's push to reduce frictions in one area is complemented by increased frictions in other areas. Simon Johnson detects a very clear bias in the meetings held by the Fed officials on the Dodd Frank legislation,

Just on the Volcker Rule — the provision in Dodd-Frank to limit proprietary trading and other high-risk activities by megabanks — Fed board members and staff members apparently met with JPMorgan Chase 16 times, Bank of America 10 times, Goldman Sachs nine times, Barclays seven times and Morgan Stanley seven times...

Based on what is in the public domain on the Fed’s Web site, my assessment is that people opposed to sensible financial reform — including but not limited to the Volcker Rule — have had much more access to top Federal Reserve officials than people who support such reforms. More generally, it looks to me as though, even by the most generous (to the Fed) account, meetings with opponents of reform outnumber meetings with supporters of reform about 10 to 1.

Are free markets in information flows good for stability?

Chris Dillow has an interesting post where he compares the role played by information cascades in triggering off stock market sell-offs and riots. In this context, in a recent paper Klaus Adam and Albert Marcet show that even a very small information cascade can generate significant asset price volatility. Can the same model of information cascades be extended to explain rioting?

Extending the same logic to social and political systems, there could be a strong argument that the latest communication technologies and information dissemination channels, which remove information market frictions, reduce their stability. In the circumstances, a familiar debate, similar to that between advocates of laissez faire and those favoring a more nuanced acceptance of free-markets, appears inevitable.

Laissez faire advocates would welcome the proliferation of these technologies as contributing to increasing the efficiency of information flows. They would argue that it will help people make more informed decisions and thereby reduce distortions and prejudices that are commonplace in social, political and economic markets. Is this assessment correct? Is unrestricted information flows and social networking platforms an unqualified good? Do market failures in financial and economic systems carry any relevance for information markets?

There are a few observations on this.

1. Information markets are vulnerable to atleast some of the same failures that characterize financial markets. Mere availability of information does not guarantee efficiency in decision making. As is the case with financial markets, thanks to the cognitive biases of human beings, the manner in which the information is presented or made available has important bearing on their individual response.

2. Social and economic systems straddle a fine line between stability and chaos. Unrestricted information flows often end up unsettling the delicate balance in such systems and chaos ensues. Without going into the merits of whether the delicate balance was sub-optimal or inefficient, it is often the case that stability is the casualty when the information market is unshackled. This assumes importance since socio-political stability is critical for any economic growth and development. The instability that followed the break up of countries like Yugoslavia is an example of this.

3. This brings us to an issue of whether some form of information latency is desirable for social and political stability. For example, in a highly heterogeneous society, democracy and the formal norms of democratic governance cannot be readily transplanted without having in place several other institutionalized checks and balances. Conventional norms of majority rule can be destabilizing in these countries, as evidenced by the civil wars in the aftermath of democratic elections held in a few African countries in the nineties.

4. As information flows unhindered and the communication channels become more active, the probability of even small events upsetting the balance is greater. Even small, often insignificant information flows, as the work of Klaus Adam and Albert Marcet shows, have the potential to generate considerable instability. A sharp increase in the quantity and velocity of information flows, as is happening now, significantly increases the probability of such dynamics being triggered off. This increases the social or political riskiness associated with traditionally unstable societies. This also means that instability mitigating institutional systems assume much greater significance in these countries.

5. Always-on and many-to-many communication channels like social networking sites amplifies the impact of free information flows. In all respects, these channels are much more disruptive of stability that mere information flows. Further, their stability creating aspects, most often end up being crowded out by the stability disrupting aspect. After all, do we not more often come across examples of an information flood clarifying issues instead of complicating them?

None of this is an argument in favor of placing restrictions on information flows and social networking sites. Far from it. It is only a note of caution and a pointer to possible triggers that can upset the stability of social, political and economic systems. It therefore becomes important that such information technology developments be accompanied by policies that mitigate the market failures that arise out of them.

Labor market matching problems

Conventional public policy on employment generation is limited to the creation of new jobs and training to equip job seekers with requisite skills to compete in the job market. Since the former is directly related to the larger issue of economic growth, the focus of government driven employment generation programs have been largely confined to the later.

However, this approach, while a necessary requirement in any employment generation policy, may be a limited view of the dynamics of labor markets. I can think of atleast two dimensions of labor market inefficiency that keeps employment market at a sub-optimal equilibrium.

1. Matching unemployed labor to employers - At any time there are unemployed people looking for jobs and buyers of this labor searching for sellers. The problem is how to match them. This is a big challenge with un-skilled and semi-skilled service sector jobs. How do we match demand for specific jobs at a specific location and on specific terms, with sellers who meet all these requirements?

Such first level of matching immediately adds people to the workforce and reduces job-search inefficiencies. Businesses benefit by way of lower search costs for employing required labor (firms incur expenditures varying from 1-3 months salary as the cost of locating the right labor pool). People stay out of workforce for longer than necessary. Can public policy lower this inefficiency?

2. Optimal matching of under-employed labor - This involves matching employed people with jobs appropriate for their skill and capability level. In other words, it enables efficient matching of labor supply and demand.

Consider the case of Ramu, working with a small, single employee mom-and-pop clothes retailer in a city. After two years in the job, Ramu acquires enough skills to assume more demanding responsibilities. He can easily fit into the role of a lower manager in a shopping mall. His place can in turn be taken by a semi-skilled or even unskilled new addition to workforce, Ravi, who recently migrated from the neighbouring district in search of jobs. Everyone benefits - Ramu benefits by way of higher wages, Ravi gets employment, mom-and-pop retailer gets employee at lower cost, and the mall gets an employee with skills and experience. Most importantly, the economy benefits by way of productivity enhancing efficient matching of two people with varying skills with jobs that are most appropriate for them.

When several millions of such matching takes place, the efficiency gains are massive. It translates into the mom-and-pop shops expanding and hiring more labor, the mall increasing its sales, consumption by the new additions to the workforce adding to aggregate demand, and so on. In other words, the removal of such inefficiencies sets the stage for a virtuous circle of economic growth and job creation. How can public policy enable the removal of these inefficiencies and facilitate efficient matching of labor supply and demand?

In both these cases, the fundamental issue is a matching problem - how do we match unemployed or under-employed workers with their potential employers? Left to itself, for various reasons, the markets cannot enable efficient matching, especially in developing economies. Therefore, what role can governments play in facilitating such matching?

I had blogged earlier about the possibility of governments facilitating this by establishing and adding value to a dynamic meta-labor supply database. This would serve as a database for individual employers or placement agencies to locate job seekers who meet their requirements, thereby benefiting both sides.

I am strongly inclined towards the view that public policy has an important role to play in facilitating this matching process. The debate should be about how to achieve it without creating any major labor market incentive distortions.

The case for health insurance and government's role

What should be the role of government and the private sector in India's secondary and tertiary health care system, especially in taking care of those who cannot afford private health care?

Traditionalists see no or limited role for private sector and advocate that government hospitals should meet the requirements. Extreme liberals on the other hand advocate a dominant role for private sector. However, reality demands a much more nuanced appreciation of the health care market and the incentives and challenges facing its participants. This post will attempt to put these issues in some perspective. I will also attempt to outline a model health insurance market, applicable to legacy-free countries like India.

1. Fundamentally, government institutions, even if they expand exponentially, are in no position cover even a majority of those in need of such services. On all primary parameters - doctor to population, beds to population, diagnostic facilities to population, and so on - we lag way behind the requirements. This deficiency will persist well into the future.

It is therefore inevitable that if the government is committed to ensuring atleast access to secondary and tertiary care facilities to its under-privileged population, not only does the public facilities have to grow fast but also be complemented with rapidly expanding private healthcare facilities.

2. However any role for private hospitals raises important questions about the details of this involvement. Crucially, how should private hospitals be involved in the treatment of such cases for those below the poverty line? This question is important because of three reasons.

One, unlike other markets, that for selling and buying health care is rife with market failures (information asymmetry problems like moral hazard, adverse selection, over-treatment), behavioural biases (healthier/younger people prefer to stay uninsured, people prefer more diagnostic tests and invasive procedures) etc. Two, as I had blogged earlier, health care will be among the few markets where productivity imporvements will remain marginal even as technology continuously expands the treatment frontiers. Three, there will be a huge and persistent supply-demand mismatch in developing coutnries which will ensure that health care remains a sellers market for a long time to come.

The first and second factors, along with demographics, have been responsible for the rising health care costs and resultant health care mess in many developed countries. The third factor will only exacerbate the already inherent distortions of health care markets and will be an important factor in countries like India.

In view of all these, the nature of private sector's involvement in the provision of affordable health care becomes important. Experience from across the world shows that health insurance is the most effective strategy to manage these risks optimally and deliver affordable secondary and tertiary healthcare to citizens.

3. Assuming government's commitment to deliver affordable secondary and tertiary health care to all its citizens and the inevitable need for health insurance, the question then is one of ensuring how we can get care that delivers bang for the buck. In other words, how do we achieve the desired health care outcomes at the lowest cost.

This would obviously require leveraging both the government and private health care facilities. The most critical factor would be the design of the insurance model. How do we structure incentives such that the doctors and hospitals confine treatment to only the necessary diagnostic tests and procedures/medications, patients do not demand more than what is required, government hospitals and private hospitals complement each other, and the insurer's administration charges are kept at a minimum?

Here is a simple model of how this insurance can be structured. The real-world model could be some variant or other of this.

Bring all citizens of the country/state into a single risk pool. All those below the poverty line and all government employees, including their families, should be part of this risk pool. In an ideal world, it would be appropriate to include all citizens and usher in a mandatory health insurance model. Further, all the schemes offered in this market should be community rated - same insurance premiums for everyone in the same age group or no differentiation based on pre-existing medical conditions.

Finalize a basic bouquet of treatments that are covered in a universal and basic insurance package and is available to all citizens at a competitively arrived premium. There can be a single or preferably multiple insurers prioviding these schemes. Then there should be a variety of top-ups available on this basic package. Government departments can offer a menu of top-ups to their employees depending on their different levels. Similarly, private employers and individuals too can purchase insurance from this market.

The government could subsidize the premiums at varying levels depending on people's incomes. For example, the poorest could have their entire premiums subsidized. Similarly, for employees, a share of their salaries could be leveraged to complement the government's share of the premium. Employees would have the option to privately top-up on their government package.

The administration of the insurance model itself could be made more transparent and protocols based, so as to minimize excesses and distortions. The entire pre-authorization process can be done transparently and rigorously audited, so as to ensure that insurers/TPAs and service providers do not over- or under-treat patients. To a great extent, as competition increases, the presence of professional insurance agencies and TPAs should contribute towards minimizing these distortions.

A protocols-based referral system can be put in place so that atleast certain categories of those covered in the government financed health insurance model are treated only in government hospitals if facilities are available. The government could, to the extent of the packages fully or partially financed by it, control the empanelment process and negotiate bulk rates with drugs manufacturers and service providers, so as encourage the insurers to quote lower premiums. The presence of a large and vibrant set of government hopspitals will provided the much needed competition to keep private service providers honest.

An additional reason for large network of government health care facilities to exist is because for much of the foreseeable future government hospitals will be necessary to service the vast interiors of the country. Private sector will find such locations commercially unviable. Insurers could deliver their services by leveraging these public facilities in remote areas.

These schemes could be sold in newly established customer-friendly insurance exchanges. Such exchanges could helps customers easily compare across similar kinds of policies, besides making clear the fine-print of these policies. The regular private insurers, who would exist and sell their respective insurance schemes, would all be eligible to bid for offering these set of insurance services. Once the insurers are designated and premiums for the basic package for both poor and different categories of government employees defined, the top-ups may be left to the markets to decide. This will ensure that in the process of rectifying market failures, governments do not end up distorting incentives wholesale and affecting market efficiency.

If required, it may even be desirable to have a risk-equalization pool, like that in Germany and a few other European countries, which would help mitigate the actuarial risks for insurers. This will contribute towards keeping premiums down and reduce the incentive for insurers to turn away patients because of their claim ratios over-shooting. All the actual claims processed each year by all insurers can be consolidated, risk incidence measured, and some pay-outs made so as to normalize risk incidence among all insurers.

Countries like the US, which already have a legacy insurance model, will invariably find it difficult to embrace many elements of this model. But countries like India, which do not have any existing insurance model, will find it easier to embrace elements that are appropriate to its requirements. Furthermore, the favorable age profile of our population too should go a long way towards keeping premiums down if the entire population is covered.

The crowds are not always wise!

James Surowiecki's best selling book, The Wisdom of Crowds, popularized the belief that the collective wisdom of a group of people was superior to the individual wisdom of even experts. It has generated considerable interest in the design of systems that seek to channelize the knowledge of large groups of people to say, predict events and prices. See Justin Wolfers' paper on prediction markets here.

It is based on the statistical phenomenon by which individual biases cancel each other out, distilling hundreds or thousands of individual guesses into uncannily accurate average answers. However, it assumes that the members of the crowd have a variety of opinions, and arrive at those opinions independently.

A new study of this phenomenon by Jan Lorenz and Heiko Rahut finds that contrary to conventional wisdom, groups insights could go awry if participants were influenced by the guesses of their peer group. They found that though groups are initially wise, "knowledge about estimates of others narrows the diversity of opinions to such an extent that it undermines” collective wisdom". Moreover, they found that "even mild social influence can undermine the wisdom of crowd effect". In this context, as the Wired article points out, computer modeling of crowd behavior also hints at dynamics underlying crowd breakdowns, with the balance between information flow and diverse opinions becoming skewed.

The authors recruited 144 students from ETH Zurich, made them sit in isolated cubicles and asked them to guess various indicators like Switzerland’s population density, the length of its border with Italy, the number of new immigrants to Zurich and how many crimes were committed in 2006.

At the end of each round of questioning, they were given small payments for coming close to the actual answer (signified by the gray bar). At left is the range of responses among participants who received no information about others. The findings of the study participants who were asked how many murders occurred in Switzerland in 2006 is shown in the graphic below.



The Wired article concludes,

"As testing progressed, the average answers of independent test subjects became more accurate, in keeping with the wisdom-of-crowds phenomenon. Socially influenced test subjects, however, actually became less accurate. The researchers attributed this to three effects. The first they called "social influence": Opinions became less diverse. The second effect was "range reduction": In mathematical terms, correct answers became clustered at the group’s edges. Exacerbating it all was the "confidence effect", in which students became more certain about their guesses."

As the authors claim, such false beliefs are commonplace in society, politics and markets. The herd behaviour of investors in financial markets is driven by excessive confidence generated by social influences. Opinion polls and the mass media largely promote information feedback and therefore trigger convergence of how we judge the facts and potentially create overconfidence in possibly false beliefs. Social fads and beliefs, some of which are of questionable value, become popular for no apparent reason.

In all these areas - markets, society, and politics - there are people and groups with an interest in influencing the beliefs of participants. They are vulnerable to being manipulated to suit the requirements of these vested interests. Such dissonances constitute failures in markets, politics and society.

Information over-load and health care

Standard explanations trace market failures in health care in general and health insurance in particular to information asymmetry (patients knowing more about their condition than the insurers and doctors knowing more about treatments and diagnostic procedures than patients) and its resultant adverse selection problems, and moral hazard (insured patients having no incentive to curb treatment and costs) concerns. Here are two less-discussed dimensions to this debate, both of which highlight the complexity involved in managing health care markets.

First, Tyler Cowen makes an important distinction between information asymmetry and information overload, and feels that adverse selection is less a problem. He writes,

"When it comes to the elderly, adverse selection as a problem is overstated. The real problem is usually a high degree of information about many conditions, so often insurance is difficult per se. It’s not the asymmetry of information that is the core issue, it is the existence of lots of information, and that is one of Arrow’s subtler points. That distinction matters a good deal for mechanism design.

An old person might know better his health care condition, but not know better his expected health care costs. That is a critical distinction. You can’t reach age 60 and credibly say: "I’ve been healthy so far, I guess my lifetime health care costs will be low." It’s not even clear whether the healthy or the unhealthy will have lower health care costs in their later years; the unhealthy might die rather quickly and decisively. Adverse selection on the grounds of health care costs need not be high and arguably actuaries can estimate those as well as the individual himself."

Another manifestation of information over-load involves the problem of patients being unable to effectively discriminate between multiple treatment options. For example, a patient exposed to two different sets of diagnosis, struggles to make a choice, leave alone the correct choice. Also, though the patient can avoid subjecting his/her body to all diagnostic tests if he/she can trust the doctor's clinical skills, such trust, for various reasons, is an increasingly rare commodity.

Further, most often, in their anxiety, patients end up following the herd and over-treating themselves. Unfortunately, the incentives of the doctors and the diagnostic service providers too are aligned towards leading patients down the path of the herd. In all these cases, it is not information asymmetry, but information over-load that either paralyses decision making or leads patients to make the wrong choices.

Co-payments and deductibles, while trying to incentivize patients to optimize on their treatment, does not always, atleast among those at the top half of the income ladder, curb over-treatment. Awareness campaigns and focussed information dissemination about medical conditions and treatment options can play an important role in helping patients make informed treatment choices.

In another post, Paul Krugman makes the point that health care recipients cannot be exact substitutes for "consumers" in the general marketplace. He writes,

"Medical care is an area in which crucial decisions — life and death decisions — must be made; yet making those decisions intelligently requires a vast amount of specialized knowledge; and often those decisions must also be made under conditions in which the patient is incapacitated, under severe stress, or needs action immediately, with no time for discussion, let alone comparison shopping.

That’s why we have medical ethics. That’s why doctors have traditionally both been viewed as something special and been expected to behave according to higher standards than the average professional. There’s a reason we have TV series about heroic doctors, while we don’t have TV series about heroic middle managers or heroic economists."

The term consumer-choice becomes meaningless in case of patients fighting to save their lives. The choice is mostly a fait accompli. As Krugman argues, it is indeed surprising that even forty years after Ken Arrow wrote this seminal paper distinguishing health care from other markets, the issue still evokes confused rhetoric. See also this post on the shockingly low levels of health care literacy even in the US.