Dec 28, 2010

Inside India: The Bureaucracy​

[This is the first of a 'Inside India' series focusing on the government system in India, mostly from the personal point of view of working with government and public sector organisations.]

The India (and China) growth story is familiar by now to anyone who reads a newspaper. Googling through pages of The Economist, Businessweek (now Bloomberg Businessweek)​ and The Wall Street Journal, to name a few, one would be forgiven for thinking India and China are vigorously scrambling up the ladder of economic growth, and using this growth to eliminate poverty and underdevelopment, regardless of the various challenges they face. But if you were to look for, say, a week, at the front-pages of Indian newspapers and magazines you would come across a different story. ​​​

Reports and proclamations on Indian growth remaining steady in the face of a global economic crisis are more often than not on p.7. What occupies p.1 are myriad tales of corruption, breakdowns in public services and constant political skulduggery with a-scandal-a-week rate at which almost everybody in public office is linked to illicit sums of money or to shady real estate deals. ​​

In fact, after about a week of scanning such headlines, one would find patterns. One, the media - and it is almost always the media, rarely a proactive government investigative exercise - breaks a sensational story on corruption in high places. Two, the Indian parliament comes to a grinding halt, as opposition parties of various hues dig into the various pressure points in the ruling party's body that this story creates. ​Three, the party in power announces, after a bit of initial denial, the launch of a commission or task force to investigate and bring to book the guilty. Four, editors and television analysts churn out pieces on how 'change at the top level' or 'a fresh breath at the top' or 'a high powered committee' is urgently needed to remedy the situation and eradicate the weeds of corruption. Fifth, the government announces the resignation or transfer of a couple of big heads.

A good illustration of this cycle is the Commonwealth Games hosted by India. Plenty has been written and said about the sloppy arrangements, collapsing infrastructure and the rigged tenders that gave away contracts to firms at overpriced rates. Stories about the Games built up over a fortnight, as the launch drew nearer, exploded and by the time it had peaked, a big head had rolled - that of the head, no pun intended, of the Games Organising Committee. The scandals around the Games then quickly disappeared from the front pages and prime time news channel slots, as if the crowds have been satiated by this act of throwing a single man to the lions.

This cycle of events suggests two things: i) corruption, particularly grand corruption, in India is now 'cycling' at an ever greater frequency of peaks and ii) 'changes at top level' type of remedies have not worked. A good reason for this is the systemic corruption at lower levels prevalent in the huge bureaucracies of the executive, judiciary and the armed forces. ​​​In fact, 'fresh breath at the top' kind of changes have no effect on this type of organisation-wide corruption much as pouring water into a hollow bucket provides the appearance of a momentary increase in the level of water which subsides once the flow of water reverts back to normal.

This sort of organised predatory corruption is under the radar and is more corrosive than grand corruption as it is the kind of evil that a common citizen faces while trying to access basic public services. In fact, one could argue that to understand and improve India's development prospects is to understand and improve India's civil service.

This office-level corruption is the result of organisation wide distortions in India's public administration system. The distortions create incentives for subjective interpretations of administrative and legal procedure and creates scope for exercising personal discretion in the dispensation of a public service to cronies. And the distortions sustain these incentives, maintaining high systemic corruption at the frontline of government service delivery in everything from primary education to law enforcement.

In other words, this is not just a case of a corrupt politician fattening himself on a government contract or two. It is the case of everyone in the administrative system, to varying degrees, eating out of the same pot and then passing the pot around to a different set of people once a new government comes in.

There is a growing literature on corruption, not least in economics. They catch perceptions of corruption or how key macro variables are affected by macro levels of corruption. [For example, take Paulo Mauro's work on how corruption adversely affects economic growth.] Research of this sort and of the kind of 'corruption perception indices' brought out by agencies like Transparency International are useful in making the case for general action against corruption as an essential element of promoting growth.

However, what are the agency level factors that lead to corruption at a micro-level and how do these mechanisms work? Economics may not be the right lens to answer these questions. Business administration and management science are better equipped at tracking and sorting these mechanisms out. [An anthropological example of this type of work is the research by Robert Wade in South India on corruption in irrigation departments that control the canals needed by farmers to water their crops.]​

Based on my experience of working as a consultant to government agencies and, now, as part of an international development agency working in India, I think that management and human resource management aspects of Indian ministries and government departments provide an explanation of why the quality of public administration in India is poor, corrupt and successful at resisting reform.


For a government ministry or department, the equivalent of a corporate policy are the Rules of Business (ROB), a relevant Act and a Compendium of Rules based on the original Act. While the Rules of Business and the Act provide broad legislative pointers to how a government agency should work, it is the Rules that provide operational guidance on procedures.

While this system seems sensible at first glance, what has happened over time is that there have been no systematic and periodic reviews of the ROB or the Act. Consider what this means: India's government departments and ministries have a policy framework and legislative backbone that has changed very little from its colonial origins. This is obviously not very useful when it comes to the agency's real day-to-day job: implementing the legislative provisions of the Act.

To work around this, every year several new Rules or modifications to old ones are created once a situation arises where an old Rule has to be modified or a new one introduced. That is, what changes are made are reactive. The body of Rules grow steadily across many such compendiums or volumes till such time that only a handful of personnel, usually old clerks who have been in their chairs for about two decades, are aware of the outlines of the entire body of Rules and can refer to the relevant Rule if the situation arises. Without systematic, periodic reviews that phase out out-of-date legal provisions and accompanying procedures and introduce new ones, the entire policy framework on which a department rests is largely based on colonial era laws with several dozen 'adjustments' to it in the form of Rules.


There are many mentions in popular literature of the infamous Indian government file that is pushed through its various loops by a well placed word or bribe or both. Since political independence, India's government machinery has amassed millions of files with documents recording every conceivable government-citizen exchange of services, grievances and complaints. The original filing system in the colonial bureaucracy was probably efficient for the times and perhaps deliberately designed to make colonial subjects run from pillar to post.

A file was created when a 'notesheet' recorded the first instance of an executive decision that needed to be taken. The lowest clerk recorded the facts of a case, the relevant legal provisions or procedures that applied and then passed it upstairs to his immediate superior. The superior officer in turn read the contents of the notesheet, jotted his opinion on the case in the margin, put his initials on it and marked it to his immediate superior. This went on till the notesheet acquired a chain of jotted remarks and signatures along the margin till the file went up to the highest level of a Principal Secretary who took the final decision if none of his subordinates had by then been able to or had the authority to take one. If the Secretary felt the need to look up a legal clause or needed more case information, back the file went looping downwards through the entire hierarchy again. The number of loops that a file had to go through depended on the underlying complexity of the business process and / or on the extent of ambiguity of the legal provisions relevant to that case. For a feel of what a typical government process looks like in India today, take the example below from an actual municipal corporation of a big ('metro') city.

Granting of a No-Objection-Certificate (NOC)

For a process like the one in the diagram, the file will bounce back and forth along a chain which will look something like this:

'Dak' or Correspondence File -> Lower Division Clerk (LDC) creates a file based on the complaint or request or claim or appeal that has just come in by post -> the Upper Division Clerk (UDC) processes the file, marks his notes on the margin -> the Section Officer (SO), a senior clerk then takes a decision on whether the file can be processed at a lower level or whether it needs 'interpretation' of the relevant legislation -> if shuffled upstairs, file goes to a Deputy Secretary -> onward again upstairs to an Additional Secretary -> onward again if need be to an Officer on Special Duty (who also doubles as an Additional / Deputy Secretary depending on the terms of service and seniority and available places left for promotion) -> and finally ends up on the desk of the Personal Assistant (PA) to the Secretary -> from where the Secretary will finally take a look at the file.

Any need for clarification or further information at any point on this chain will mean that the file is pushed back to the preceding link in the chain. ​Add several such loops if the file contains issues that need approvals or clearances from other civic agencies, departments or ministries.

Over time, this system has remained practically unchanged. [An indication of this is the recent move by the bureaucracy to mask the margin jottings that reveal the decision making chain from the clutches of India's new Right to Information Act under which it is mandatory for a government department to reveal inside information to a citizen on request.] With the phenomenal increase in India's population and the fact that government still provides almost all services to citizens, not just public goods and especially outside the big cities, the number of files has snowballed to a very large, unknown number. Unknown because there has been no census or government-wide time and motion type of study to track document load at each office.

The original filing system was organised much like an ancient university library with Index Files updated regularly for navigating the file classification system. Had this sort of system been diligently maintained, perhaps things would have lent themselves relatively easier to future automation.

However, the crucial Index Files are also not updated regularly which means that they are no longer a complete or accurate record of the archival system in government offices. The quality of the Index File rests on the regularity with which an Upper Division Clerk (UDC) or, more rarely, a Section Officer (SO) updates it. In practice, the File Index may contain a trace to the most-frequently- referred-to files. The rest of it lie there gathering dust or are thrown away in corners of the building or sometimes even trashed (to be honest, if files keep growing and there is no archival policy, where's the space?). There are virtually no computerised document management systems, though a few public sector enterprises are experimenting with such systems in recent years.


Performance metrics have to strike a balance between what is tangibly measurable and what is the right outcome to measure. Over time, this balance has gone haywire in government organisations in India. The all-important measure of performance of a department or ministry is the quantum of money spent in a financial year as documented by the all important Utilisation Certificate (UC). ​The UC certifies that a government department or ministry has spent a tranche of funds that has been allocated to it for a specific programme (or 'scheme' as it is known in India). The UC measures what its name suggests - how much of the money allocated to an office has been spent. It does not concern itself with the impact of the funds spent.​
In the Soviet-style centrally planned economy of the fifties, this may have been the only variable to track, though the logic is still hard to fathom, but in today's Indian economy where many products and services that were once produced only by government are now supplied by the private sector, this is a partially blind approach to measuring the performance of government programmes.
Scores of government ministries, line departments and other agencies in India still report 'performance' in this antiquated format, diligently filing reports on budgets spent without any data on whether the expenditure achieved any of the planned outcomes. Programme indicators are reported separately, if at all, unrelated to the programme expenditure and often at such levels of aggregation that it is difficult to comprehend the impact they signify apart from statements.
In other words, the annual report of a department will often quote something like"X training sessions completed for women artisans" while reporting on a programme, along with the total percentage spent out of the funds allocated to that programme without:
- explaining the which/why/ how of the outcome indicators chosen
- the innovations, if any, introduced in the processes needed to roll-out the programme
- the data collection approach and mechanism
- the type (internal / external) and frequency of monitoring and evaluation reporting and the resultant trends (e.g. did the key indicator have sharp drops in the middle of the programme and then rise up to its current level?)
- synergies or conflicts with other programmes run by the department or by other departments
The result is that what passes off as 'performance' monitoring in government programmes is a basic accounting appraisal in the shape of "how much was given to them versus how much did they spend" variety of reports.
At an operational level this translates into counting mostly input level indicators (number of training sessions, number of workshops, number of filed agents hired etc) rather than results of the programme (how many children cleared the test for numeracy after the training etc). ​
The Indian political system encourages this sort of skewed reporting. Indian politicians are wont to announce triumphant new 'schemes' once they are voted to office with a certain budget value earmarked. There is practically no discussion on underlying management processes and outcomes to be monitored and achieved. In fact, most of these announcements are in the form of renaming an existing programme with an increased budget allocation and launching it with a publicity blitz. ​


When Qutb-ud-din Aibak died in 1210 A.D., who became his immediate successor?
(a) Iltutmish
(b) Taj-ud-din Yildoz
(c) Nasir-ud-din Mahmud
This multiple choice question is taken from the History paper in the Preliminary Examination ('Prelims') for the Indian Administrative Service (IAS) cadre, the most influential branch of the bureaucracy today. While the examination as a whole is gruelling and involves at least a couple of years of preparation, it has no relation to public administration or to the technical aspects of regulating, say, energy or telecom markets. Candidates ​can, and do, choose from a wide variety of liberal arts and science subjects. Masses of candidates, especially from states where technical education facilities are poor, choose subjects such as Hindi, English literature, History, Geography or Geology and clear the exam. These candidates then go on to handle ministries where specialist knowledge is often required. This situation is largely similar for many such premier 'Group A' services in the bureaucracy, although for very specific positions such as public sector bank officer examinations or positions in the civil aviation sector, there is no choice but to test the candidates on narrowly technical areas suitable for the job. 
What this results in is a deluge of civil servants with academic training that often has no relation to the kind of problems they manage in their career. A civil servant with a combination of geology ​and Sanskrit in the entrance examination could end up in her senior years heading the agency that regulates electricity tariffs. While one can argue that once selected, all civil servants are put through their paces for a year at an academy, this training is similar to an extended induction session that takes the new entrants through the minutiae of office and administrative procedures. Exhibit A is India's Department of Information Technology which is filled with civil servants of designations such as 'Scientist -A' but all of who are busy with routine paperwork instead of genuine research and development. This leads to a large information asymmetry between a government official tendering for a service and the private sector firm that provides the service. 


If there is one fact about Indian government offices that you could bet on and consistently win, it is the fact that any ministry or department or agency office has on average a 20 per cent vacancy rate. Add to this another fact: at least half of those vacancies would be at the managerial cadre ('Groups A and B Level Officers'), not at support staff level, which impacts the quality of administration significantly.  This is a conservative estimate, some agencies have much more.
A thumbrule to predict the proportion of vacancies in positions for sanctioned posts (i.e. positions for which the government has officially allocated a line budget item for pay and benefits) is the political importance or rent-seeking ability attached to that office. ​A government office at federal level, to do with government procurement or regulation, is less likely to have vacancies than a district level office which simply implements programmes.
Shouldn't it be the other way around? After all, even if we assume that a developing country government do not have budgets to fully staff every office​, shouldn't resources be prioritised towards offices engaged in actual service delivery at ground level? No, because these lower level offices do not have the influence to generate rents so much as an enforcement or regulatory office located in New Delhi close to big businesses and corporate firms in India lobbying for minute but crucial changes in the policy landscape (the 2G spectrum allocation corruption scandal is a good example).
As an example, let us consider the network of vocational training institutions in India called Industrial Training Institutes (ITIs). One would think that with the daily additions to the discourse on competitiveness and the need to develop a skilled workforce, the ITIs would be a natural starting point to intervene by the government. Now, think about what the figures below show.

The data above relates to only management cadre officials, Assistant Director and above. Most positions have been lying vacant for 3-6 years. At the Director level - the institute head - some positions have been vacant for more than nine years. This is of course in addition to the other problems with these sort of government run institutions - outdated equipment, onerous reporting and accounting procedures, lack of performance pay for achieving targets but plenty of disciplinary measures for a minor deviation against some guideline or law written ages ago and other factors that are so typical of government service delivery in general, stifling initiative and rewarding conservatism and mediocrity.
The problems listed under these five points lead to corruption because they make it almost impossible for a clean, rational process with definite outcomes to be developed. Because of unclear policies, personal discretion is introduced by the bureaucrat (and politician) in interpreting the target recipients of a service. Because of weak Management Information Systems and a half-blind performance appraisal mechanism, targeting and outreach of government programmes and services can be diverted to special interest groups or not delivered at all by the bureaucracy. Because of a bureaucratic organisation that is riddled with vacancies especially at management level and a colonial era recruitment system, the capacity of the system to deliver even basic services is severely limited. All of these factors when combined with a population that is largely rural and poor lends enormous power to an Indian civil servant who then uses these to rationally maximise personal wealth and position in the bureacracy and selectively deny or provide services to generate rents.
A permanently weak system is a boon for the bureaucracy in terms of the magnitude of personal discretion it can exercise. Clearly, someone within the bureaucracy must be crazy to try and reform this system when the benefits accruing from a damaged system far outweigh the costs of fixing it. 
Yet top level policy pronouncements such as the Prime Minister's National Competitiveness Manufacturing Programme do little more than increasing the budget for a specially earmarked programme without reforming fundamental management level bottlenecks such as these. ​
The effect ​is that of pouring water into a hollow bucket.​

No comments:

Post a Comment