Public programs are designed on assumptions - nice, tidy, convenient assumptions. Then they hit the real world and very little goes as planned. The culprit, some philosophically inclined would argue, is human behavior. After all, human beings are impossible to predict. They can react in ways entirely unexpected and fairly baffling …
… until you dig deeper.
We found this happening in one of the most commonplace and straightforward of public programs: textbook provision to schools. In 2008, the Government of Sierra Leone implemented a program for delivering textbooks to primary schools – one core set of textbooks for each student. There were not enough books for all primary schools and the government was interested in establishing the relationship between textbook provision and learning outcomes, so implementation was done using a randomized trial. The school sample was randomly divided into a treatment group (books) and a control group (no books).
Straightforward, right?
The first touch of mild-ish puzzlement came when we found no impacts on student performance. This was disappointing but not entirely unexpected. At least one other study – in Kenya – found similar results (at least for most students).
What was stranger – at least to the naïve researchers – was what had happened to the program textbooks. In many treatment schools, program textbooks had not been distributed to students; instead they were being stored on school premises. They hadn’t disappeared; we found them! On follow-up visits, we saw storage rooms filled with textbooks, while in classrooms students were still sharing - 3 or 4 students to 1 textbook (or worse). What’s more, students in treatment schools were not being allowed to take textbooks home.
What? What were head teachers thinking keeping textbooks from students who clearly needed them? What could prompt head teachers to (uselessly?) let these textbooks gather dust in some storage room? We explored several theories and carefully analyzed the correlates of book storage. We asked teachers if they thought the books were useful; they did. We asked students if they already had books at home; they didn’t. It turns out that head teachers who had high uncertainty about government transfers of textbooks were significantly more likely to store them: We measured this uncertainty by whether or not the head teacher knew how many textbooks were allocated to the school the previous year by the government.
Once we took this underlying ‘uncertainty’ into account, the storing behavior started to make sense. Think of consumption theory: if there is uncertainty about future transfers then current transfers are seen as a one-time (or transitory) shock. If this is the case then the impact of current transfers on current consumption (defined as intensive use of textbooks by students) will be limited. This is because when there is uncertainty, head teachers have incentives to store part of the current transfers in order to smooth consumption over time, particularly for inputs like textbooks that depreciate quickly when used. Remember the Permanent Income Hypothesis? This is a permanent public input hypothesis.
Taking ‘uncertainty’ into explicit consideration makes the seemingly irrational, rational. Clearly, head teachers in this context are acting as forward looking agents who – in the classic spirit of homo-economicus – are deciding how to respond to the public program based on their resources, constraints, and expectations.
And these results illustrate what this decision-making looks like when prevailing expectations around government largess and reliability are bleak. And why wouldn’t they be bleak? Our survey data shows that in 2009, 17 percent of the head teachers and 36 percent of classroom teachers reported not receiving their full pay in the last year. In Zambia, the unpredictability of government transfers to line ministries was well documented a decade ago. Anecdotal evidence tells us this is widespread.
This smoothing behavior isn’t unique to textbooks. In a project providing grants to schools in the Gambia, schools “were directed to use the grant towards some aspect of the school development that relates directly to teaching and learning.” Yet a quarter of schools reported their biggest expenditure was on infrastructure, durable inputs rather than quickly depreciated learning materials.
Is this simply ill-informed decision-making OR in the face of uncertain future transfers, are school committees trying to convert current transfers into goods that can be consumed not just in the present but also in the future?
When program goals are subverted by beneficiaries (or intermediate agents, as in this case), it is common to blame low capacity, myopic decision-making, lack of information, human error, corruption, or increasingly (and more charitably) the poverty-induced burden on mental bandwidth. But what we find here is rational behavior.
Our modest argument is that public provision of inputs that depreciate quickly when used, like textbooks, will be fully consumed only if agents have expectations of replenishment. Hence, for public programs that are designed as recurring transfers, there is a need to establish a reputation of consistent delivery and to reliably communicate the timing of the next transfer to agents. Otherwise, an environment plagued with uncertainty can easily wreak havoc on the nice, tidy, and convenient assumptions underlying program design.
… until you dig deeper.
We found this happening in one of the most commonplace and straightforward of public programs: textbook provision to schools. In 2008, the Government of Sierra Leone implemented a program for delivering textbooks to primary schools – one core set of textbooks for each student. There were not enough books for all primary schools and the government was interested in establishing the relationship between textbook provision and learning outcomes, so implementation was done using a randomized trial. The school sample was randomly divided into a treatment group (books) and a control group (no books).
Straightforward, right?
The first touch of mild-ish puzzlement came when we found no impacts on student performance. This was disappointing but not entirely unexpected. At least one other study – in Kenya – found similar results (at least for most students).
What was stranger – at least to the naïve researchers – was what had happened to the program textbooks. In many treatment schools, program textbooks had not been distributed to students; instead they were being stored on school premises. They hadn’t disappeared; we found them! On follow-up visits, we saw storage rooms filled with textbooks, while in classrooms students were still sharing - 3 or 4 students to 1 textbook (or worse). What’s more, students in treatment schools were not being allowed to take textbooks home.
What? What were head teachers thinking keeping textbooks from students who clearly needed them? What could prompt head teachers to (uselessly?) let these textbooks gather dust in some storage room? We explored several theories and carefully analyzed the correlates of book storage. We asked teachers if they thought the books were useful; they did. We asked students if they already had books at home; they didn’t. It turns out that head teachers who had high uncertainty about government transfers of textbooks were significantly more likely to store them: We measured this uncertainty by whether or not the head teacher knew how many textbooks were allocated to the school the previous year by the government.
Once we took this underlying ‘uncertainty’ into account, the storing behavior started to make sense. Think of consumption theory: if there is uncertainty about future transfers then current transfers are seen as a one-time (or transitory) shock. If this is the case then the impact of current transfers on current consumption (defined as intensive use of textbooks by students) will be limited. This is because when there is uncertainty, head teachers have incentives to store part of the current transfers in order to smooth consumption over time, particularly for inputs like textbooks that depreciate quickly when used. Remember the Permanent Income Hypothesis? This is a permanent public input hypothesis.
Taking ‘uncertainty’ into explicit consideration makes the seemingly irrational, rational. Clearly, head teachers in this context are acting as forward looking agents who – in the classic spirit of homo-economicus – are deciding how to respond to the public program based on their resources, constraints, and expectations.
And these results illustrate what this decision-making looks like when prevailing expectations around government largess and reliability are bleak. And why wouldn’t they be bleak? Our survey data shows that in 2009, 17 percent of the head teachers and 36 percent of classroom teachers reported not receiving their full pay in the last year. In Zambia, the unpredictability of government transfers to line ministries was well documented a decade ago. Anecdotal evidence tells us this is widespread.
This smoothing behavior isn’t unique to textbooks. In a project providing grants to schools in the Gambia, schools “were directed to use the grant towards some aspect of the school development that relates directly to teaching and learning.” Yet a quarter of schools reported their biggest expenditure was on infrastructure, durable inputs rather than quickly depreciated learning materials.
Is this simply ill-informed decision-making OR in the face of uncertain future transfers, are school committees trying to convert current transfers into goods that can be consumed not just in the present but also in the future?
When program goals are subverted by beneficiaries (or intermediate agents, as in this case), it is common to blame low capacity, myopic decision-making, lack of information, human error, corruption, or increasingly (and more charitably) the poverty-induced burden on mental bandwidth. But what we find here is rational behavior.
Our modest argument is that public provision of inputs that depreciate quickly when used, like textbooks, will be fully consumed only if agents have expectations of replenishment. Hence, for public programs that are designed as recurring transfers, there is a need to establish a reputation of consistent delivery and to reliably communicate the timing of the next transfer to agents. Otherwise, an environment plagued with uncertainty can easily wreak havoc on the nice, tidy, and convenient assumptions underlying program design.
Join the Conversation