Innovation in government to improve services and lower costs

By Arnaldo Cruz- Co Founder of ABREPR

Despite having some flexibility with program implementation, in most cases government decisions in Puerto Rico are arbitrary. That is, decisions are made without any evidence that it is the best alternative for the government. Alternatively, the government could base its public policies, laws and regulations using empirical evidence rather than a guessing or a hoping for the best philosophy. It might sound complicated, but in this essay, we present an idea that with very little efforts could improve government performance, modify the behavior of employees, contractors and constituents, without requiring legislation or regulation. That concept is called nudge, developed by Nobel economics laureate Richard Thaler. The concept is inspired by the ideas of the behavioral economics, a subfield of economics, that explores why humans are not as rational as in traditional economic models.

A nudge is a relatively subtle policy change that encourages people to make decisions that are in their own best interest. Should I go to the doctor? Should I pay my taxes? Should I save for retirement? A nudge changes the way those options are presented making people more likely to choose a desired outcome, like saving more or paying overdue taxes. Although a nudge alters people's behavior in a predictable way, it does not prohibit their options or significantly change economic incentives. In other words, a nudge is not a mandate. Including the voter registration form when you renew your license is a nudge, requiring by law to register is not. To be consider a nudge, the intervention must be easy, not forced and inexpensive to the individual.

The nudge concept has gained worldwide attention, thanks to years of work done by experts in economics, psychology and law. This popularity has inspired the creation of nudge offices in several governments that use nudges to create or modify public policies. The UK's Behavioral Insights Team (BIT) was arguably the first formal government-led nudge unit. By 2010, the BIT was composed of a group of eight people. Since then, their work has contributed to a $70 million per month increase in government revenue, 96,000 additional citizens per year register as organ donors and better racial diversity in the police force, all with a budget of less than $654,000 a year. By the end of their sophomore year, the BIT had saved the government 22 times the cost of operating the office. Today, the BIT is a non-profit company, in partnership with the UK government and the Nesta Foundation. Countries such as India, Indonesia, Peru, Singapore and many others have begun to apply nudge to their policies and programs. International institutions such as the World Bank, IMF, OECD and EU have also established nudge units to support their programs.

To better understand the concept, we will provide several examples. Great Britain's Treasury Department traditionally sent letters to companies who were delinquent with their taxes. Historically 33% of companies paid their balance once they received a letter from Treasury. The BIT unit decided to perform an experiment by sending letters with different messages to various companies in order to identify which turned out to be the most effective at getting them to pay outstanding balances. For example, some companies were sent the traditional letter while others, a letter with a message stating: "7/10 businesses in your neighborhood are in compliance, you’re the minority." It turned out that 39% of the companies that received the second letter paid their taxes, versus 33% that received the traditional letter. Once the Treasury sent the second letter to all companies it recovered about $200 million over the next 12 months. In Polonia, this nudge increased payments by 17%.

In another example, the government of Peru established its nudge unit in 2014 (MineduLab), focused on education. One of their most successful experiments was related to the attendance of teachers and school principals. Advised by MineduLab, the education department sent different emails to teachers and school principals warning them about their absenteeism and comparing their absences to their colleagues in their nearby school district. Although the emails had no impact in teacher attendance, they did increase the average attendance of school principals, from 83% to 87%. This corresponded to approximately 7 more days at the school per principal, a substantial effect, especially when it didn’t cost much to the Peruvian government.

Another good example of nudge comes from the Canadian government with organ donor registrations. The government’s nudge unit tested several interventions with the donor registration process. They developed simpler versions of the registration form, altered the timing to approach potential donors, and changed the messaging of their advertising. The government tried six different strategies. This allowed the government to compare registration rates with a control group, to see which interventions had the greatest impact. In the end, the government identified 3 strategies that were able to increase donor registrations by 143%. As a result, they extended these changes across the state, which then increased the number of registered donors by several hundred thousand.

In Puerto Rico we could also benefit from this concept. For example, the Department of Transportation and Public Works may communicate to employees at CESCOS (DMV’s) how efficient their office is in relation to other CESCOS in the area. Because the vast majority of CESCOS now have an electronic ticketing system, the data could be extracted, analyzed and communicated to all employees, and made available at citizens at the waiting room. This could incentivize employees to provide better service and optimize the work of the office. This same concept could apply to government permit offices, costumer service offices of public corporations such as PREPA, employment offices and even police headquarters. These nudge techniques could also become a powerful tool for dealing with employee absenteeism at the agencies.

Another great implementation opportunity could be in the ‘s health system. The government could work with private insurers to include nudges in the system that could improve patient health and reduce costs. For example, studies have found that patients are more likely to be adherent to generic medications, but physicians often prescribe the brand name medication. By displaying generics as defaults in the electronic health record, you increase the generic prescribing rate and thus save millions of dollars. As well, you can design prevention programs to shape choices of Medicaid patients so that healthier lifestyle alternatives become a thoughtful deliberative decision, are easier to choose, or feel less costly in terms of time, emotional commitment, or risk to the patient. The Health Insurance Administration (ASES), a public corporation that administers the Medicaid program, has every insurance claim for all Medicaid recipients and providers in an electronic database. Using data science, the government could find patterns of sevice overuse in the system, either from the patient's perspective (patients who constantly go to emergency room for routine symptoms) or service providers (doctors that prescribe too many antibiotics). Once you find these patterns, you can send emails or texts to incentivize a change in behavior. For example, patients with recurrent emergency room visits could have their primary doctor appointments automated (primary care visits cost a third emergency room visit). You can also make changes to rules, forms, or processes. All in order to find the solution that brings the best results at the lowest cost.

In Puerto Rico there have been some variations of nudge in the past. For example, in 2013, the government decided to pay the fee of the college entrance exam (College Board) to all public-school students. Between 2009 and 2012, the number of students taking the exam ranged from 20,144 to 18,464, respectively. Meanwhile, in 2013, some 26,485 students took the test. Subsequently, in 2014, that number increased to 55,446 students, a dramatic increase associated with this policy change. Notwithstanding, this cost the government between around $5 million. A question the government should have asked was whether the cost was the only barrier to students taking the test. That is, if the government had done another type of nudge with parents and students, such as sending text messages indicating the importance of the exam or filling out the application on behave of the students using data from the Department of Education. Perhaps that type of nudge would also have increased the participation rate, without increasing costs for the government. The truth is that we do not know, since no analysis or experiment was conducted before making the decision. Although we are only talking about $5 million in this case, think about all similar decisions the government makes every day based on a hunch that cost way more. In the other hand, if we knew that these nudges actually work, we could apply them to college applications and financial aid. Let's imagine that we automatically fill out FAFSA applications for all public-school students using existing data from Treasury and the Department of Education. How many additional students could have access to financial aid for college?

Examples such as these exemplify the huge benefits a nudge unit can bring to the government and the citizens who receive services at a nominal cost to the taxpayer. A nudge unit can help the government identify (evidence-based) the best programmatic strategies and design options in a pilot scenario before applying it to the entire Island.

It's not that difficult to get started.

Most units have started with two or four employees. Profiles include public policy experts, psychologists, sociologists and economists. The important thing is that there is a commitment from the government and some patience, as the results do not arrive immediately.

Where should this new unit reside? Difficult to think of a government agency that has the culture, resources and expertise to do this type of work. In many countries the office has been linked to the office of the president directly. An alternative is for the unit to be a joint project between the executive branch and the Fiscal Oversight Board. That way there is mutual responsibility, both for funds and for results. Having the Board and the governor committed will help advance the agenda, integrate government agencies and navigate the authorization process of the experiments efficiently. At a minimum, the office should begin with a commitment of funds of at least 3 years, which provides enough time to show results.

Before the new office can start doing experiments, it needs to establish the organization structure, which includes recruiting the team, establishing an advisory committee (government, academia, research centers and the private sector), raise awareness of the concept, identify the 3-4 public policy areas to start and identify the data sources needed to measure, evaluate and design experiments.

Because of the historical moment Puerto Rico is going through right now, it is imperative to advocate for data-driven and science-based decision-making to improve the public policies that affect us all. For many reasons, this sorry state of public sector innovation cannot stand in Puerto Rico. We must pave the way for innovation, test novel solutions to problems on top of existing structures and generate compelling evidence for how to best serve the people of Puerto Rico.

Arnaldo Cruz