How time subsidies increased fertilizer use by up to 11%

Intervention · Business


Historically, fertilizer use has been low in Africa, even though there is evidence that it is associated with greater attention to weeding, better seed, improved irrigation practices, and generates high returns on investment.1 This partly because behavioral biases lead to a preference over immediate gains rather than investments over a longer period of time. Taking a behavioral approach yields a better way to solve this issue, especially when considering that this behavior is the opposite of what we’d expect to see. 

Duflo and researchers created three implementations of the SAFI program, all conducted in Western Kenya. In this program, farmers are offered free delivery of fertilizer. The first condition is the program as it’s meant to be run at the time of harvest. In the second condition, they offered the same program at harvest but during the heavy rain season. Lastly, they offered the program before the harvest began, testing whether farmers were present-biased.


Effective interventions start with a nuanced understanding of how decisions are made. Our mission is to help large organizations be better and do better, using behavioral science.

Learn about what we do

Rating: 4/5 (Some significant effects; Positive impact; Expensive for policymakers; Effects didn’t last)

Variations of the SAFI Program on Fertilizer Use Among Western Kenyan Farmers
Without SAFI program  34% of farmers in the control group used fertilizer
With SAFI program 45% of those offered fertilizer deliveries reported using fertilizer

Key Concepts and People

Hyperbolic discounting: Our tendency to choose immediate rewards over future gains, in spite of the latter being greater. 

Libertarian paternalism: First proposed by Richard Thaler and Cass Sunstein, it is a behavioral change methodology that maintains both the ability to alter behavior without coercion and the individual’s right to choose.

Esther Duflo: The main researcher of this paper, Esther Duflo has dedicated her career to researching the economic environment of the world’s poorest people. She won the Nobel Prize in Economics in 2019, achieving the title as the youngest person ever to win the award, in collaboration with Michael Kremer (another co-author) and Abhijit Banerjee.

The Problem

Subsidizing fertilizer can have downsides

In order to increase agricultural output in developing countries, many researchers point to modern solutions. In particular, fertilizer. It can explain the incredible growth in Asia’s agricultural yields compared to Africa’s stagnant yields. While it might initially make sense to give out more fertilizer, that can pose its own problems: subsidies often have unequal benefits, as wealthier farmers benefit the most. And loans to buy fertilizer are often given out to those who are politically connected. Subsidies can also lead to greater government involvement in the distribution, politicization, and selection of fertilizer.

Low investment in fertilizer, despite its high returns

In the context of the anti-subsidy position of international financial institutions and economists, a better answer may be to give out fertilizer. In a previous study by the same researchers, they distributed fertilizer to farmers in Western Kenya in limited quantities and showed returns of 36% over a season on average, which translated to an annual return of 70%.2 However, there has still been low investment among this population. 

Can discounts overcome biases?

The research team proposed that behavioral biases are responsible, namely procrastination, hyperbolic discounting, and naivety. Utilizing the principles of libertarian paternalism, they predicted that small, time-limited discounts may result in higher welfare compared to laissez-faire policies or heavy subsidies.1


The Basic SAFI Program

In essence, the SAFI program is as follows: at harvest, farmers are offered free delivery of two types of fertilizer. A field officer visits immediately after harvest and offers to sell the farmer a voucher to buy fertilizer (at regular price) with free delivery. The farmer decides during the officer’s visit whether or not to participate in the program. They were not limited to the amount of fertilizer available for purchase. 

This program helps to reduce the utility cost of fertilizer use in two ways. First, it saves the farmer the (on average, 30-minute) trek to the market. Secondly, it saves the farmer time trying to decide the type and quantity of fertilizer they need to get. 

The two program timelines

Researchers ran the basic SAFI program during two time periods: the first was in 2003 after the short rain harvest, in order to allow farmers to fertilize in anticipation of the 2004 long rain season. The second occurred at harvest time after the short rain season in 2004. In Kenya, the two agricultural seasons are the more productive long rains (spring through summer), and the less productive short rains (summer through winter).

The second version of the SAFI program, run at harvest time after the shorter rainy period in 2004, is identical - except field officers visited farmers before the harvest. They asked farmers when they would want the officers to return to receive a SAFI program. Participants for all conditions were randomly selected from a sample of parents of fifth and sixth grade children in sixteen schools throughout Busia County, which borders Uganda and Lake Victoria.

The EAST framework

Offering free delivery of fertilizer not only makes it easy for farmers to obtain the product, but it also saves them significant time that could be spent doing other things. Further, the SAFI program is an attractive option for Western Kenyans looking to make a wise investment. Given the qualities of the intervention, it’s clear that the EAST Framework influenced researchers’ decisions. This framework highlights four ideals for policy making: a policy should be Easy, Attractive, Social, and Timely. 

Results and Application

The results suggest that the SAFI program had a significant effect on the use of fertilizer throughout the region. In the first season, 45% of the farmers offered the program reported using fertilizer in that season, compared to 34% in the control group.

For the second growing season, the program yielded an increase in fertilizer use of 10.5%, which was only significant at the 10% level. When fertilizer was offered before the harvest, as in the case of the condition based on forecasts rather than actual results, about half (44%) of farmers asked the field officer to return right after harvest, and 46% of those bought fertilizer in the end (about 20% of the total). Conversely, in the same condition, 52% requested late delivery, and 39% of those eventually purchased fertilizer (about 20% of the total). 

These findings suggest that a properly timed intervention that reduces the utility cost of using fertilizer can meaningfully enhance adoption rates. However, they did observe that the SAFI program did not have lingering effects. After two subsequent seasons, fertilizer use went back to the level of the comparison groups. While the SAFI program increased fertilizer use substantially, adoption still remained low. The program is also too expensive to be cost effective, but the authors urge policymakers to consider similar small, time-limited subsidies.

Public Policy Sending COVID tests by mail (for free) could increase how often people get tested and potentially reduce prevalence of the virus
Development and Social Protection Having working parents select groceries ahead of time and providing free delivery could help them spend more time with their kids and choose healthier foods
Retail & Consumer Policymakers could incentivize free delivery of environmentally-friendly home products (i.e. a plant-based sponge versus a plastic one), and continue to charge for traditional products


  • Research focused on a non-WEIRD (Western, Educated, from Industrialized, Rich, and Democratic countries) population 
  • Investigators considered the unique cultural climate of Western Kenya and respected their way of life
  • Participants were randomly selected, although it’s unclear why they chose parents of this age group

Does the intervention demonstrably improve the lives of those affected by it?
Fertilizer use can increase agricultural yield and farmer’s overall financial well being
Does the intervention respect the privacy (including the privacy of identity) of those it affects?
No identifying information was presented in the paper
Does the intervention have a plan to monitor the safety, effectiveness, and validity of the intervention?
Yes, the effects of the intervention were monitored for two seasons afterward

Does the intervention abide by a reasonable degree of consent?
Participants were free to leave at any time
Does the intervention respect the ability of those it affects to make their own decisions?
Farmers had autonomy to choose whether or not to buy and fertilizer
Does the intervention increase the number of choices available to those it affects?
Farmers were given more choices in fertilizer acquisition 

Does the intervention acknowledge the perspectives, interests, and preferences of everyone it affects, including traditionally marginalized groups?
Researchers reportedly maintained cultural sensitivity and had depth of knowledge of the cultural intricacies in the region
Are the participants diverse?
Room for Improvement
Participants were from the same 629 sq mi county and of similar socioeconomic status and race
Does the intervention help ensure a just, equitable distribution of welfare?
By creating greater agricultural yields, farmers could earn more money and contribute to a wealthier economy, which would benefit all

Related TDL Content

Should We Pay Students to Go to School? - Can subsidies keep kids in school? In a study conducted in Mexico, it turns out that they actually can reduce dropout rates — but not necessarily at improving other predictors of success like grades and attendance rates. Read more to find out why that is.

Behavioral Science and the Future of Agriculture - In the face of the climate crisis, farmers need to adopt new, sustainable technology. Unfortunately, farmers’ biases such as cognitive dissonance and heuristics can impede this progress. There is a solution, though, and it’s simpler than you’d think: education.

Nudging Towards A Sustainable Future - Even though Duflo and associates discovered that the SAFI program may be too expensive for policymakers, nudges may serve as a cost-effective solution to improve conservation and poverty alleviation efforts. 


  1. Duflo, E., Kremer, M., & Robinson, J. (2009). Nudging farmers to use fertilizer: Theory and experimental evidence from Kenya. National Bureau of Economic Research. 
  2. Duflo, E., Kremer, M., & Robinson, J. (2008). How high are rates of return to fertilizer? evidence from field experiments in Kenya. American Economic Review, 98(2), 482–488. 
Notes illustration

Eager to learn about how behavioral science can help your organization?