Practice Insights > Notes on Strategy

Analysing Trade-offs in Development Practice

By V Santhakumar

1. Introduction

In the development practice, is there a trade-off between the cost of an action and the benefit from it? When one has already decided to spend a certain amount of money or time, or both, to help others – this is true even for those who have decided to devote their lives to development practice and may have sacrificed a better income elsewhere – why should they consider costs and benefits of it? Will such an analysis not deter them from doing good if the tradeoff is discouraging?

For anyone, the resources that they can spend on helping others are limited. This is true for altruists and the government too. Because all needs cannot be met with the available resources, these should be spent on the most important needs. When the most important needs are met, the beneficiaries themselves may be able to address some of the secondary needs. And whatever be the personal objectives of the development practitioners, there will be a greater impact of their actions on the society if their time and money are spent on more beneficial actions. Hence, a consideration of how to spend the money by comparing alternative means – or trade-offs– is useful in the development practice. There can be some generic considerations in this regard.

2. Addressing the root cause can be difficult

While trying to help others, there may be a temptation to help their immediate problems – providing food or treating a disease or giving money to tide over an emergency crisis. These problems that affect the survival of people cannot be overlooked. However, there can be various underlying causes for many such problems, such as unemployment, lack of skill or low productivity, or undesirable behaviour. In many cases, spending money to address the root causes of problems will not only ensure the long-run welfare of the beneficiaries but is also cost-effective, that is, beneficiaries can be helped with a lesser amount of money and therefore the help can be extended to more people. For example, there is a set of people who are provided with free medical treatment whenever they are sick. Understanding the behaviours or factors that are affecting their health and enabling them to change those, may reduce the need for such free or subsidized treatment, which though cannot be avoided when they are sick.

However, there is another aspect of this issue. Providing emergency help is easy, but addressing the underlying cause can be more difficult. For example, running de-addiction clinics for alcoholics or drug addicts can be cheaper and easier than providing treatment for their health-related issues and for solving the other ill effects of substance abuse on them. These other issues may require the addressing of both the immediate and the root-cause problems. For example, the immediate action in the case of domestic violence against women is to rescue them but the long-term, root-cause solution would be to empower them to be able to earn and live independently.

These two aspects – the need to address the underlying cause and probably the greater difficulty in doing so – are the important considerations for the trade-off involved in development practice. There may not be any general principle in this regard and the desirable course of action may vary depending on the context.

3. The Need to focus on workable solutions

A related issue of the trade-off is the selection of workable actions or the practical considerations in selecting the means of development practice. Many development practitioners are aware of the complexity of the problem at hand, and the need to address/attack it on multiple fronts, but the considerations of practicality may dissuade them to focus on more than one or a few means. Providing a simple solution to the maximum number of people is a common practice. For example, extended provision of deworming tablets to children is more workable than, say, doing something to reduce open defecation in the area (which may be the major cause of worm infections). Even well-meaning development practitioners are interested in creating a visible impact (preferably, in the shortest possible time) and hence workable means have a greater attraction. There is no justification for spending money on unworkable strategies without any certainty of outcomes. However, the downside of this approach is that many people may shy away from attempting difficult actions which may have a long-lasting impact. This is a personal choice for many development practitioners and altruists. How far one is willing to tolerate uncertainty in trying out difficult and untested routes of improving the lives of other people is the crucial question here. However, the trade-off in this regard need not remain static. More information; better understanding of effective ways (say, on how to tackle open defecation); and, the knowledge of developing technologies that improve the reach to the beneficiaries may change the trade-off in this regard, over time.

4. The Need to consider alternatives to reach a goal

It is obvious that there is a need to be concerned about the effectiveness and efficiency in development practice and altruist actions, that is, to assess that we get maximum social benefits from a given amount of resources that are spent. In other words, we spend the minimum amount of resources for a given level of benefits. Therefore, in order to be effective and efficient, considering alternatives is important. There are many ways of improving the health of pregnant women from poor families (so as to reduce maternal and/or infant mortality) – giving out cash, supplying more food items at reduced prices and supplying cooked food to Anganwadis, to name a few. All of these may not be effective and efficient in every given context. By going ahead with pre-conceived ideas, we may be neglecting more effective and efficient means, leading to a wastage of social energy and resources, which could have been used elsewhere. Hence, it is important to explore and identify a number of alternative solutions to address a particular social problem. Then, the most effective and efficient solution needs to be identified through the consideration of the social benefits and costs.

The assessment of social benefits and costs is a tedious process. Though this ought to be carried out transparently by government agencies for their public projects¹, it is seldom done and the decisions hardly ever take these into account². Moreover, there may be a need for technical or economic expertise for carrying out such an assessment. In this note, the attempt is to highlight a few lessons on assessing social benefits (and costs, where relevant) that must be kept in mind while planning development actions.

5. Assessment of social benefits is different from financial analysis

This assessment of social benefits and costs of altruistic or development actions is different from (though there are similarities with) the financial analysis carried out by individuals and private firms regarding their investment decisions. In a financial analysis of any investment, the costs borne and revenues received by the individual or organizations are analysed to see whether the revenues received justify the costs incurred. However, an altruist spending money on the education of children from poor families or on the healthcare of a deprived section of the society may not get any revenue (sometimes, a small share of the cost may come as revenue). It is, therefore, obvious that a financial analysis will show most development actions as loss-making. Hence, social benefits are assessed not through financial analysis but by considering all the benefits accrued to the society and all costs incurred by it.

5.1 What are social benefits?
The increase in the happiness or satisfaction of anyone in society as an outcome of an altruist action should be considered as a part of the social benefit. It is easy to identify if the intervention leads to the increased consumption of items or services demanded by the people. It is not very difficult to understand this benefit when the intervention leads to increased use of education or healthcare by the people who were not getting these services adequately before the intervention or if some people start earning more. Similarly, if the intervention is to create a better-quality road (so that time used for travel can be saved), it may also enhance the utility or happiness – since the time saved can be used to earn an income for higher consumption or have more leisure. This is true in the case of creating a facility that provides entertainment or intellectual enrichment like a library. If the intervention leads to the creation of a park, or the preservation of a small patch of forest, that can also be counted as increasing the consumption (or happiness arising out of the consumption) of certain environmental amenities and services. All these interventions enhance the consumption (or utility/happiness) by a section of the society and hence, have social benefits, even if the interventionist or the development practitioner does not get any revenue or money back³.

5.2 Quantifying benefits
Once we identify the benefits, the next step is to quantify each of these. Here, we may be concerned about the physical quantities, the number of water litres per person, per day to be made available; the additional number of children who can be provided schooling of a pre-determined quality; the reduction in dropout rate; the number of pregnant women getting additional nutrition; or, the reduction in infant mortality rate, are some examples of the benefits quantified. We may note that development interventions usually make an additional impact like more children are brought to school or the learning outcomes of children who are in school are improved. It is this `additional’ impact or the impact over the situation `without’ the intervention that we are interested in. Hence, the analysis of an intervention is always in comparison with the situation of `without’ the intervention. However, there is a common mistake that people commit here. We may consider `without’ the intervention situation as equivalent to that before the intervention. This need not be correct in all the cases.

Even without the intervention, social context is changing or evolving due to other factors. Let us take the example of an intervention to improve the learning achievements of children in a district. This is attempted through training teachers and/or by distributing better learning materials. The intervention is expected to make a significant impact in two years from the start of intervention (say, 2014). One may be tempted to think that the comparison is between the learning achievement in 2014 and those in 2016 (so that the aim is to reach an achievement in 2016 which is significantly higher than that in 2014). This is incorrect. What is to be compared is the achievement in 2016 without the intervention and in the same year, with the intervention. If we are comparing the achievement in 2016 with that in 2014, we are presuming that there has not been any change in the achievement between 2014 and 2016 in the absence of the intervention. But in reality, even without such an intervention, there may be some change in the learning outcomes through the impact of factors other than the proposed intervention, for example, the education of an elder sibling, the income of the household, and similar factors may have a positive or negative impact even without the intervention. This is described in Table 1. This change needs to be accounted for, and it cannot be attributed to the intervention. If we do so, there is an `attribution error’.

Table 1

*Start of the intervention
^Completion of the intervention

Here, learning achievement is defined as the average percentage of children who pass standard tests for a particular grade or level. It is clear from the table that the actual impact of the intervention is 23 (60-37) and not 32 (60-28) – the result that we get if we compare the situation with, and before the intervention. Hence, we need to forecast or predict the situation as it would be in the absence of the intervention. There are many sophisticated methods of forecasting/prediction, but these are not discussed here. A simple way is to extrapolate from the situation for years before the starting of the intervention. We may know the learning achievements in the previous years – 2010 to 2014. These can be extrapolated (graphically or statistically) to predict the situation in 2016 (that is, without the intervention). Then, it can be compared with the actual situation in 2016 with the intervention. In certain cases, we may be comparing the situation before and after the intervention. Here, the hidden assumption/expectation (which needs to be articulated) is that there would be no change in the status quo if there were no intervention. In the example mentioned here, this would mean that the learning achievements in 2014 and 2016 remain the same in the absence of the intervention. There could also be a deterioration of the situation in the absence of the intervention, and in such situations, the impact must include the avoidance of such deterioration and not just an improvement in the outcome.

The fact that all benefits may have to be assessed does not imply that all such benefits have to be quantitative in nature. For example, the measurement could be in terms of changes in perceptions, like, currently 80% of the parents in a village are unhappy with the performance of the village school, and the objective of an intervention could be to reduce it to 30%. This reduction in the negative perception could be the potential benefit of an intervention. However, some measure of assessment – either quantitative or qualitative – may be needed here.

5.3 Incremental and non-incremental benefits
The benefits created through an intervention can be categorized into two. Let us consider one simple example, the people in a village suffer a shortage of drinking water. They (usually, women and girls) travel a substantial distance to collect some quantity (say, 10 litres per person, per day which is not adequate for all their needs) of water during summer. It is in this situation that a development practitioner aims at implementing a water supply scheme, distributing 40 litres per person, per day. There are two benefits of this scheme. First, the cost or the effort of collecting the 10 litres, which they were now getting has come down. They get it closer to home without having to walk the long distance. This is called the non-incremental benefit. The benefit of this part is the cost/effort saved in getting the 10 litres of water from a distant source when the same quantity is available in the village through the new scheme. Secondly, each person gets 30 additional litres of water per day. This is the incremental output. Earlier, people wanted to get more water but it was not available. The intervention made it available. That is the benefit of this incremental part. Total benefits include non-incremental and incremental parts. However, we have not valued this benefit yet. That is taken up in the following section.

5.4 Valuation of benefits
Let us take the example of a village where people do not have drinking water and there is no healthcare centre. People travel four to five kilometres to collect some minimal amount of water during summer. The cost associated with travelling to the private clinic located in the nearby town is sufficiently high and so 90% of them do not go there when they are sick. A development practitioner is interested in spending a fixed amount of money for these villagers. What should be the priority? Solving the drinking water problem or the absence of a health clinic in the village? There may be several cases of this kind, and a comparison may be required. Here, we are comparing social benefits of different kinds – say, that between the availability of more drinking water and that of healthcare. We know that two entities can be compared only if they are of the same unit. Hence, these social benefits may have to be converted into the same unit for comparison. The unit normally used is money. We may be able to compare these two benefits if both can be converted into money.

Some issues related to the valuation of benefits are discussed here. We are not presuming that all development practitioners are in a position to (or need to) calculate of social benefits (and costs). However, it may be desirable if all understand the basic ideas underlying the valuation of such benefits.

What is the value (in terms of money) of a social benefit? If we go back to the example of water supply, the value of non-incremental benefit is the cost or effort saved – let as assume it as A. If we have some information on the cost or effort that people spend on collecting water from a longer distance, the savings in this regard can be reckoned as the value of this benefit. This is relatively easy to calculate. However, what about the benefit of getting additional water which they wanted but was not available earlier?

Let us assume that the cost of providing this additional water (30 litres) for the development practitioner was 0.1 rupees per litre. Hence, he charges the same amount as the price of this water. Does this price indicate the value of each additional litre of water available to the people there? If not, what else is the value? When this water was not available, people would have been willing to pay an amount to get some more water. For example, let’s assume that they would be willing to pay one rupee per litre for one more litre of water when they were bringing only 10 litres from the distant locality, in other words, they were willing to pay rupee one for the 11th litre.

It should not be that difficult to understand this concept of willingness to pay. Assume that we as urban consumers face a severe shortage of water supply on a day due to the bursting of a pipeline in the locality. We may be then willing to buy bottled water at rupees 10-15, per litre to meet our basic needs. Hence, our benefit of regular municipal water supply includes even this amount (Rs 10-15 per litre) that we save, since we don’t have to spend this much money due to the municipal supply.

Coming back to the example of villagers, they may be willing to pay Rupee one, per litre for the 11th litre. However, they may not willing to pay a similarly higher amount for the 12th litre. It could be lower since they could meet a part of severe needs for water by consuming the 11th litre. This trend may continue. What they would be willing to pay (WTP) for the 30th litre, could be significantly lower than the WTP for the 11th litre.

Hence, the value of the incremental benefit is to be assessed on the basis of what people are willing to pay (WTP) for this additional amount of water. As noted earlier, the WTP may come down as more and more water is available. An easy way to calculate this value would be to take an average of what they are WTP for 11th L, and 30th L, and multiply it with 30 to get the total amount that they are WTP for 30 additional litres. Let the value of this incremental benefit be B. Hence, the value of the total benefit of the new water supply scheme is A plus B.

We have taken a simpler case here to discuss how to measure the value of a social benefit but there are more complex situations, such as, what may be the value of the social benefit of creating a park in the city by an NGO, an individual or a corporate?

People want certain goods and services. Under normal circumstances, they would be willing to pay or willing to make the effort or sacrifice something else, say leisure to get these. If they cannot afford to pay for an item or service that they need, they may `suffer’ from the lack of it. For example, people with a disease need medicine, and if they cannot buy it, they will suffer. By looking at the situation and the costs (including the suffering) that they bear, one may get some indirect picture of their willingness to pay, say, for the service of a doctor. However, the fact that somebody demands (or is willing to pay for) a service does not necessarily make it available. There are many real-world situations where a service is not available even if there is demand. For example, in villages, there are many people with dental problems, but no dental care is available. The reason for the absence of the provision of such a service could be that the amount the people are willing to pay for it is lesser than the price at which it can be supplied. This is especially so since there may be fixed costs in supplying a service in a locality, and if there are not enough potential consumers willing to pay adequately, no service provider may be willing to start serving there. This could be one area for development practice. However, the benefits of such an intervention are assessed by looking at what people are willing to pay for a specific amount of this service and/or what they bear due to the absence of the service.

5.5 What if the benefits cannot be valued?
For a number of development actions, it may not be possible to assess the benefits in terms of money. For example, the benefits cannot be measured easily for a project to reduce infant mortality rates or to improve the nutritional status of children in Anganwadis. There are complex procedures for doing such measurements and are difficult to apply in many cases. However, this does not mean that we need not be conscious of the trade-offs involved in such projects. In the case of certain interventions, the benefit can be quantified even if it cannot be valued easily. The number of children vaccinated, the number of pregnant women whose nutritional status has been improved, the number of teachers whose quality is enhanced, etc., could be examples of such quantifiable benefits.

In such cases where benefits cannot be measured, one strategy could be to use the cost-effectiveness approach. Here, we use the means, which minimize the cost (or resources) to reach a particular objective or select those means which give maximum output for a given amount of resource. For example, there can be several ways to improve the nutritional status of the children attending Anganwadis. These may include, providing cooked meals; supplying ready-made products like nutritious biscuits and vitamins; giving additional food materials to their households; transferring additional money to their parents; etc. The cost-effectiveness approach should enable us to select the cheapest possible means to enhance the nutritional status. By using minimum cost here, it is expected that the resources saved can be used for other equally important social needs. However, here we should be in a position to quantify the outcome (or improved nutritional status). For example, the number or percentage of children in an Anganwadi who a physician considers as `malnourished’ can be the benchmark indicator, and the progress in this regard through the intervention can be measured as its output or benefit. In certain cases, technically sophisticated measurements may be needed. For example, interventions that lead to a general improvement in the health of a population may have to be measured in terms of the number of additional years of healthy life (or indicators, such as healthy life days or disability-adjusted life years) achieved through such interventions. There should be a reasonable indicator of the benefit even if all dimensions of it cannot be measured. If the intervention is to enhance the learning levels of children, their performance in standardized tests with and without intervention can be used as a measure of the outcome or benefit.

The costs of two projects which aim at different objectives cannot be compared easily. Consider the following three projects: (i) improving the vitamin A deficiency of 50 children, (ii) improving the vitamin A deficiency of 70 children and (iii) improving the vitamin A and Protein deficiency of 50 Children. It is not easy to choose the project with the minimum cost among these three because these are projects with different objectives. Similarly, the cost-effectiveness of the same health intervention carried out on two sets of the population could be different since, the additional improvement in terms of health, that can be achieved may vary between different populations depending on a number of other people-specific (including social and/or context-specific) factors.

When each of the alternative interventions considered makes a different level of impact, how do we pick the least-cost or cost-efficient one, is an issue. In certain cases, a simple normalization procedure may be used. For example, one may calculate the cost per unit of outcome (like the number of children vaccinated, the percentage of average increase in test scores for the children in a school, etc.) For example, if project A worth 2 lakh rupees can address the malnutrition of 100 children, and project B worth 5 lakh rupees is adequate to address the same problem for 1000 children, simple ratios of cost and output, or cost per child, under projects A and B can be compared.

Author

V Santhakumar, Professor, Azim Premji University