Grand Unified Framework to understand elements of state capacity in the context of school education
Why a framework for state capacity?
In another post, I discussed the nature of problem in public education - the importance of correctly identifying the nature of problem, how a misdiagnosis can lead to incorrect approach to reform etc. I highly recommend reading that post (link shared below) before continuing with this.
What’s the binding constraint in public education or in more layman terms, what’s the bottleneck or the major reason…medium.com
In summary, the crux of the argument in the above post is that the problem in education is not necessarily with those visible problems like lack of infrastructure or lack of teacher motivation etc. The inference was that the problem is much deeper and that weak state capacity is the binding constraint in public education and NOT anything other as normally pointed out.
The problem with the arguments like ‘weak state capacity is the binding constraint’ or ‘it’s a structural problem’ is that they are too vague and can mean anything. As Kaushik Basu says
When you point weak state capacity as the constraint, the other argument is that — India conducts world’s largest democratic elections with highest levels of transparency. Clearly, it’s not the case of weak capacity. If India can conduct such complex task of elections, why not education?
Lant Pritchett’s topology of tasks performed by the government gives a partial answer to the above question - tasks that require discretion for the frontline worker to be able to carry them out, are more difficult. It helps us understand the black box of state capacity to an extent. But then the question is — if discretion is the key, why not just give it? Well, if you give discretion, people misuse it. Oh, wait! Not everybody does so. There are many such followup questions.
We need to thus answer such questions accounting for more details into our model. In this context, it is thus important to delve deeper and understand the components of which ‘state capacity’ is made of, so as to a get a broader sense of its functioning.
The other reason to have a framework of state capacity, however imperfect it might be, is to have a sense of our destination and if something goes wrong, to be able to figure out the origin of the problem. Having a sense of destination will put the micro initiatives in perspective else it’s just moving aimlessly. It is analogous to a journey to a destination with a map. If you don’t have a map but are still moving (micro initiatives), you may well be progressing but you don’t know where! On the other hand, if you have a map, the motion through the micro initiatives will give a sense of direction and status of current location and the task ahead, putting things into perspective. In reality, we might not have the luxury of having a detailed map, like the Google Maps, but something on these lines is desirable.
With this context, let us move on to build a broad framework to understand the components of ‘state capacity’. The idea here is to break up the ‘state capacity’ into components and understand the implementation of a public policy through these components. The state capacity challenges are then modeled as interactions between these individual components.
Framework for state capacity
The framework for understanding state capacity in education is documented in my book UnpackED: The black box of Indian school education reform. The text below is an excerpt from my book, with some updates.
Grand Unified Framework (GUF from now on) in the context of education has 8 components that are shown in Figure 1. Some of them may have overlaps but these are categorized into 8 components for the sake of simplicity. The net result of interaction of these 8 components decides the effectiveness of policy implementation.
- Nature of the personnel
Let me introduce a new term called ‘local capacity (LC)’. Local capacity is different from state capacity. Local capacity as the term indicates is about a local system. It’s simply a measure of —if given a task, how confident are you that the personnel in lower bureaucracy will do their job without resorting to corruption or work evasion?
In an ideal ‘every one is a Mahatma’ world, a front line worker would do every job with responsibility and wouldn’t need any monitoring. Such systems can be termed as ‘high local capacity’ systems or simply high capacity systems, for the sake of simplicity. In a non-ideal world, you would need monitoring, checks and balances to weed out inefficiencies. The divergence between this ideal world and non-ideal can be thought of as resulting from the ‘nature of personnel’, the first component of our model for state capacity
Nature of the person dictates whether she performs duties without resorting to corruption or evasion of work. This nature of a person further has 2 components — a part of person’s nature that’s inherent to her, and another part which is shaped by the system in which the person is present in.
The inherent honesty levels of people differ — some people are inherently honest and they don’t deviate despite external compulsions. The thresholds of this vary across people.
The second aspect of the nature of the person is the part of the personality that is shaped by external compulsions or constraints. These constraints influence the overall behaviour of the personnel by operating through this part of nature which reacts to incentives and disincentives.
One must note that the inherent nature of the person can also be altered due to prolonged disincentives. For instance, even if a person starts out with some level of honesty, a trait of inherent nature, prolonged disincentives for being honest and scope for corruption can make them less sensitive to corruption and work evasion etc.
If we set up appropriate incentives, it influences that part of nature that’s susceptible to external constraints, and if it is continued for long time, it strengthens the inherent levels of honesty (loosely speaking). After some time, the default instinct will be to NOT resort to corruption or break rules or evade work. This is the nature attained after what some call the ‘cultural shift’.
In summary, the nature of a person can be thought of as a mixture of these two sub-elements, inherent honesty traits and responsiveness to external constraints, with varying weight-age of each of these across people.
2. Skill of the personnel
It is not just enough for personnel to be honest and responsible, she should also have the necessary skill to perform the job. Skill of the personnel to perform the task is the second component in our framework.
The combination of first and second components constitute the individual level characteristics of personnel and form a black box. These two components — nature of person and skill of personnel together form the ‘local capacity’. The challenge of external system is — how can it influence this setup to achieve the best possible outcomes?
3. Autonomy given to the personnel
The third component of our framework is the autonomy or discretion given to the personnel to be able to take decisions while performing their task.
Having autonomy will reduce the time of decision cycles and improve the quality of service delivery as one can customize the responses to the context. On the other hand giving autonomy might also lead to misuse.
4. Nature of the task(s) — level of discretion required to perform the task
Nature of the task(s) performed by the personnel is the fourth component in our framework. The third component is about ‘whether the personnel has autonomy or discretion’. This fourth component is about whether autonomy/discretion is required to perform the task.
This follows Lant Pritchett’s classification of the nature of tasks performed by the government based on discretion required for frontline worker to perform the job — postman job requires little discretion while performing the tasks while doctors and teachers require high discretion while performing their job.
Discretion required to perform a task is a subjective aspect which changes with time. Many aspects that would have other wise required human intervention are being automated using technology. In an extreme case, every thing including education service delivery and health care would be automated, replacing teachers and doctors, solving the monitoring issues altogether. In reality, the tasks are in the spectrum of ‘amenable to automation — not amenable to automation’.
5. Monitoring norms inherent to the lower bureaucracy
These are the formal and informal norms that are inherent to lower bureaucracy. Remember that these aren’t necessarily mandated by a central authority as part of implementation design of a public policy.
The formal monitoring norms set the incentive structures under which the personnel operate. In terms of economics, they determine the ‘stakes of non-performance of a job’ and ‘incentive for performing a good job’.
The informal norms can be thought of as arising out of trust between frontline personnel and monitoring authorities at local ecosystems. Such informal monitoring norms can influence the quality of service delivery.
Prof. Akshay Mangla of Harvard University has studied the phenomenon of unwritten bureaucratic rules in the context of Indian education. Analysing the difference in learning outcomes of Uttarakhand and Himachal Pradesh, which are other wise similar in socio-economic characteristics and administrative structures, Prof. Mangla argues that
bureaucratic norms — unwritten rules that guide public officials — influence how well state agencies deliver services for the poor
6. Implementation design of public policies
We discussed the inherent norms of lower bureaucracy in the fifth component. These are the norms of operation at the local level. Apart from that, lower bureaucracy is also influenced by the norms of implementation design of specific public policies.
Public policies often have monitoring and reporting procedures embedded in their implementation design. For instance, suppose say that the permission to finalize the site for a child-care centre is strictly to be made at the district level, centralizing the decision making, it conflicts with the ethos of a lower bureaucracy which functions based on autonomy based structures. Probably the autonomy based ethos are due to a large set of other policies where the lower bureaucracy has power to take similar decisions but only the child-care centers policy deprives them of such power.
7. Skill to frame implementation design and monitoring norms
A well-designed policy need not just be technically sound, it should also account for the existing norms of the organisation, factoring in the effects of implementation design’s interaction with the existing framework at the lower level.
While we can iterate on policies to fine tune them, the way technological products are built by developing several prototypes, we must note that iteration of policies sometimes have significant costs associated with it, as each change in a policy has the potential to affect a large population, while such costs are low in case of software products. Hence, we must carefully anticipate and account for all possible aspects while framing rules and regulations of the implementation design. This requires specialists with strong domain knowledge and understanding of the nuances of implementation policy design and effects of its interaction with local systems.
For instance, consider the example of Continuous and Comprehensive Evaluation (CCE) in India where students are graded based on cumulative scores recorded in low-stake tests at shorter intervals as opposed to earlier practice of one test at the end of the year or twice an year. CCE may be desirable in theory but if the implementation process is made too cumbersome for an average teacher to carry out activities in class, it defeats the whole purpose. Similarly, functionaries at mid-level or higher level should also have the skill to design and adapt norms of functioning as per context.
Requirement of skills to design implementation policy and norms of organisation can be thought of as local capacity at higher levels. Just as professionalisation of lower level bureaucracy helps (Francis Fukuyama: What is governance?), professionalisation of bureaucracy at higher and mid-levels by equipping them with appropriate skills to design norms and implementation design, can also enhance effectiveness of implementation of policies.
One should note that administrative issues related to inefficiencies in flow of funds are a overlap of sixth and seventh components.
The seven components — nature of functionaries, skill, autonomy, nature of work, inherent norms, implementation design, capacity to design norms, are related to the internal dynamics of the bureaucracy that shape its capacity.
8. Environment where policy is implemented
The eighth component of our framework is the environment in which policies are implemented. One should note that while the first seven components were part of the education bureaucracy, this component is external to it.
The environment where policy is implemented or service is delivered interacts with the bureaucracy at three levels.
i) Resistance to state penetration: Joe Midgal used the term ‘resistance to state penetration’ in his seminal work “Strong Societies and Weak States:
State-Society Relations and State Capabilities in the Third World” to outlines the effects of interaction between the state and society.
We can give meaning to this term based upon context. In education, for instance, a bureaucracy may achieve to higher outcomes while educating a child from a rich family as compared child from an underprivileged community which is against educating children. In other terms, a same teacher might produce different results in these two contexts, the difference largely driven by the community and not necessarily the bureaucracy.
Suhas D. Parandekar and Elisabeth K. Sedmik of the World Bank highlight the contribution of ‘cultural factors’ to the high scores of Vietnam, explaining its anamoly of high scores among countries of similar profile.
The arguments that perceived good quality of private schools is a result of the type of students who attend such schools, and not necessarily due to the school come under this component. Similarly, the evidences which suggest that having educated mothers helps students also comes under this.
The analogy to legal enforcement is — if every citizen decides to deliberately violate a law (high resistance to state penetration), there isn’t much the police force can do. We often miss this factor in thinking about the effectiveness of enforcement.
ii) Enforcing accountability: Communities interact with bureaucracy to enforce the accountability. The nature of such interaction influences the quality of service delivery. Albert Hirschman’s ‘Exit, Voice and Loyalty’ needs no introduction in this context.
iii) Shaping the intangible incentives: Communities also shape the incentives of the personnel in intangible ways. For instance, if teachers are respected in a particular community, it may act as a motivation to them.
We discussed 8 points of state capacity till now. A quick recap, the 8 components are
- Nature of the personnel
- Skill of the personnel
- Autonomy given to the personnel
- Nature of the task — does performing the task requires discretion?
- Monitoring norms of lower bureaucracy
- Implementation design of public policies
- Capacity to design monitoring norms and implementation design of policies — local capacity at higher bureaucracy levels
- Environment or communities in which the service has to be delivered or policy has to be implemented
We finally have our broader framework of state capacity ready. We now proceed to understand the challenges in education service delivery as challenges arising out of interaction between the above 8 components.
Challenges in education service delivery as interactions between components of state capacity
Challenge I. Interaction between Autonomy and Local Capacity
To be clear about the terminology, local capacity corresponds to the 1st and 2nd components of our framework, nature of the personnel and skill. If a given task is done well without monitoring, which is a combination of nature of person and skill, then we call such such local systems to be having strong local capacity.
Autonomy is the autonomy given to the personnel to take decisions while carrying out the tasks.
Giving autonomy to personnel may help increase efficiency but it’s also prone to misuse. When to give autonomy then? The local capacity comes into picture now.
Giving autonomy to high local capacity systems enhances the quality as the probability of performing the task without resorting to corruption or work evasion is high, even with little monitoring, is high in such systems. On the other hand, autonomy to low capacity local systems can deteriorate the quality.
Hanushek et al paper on school autonomy — “Does school autonomy make sense everywhere?” is a prominent paper point out the possibility of negative effects of autonomy. Using PISA scores and responses to questions on schools’ power in decision making (also part of PISA), the authors find that
“ autonomy affects student achievement negatively in developing and low-performing countries, but positively in developed and high-performing countries.”
It’s because, as they discuss, if given more autonomy, a lazy, non-school attending teacher might use misuse the autonomy by reducing the syllabus and so on.
Another recent study on impact of decentralization in Indonesia suggests
“ no overall effect on achievement, but a negative effect on teacher effort, particularly in rural areas and among schools with inactive school committees.”
Bloom et al. paper on school management is also worth noting here. It uses data of 1800 schools across 8 countries and finds
Higher management quality is strongly associated with better educational outcomes
Almost half of the difference between the management scores of autonomous government schools and regular government schools is accounted for by differences in leadership of the principal and better governance.
In other words, the local capacity at school level, the accountability structures etc seem to be the key factor in determining the effects of decentralization. Intuitively, giving more autonomy to school systems with high local capacity enhances quality where as giving autonomy to ailing systems is detrimental. Good schools typically tend to have good management structures.
So, for a given system, at a given point of time, there is only so much autonomy that can be given to be able to maximize the quality.
Francis Fukuyama calls this as the inverted U curve of quality vs. autonomy.
Does increasing autonomy enhance quality for all types of tasks? Not necessarily. It depends on the nature of the task, which brings us to the next point.
Challenge II. Tripod of ‘Local Capacity — Nature of Task — Monitoring Norms’
We recapped local capacity above. Nature of task is determined by the level of discretion required by the frontline personnel to perform the task. Meaning of Monitoring Norms should be self-evident. How does the interaction of these three elements play out?
Let’s start with the inverted U curve (Figure 2). It says that in contexts with weak local capacity, inefficiencies arising out of it can be dealt by reducing autonomy, which means monitoring the tasks through rigid rules.
Decreasing autonomy may work in cases where the task itself doesn’t require discretion for effective implementation or in other words (tasks amenable to automation). What if the task itself needs discretion to be performed effectively (not amenable to automation) but we have a case of weak local capacity and our inverted U mandates against giving discretion?
We are in a situation where, if we give autonomy, it will be misused. If we don’t give autonomy due to local capacity constraints, then the task can’t be performed meaningfully.
For instance, the job of a postman which doesn’t require discretion can be monitored through rigid rules in weak local systems. What about the case of teachers whose job requires discretion?
The nature of the job of a teacher (2nd leg of tripod) requires discretion to adjust to children’s queries and situation in the class but the weak local capacity (1st leg of tripod) forces the monitoring norms (3rd leg of tripod) to be rigid rule based, resulting in a conflict. This conflict is at the center of state capacity issues in delivering education.
Thus, as discussed in an earlier post, we can increase the quality of delivery by adjusting the autonomy as per the local capacity of the system. But if the nature of the task itself requires autonomy to be performed meaningfully, then decreasing autonomy isn’t a sensible instrument to enhance quality, enhancing the local capacity is the only option.
Challenge III. Variance in nature of tasks of a same personnel
The third challenge is due to the tasks of multiple natures performed by the same personnel.
In the second challenge discussed above, we assumed that the front line worker performs task of only one nature. What if the personnel performs multiple tasks and they are of different natures — some requiring autonomy to perform, while some not requiring autonomy?
It leads to two challenges —
i) It leads to de-prioritisation of tasks that require discretion because such tasks aren’t easily trackable in short term.
(ii) Conflict between aptitude or mindset required to perform the tasks
(i)De-prioritization of tasks
Let’s say that a personnel has two tasks — one requires discretion and the other doesn’t. Tasks that don’t require autonomy are easy to monitor because there are clear metrics to measure progress and mechanisms to attribute the results to the personnel are clear. They can be monitored through rigid rules. However, the tasks that require discretion are difficult to track meaningfully in short term and also the mechanisms to attribute the results to the personnel may not be clear, at least in short term.
The case of Cluster Resource Centre Coordinators (CRCCs) in India is a classic example. CRCCs are supposed to be supporting structures to teachers helping them with aspects related to teaching; coaches in other words. This task of CRCCs requires discretion to be able to respond to each teacher differently. It’s difficult to monitor such task by rigid rules or mandate CRCC to give same instruction to all teachers. Along with this, CRCCs are also made to collect data for different purposes and report to higher officials. This task in the basket of tasks of CRCCs doesn’t require discretion and it’s monitored regularly owing to its time sensitive nature.
In a time use study of CRCCs, Yamini Aiyar et al. found that CRCCs spend less than 10–20% of their time in classrooms. It means that they spend 80–90% of their time in collecting data. It’s because the task of collection and reporting data is time sensitive and easily trackable. On the other hand, judging whether a person has effectively coached teachers in a class room isn’t easily trackable. It takes time for such results to emerge and even if the results are bad, CRCCs may not be pin-pointed for the blame. Thus, task that can be monitored by rules and hence urgent got prioritized over tasks that aren’t amenable to monitoring through rules.
What doesn’t gets measured doesn’t get improved! Effectively what happens is that only those tasks get tracked which can be monitored through rules and have clear metrics. It pushes the tasks that aren’t amenable to rule based monitoring (require discretion) down the priority list of the personnel.
The piling up of work load needs a special mention here. Whenever a new policy is announced, it effectively ends up adding a new task to the basket of tasks of a personnel — addition of 6th component (implementation design of policy). Typically, new personnel aren’t recruited with a new policy. So, over time, the tasks pile up on the personnel at lower level.
The phenomenon of deprioritization worsens with increase in such work load. When a personnel has too many tasks, tasks that are easily trackable by her higher authorities figure high among priority list. They become tasks of high stakes. Discretionary tasks that can’t be tracked in short term end up being low stakes tasks and hence don’t figure in their schedule.
(ii) Conflict between aptitude required to perform tasks
Let’s consider a person who has to perform both ‘routine task’ (task that doesn’t require discretion to perform) and also a ‘task that requires discretion & judgement’.
The first type of task doesn’t require self-responsibility because such aberrations can be handled through monitoring. The second type of tasks require some form of self motivation to effectively address the dynamic situations.
If you monitor the person only on the former task (routine task), through rigid rules, for a long time, it can lead to long term consequences where she might end up thinking that she shouldn’t exercise discretion while performing tasks. Through monitoring by rigid rules, certain type of attitude is imbibed. Spill over of such aptitude into tasks which require use of discretion affects their performance.
Yamini Aiyar et al. call this the ‘post-office state’ in their CRCC study. It reports
“CRCCs perceive themselves as cogs in the administrative machine, which inevitably leads them to internalise and interpret the challenges they face in their jobs as something that remains outside of their control. This legitimises a culture of apathy and lack of responsiveness toward understanding and directly addressing the learning deficit”
The challenge thus is — how do we monitor people who perform tasks of varied natures?
One solution is to re-assign tasks and ensure that one person performs tasks of only one nature.
Do we have such flexibility and bandwidth?
Challenge IV. Variance in capacities of local systems or personnel
In the third challenge, we discussed the challenges due to variance of tasks performed by an individual. If we look from a macro scale, we note the fourth challenge, variance between local capacities of systems (local) and personnel.
In simple terms, in a large state, let’s say there are 1000 local systems, of which 200 have high local capacity, 300 have weak local capacity, while the rest are in between. The constraints of transparency and fairness mandate the higher authorities to subject all the 1000 local systems to the same treatment.
Keeping in view of the 200 local systems with high capacity, if autonomy has to be given regarding certain aspect to enhance the quality, it would mean that the quality will deteriorate in others. If autonomy is reduced, it deteriorates the quality in 200 local systems. While the trade-offs can be made based on effects on overall quality, we note that ‘local systems’ are the units here and these distinct units are at loss.
Similarly, within a local system, if trust-worthy personnel have to be monitored based on trust giving more autonomy, it might lead to accusations of partiality from those who aren’t given autonomy.
The challenge thus is —
how do we best deal with the variance in capacities across the system?
Challenge V. Interaction between rules of specific public policies (6th component) and nature of norms in lower bureaucracy (5th component)
If the lower bureaucracy in a particular local system works on unwritten rules based on trust and if the implementation design of particular scheme gives little autonomy to the frontline personnel mandating rigid rule based monitoring, it interferes with the functioning of the personnel in a negative sense.
Similarly, if autonomy is mandated to personnel where the lower bureaucracy works based on rigid rules, it might create some tensions!
Challenge VI. Interaction between expectations of a personnel from the higher authorities (5th component — monitoring norms) and expectations from the environment where it’s implemented (8th component)
The 8th component in our framework is the context or environment in which the task is being implemented. In our case, it is loosely the society or community.
The personnel are subject to vertical accountability by the virtue of being in the bureaucratic hierarchy and are at the same time subjected to horizontal accountability by the virtue of being part of the community where education is being delivered. The misalignment of expectations from these two can also lead to tensions and inefficiencies.
The challenges in education service delivery can thus be boiled to 6 forms resulting from interactions of 8 elements in the framework of state capacity. The 6 challenges are
- Autonomy vs. Local Capacity
- Local Capacity — Nature of Task — Monitoring Norms
- Variance of nature of tasks performed by a personnel, leading to deprioritization of tasks that require discretion to be performed
- Variance in capacity of local systems leading to design of sub optimal rules.
- Alignment between rules mandated by implementation design of public policy and inherent norms of lower bureaucracy
- Alignment of expectations between monitoring authority and communities — vertical accountability vs. horizontal accountability
The terms governance, state capacity, accountability, systemic problem are used interchangeably to communicate an abstract theme.
Some times, questions are also raised on the arguments of ‘weak state capacity’ being the binding constraint because bureaucracies that don’t perform certain tasks well, also perform certain other tasks in an impeccable manner. The case of Indian bureaucracy conducting elections while not delivering education is a case in point.
In other cases, lack of political will is cited as the binding constraint. However, there are cases where initiatives even with government’s will haven’t been successful. Also, we intuitively know that the task of conducting elections is different from that of education service delivery.
In order to address these questions, it is essential to break down the abstract theme of state capacity/governance into understandable components and explain the challenges in service delivery using these components.
The above framework tries to disentangle these abstract notions into 8 specific components of state capacity and models the 6 challenges in the education service delivery as specific interactions between the components, using such framework.
This framework gives a theoretical understanding of the nature of processes involved in education service delivery understanding which can help in diagnosing bottlenecks for a given context. It might also be extended to other areas like police with necessary tweaks.
It is important to reiterate that this framework is only an initial attempt to explain the mechanisms involved in implementation of policies. It’s no where near perfect or fool proof. It can be used as a base to build further. One can also tweak this by testing some of the arguments — for instance, we argued that if an additional work is given to a personnel in form of a new scheme, without increasing work force, the quality of her performance on other tasks is going to get affected. It’s an empirically verifiable hypothesis exploiting natural experiments.
The two unique aspects of this framework are that it considers state capacity as a dynamic quantity and not a static quantity. It theorizes that the quality of implementation will suffer with additional work load to lower functionaries without corresponding increase in personnel or if the implementation design interferes with the norms at lower level. It also incorporates the community (resistance to state penetration as Joel Midgal may call) in its model, sharing the responsibility of the outcomes between the education bureaucracy and the community.
Understanding framework of state capacity helps us realize the nature of challenges. It is the first step towards reform. The next step should be to leverage this model to device a framework for approach to reform. One such framework is proposed in my book UnpackED: The black box of Indian school education reform, pages 186–204. It might have to be tweaked around a bit. I will post the updated framework here as and when it takes a concrete shape.