slide 1


Five Principles for Improved Systems Thinking

The disparities we witness in economic, social, and health outcomes are a manifestation of deep structural inequities in the United States’ social hierarchy. These structural inequities appear in various forms but are woven into the very fabric of American society and can be observed across our legal, educational, business, government and health care systems. They occur because of an imbalance in the distribution of political and economic power where favored groups have historically set the rules and limited equitable access of marginalized groups to wealth, information, connections, and resources.

Dismantling the structures that reinforce these inequities is a long game, particularly because inequity is embedded within American institutions and culture. We must understand where and why they emerge as a critical step towards addressing them. Therefore, change agents must embrace new approaches to thinking about knowledge and causality, systems thinking. A system is defined as an interdependent group of components working together as a whole to accomplish a goal. For example, police, lawyers, judges, courts, prisons, and come together to make up the “criminal justice system”.

In the remainder of this article, I will describe five principles of systems thinking that help make the dynamics and patterns of system behavior intelligible to inform our understanding of how to create more equitable systems. These principles provide a different way of looking at inequity and a frame that sheds light on beliefs, relationships, and interactions that are often overlooked within approaches that lead to a one-dimensional understanding of an issue.

1. A system is not just the sum of its parts

All systems have a structure, and those structures are important. It is the organization and relationships between a system’s parts as much as the parts themselves that determine how the system behaves and the outcomes it produces. The behavior of the overall system can be quite different from the behavior of its individual parts or the sum of these parts. The interactions of the parts within the system create emergent properties which cannot be found in the parts of a system themselves. These emergent system properties are as much a product of the interaction and organization of a system’s parts as the nature of the parts themselves.

In the United States, advantages and disadvantages based on race and ethnicity are primarily driven by structures of access and opportunity within society. These structures shape and constrain the choices and life chances of racial and ethnic minorities. A systems perspective assumes there is a fundamental interdependence between the factors that shape access and opportunity, recognizing the relationships between racially and economically isolated neighborhoods and employment, mental health, and educational outcomes, for example. We must always account for systemic effects as well as the effects of the parts of a system, otherwise solutions will fail in their intended goal of shifting system behavior toward greater equity.

2. Cause and effect are not linear

The traditional view of causality, often popularized by logic models, is linear with an emphasis on identifying and isolating individual, closely aligned causes for any given effect. The linear model of causation is so deeply ingrained that it manifests itself in a series of implicit assumptions about how the world works. Causality flows in one direction – from inputs to actions to outputs and outcomes.

Systems thinking is a shift toward recognizing interdependence, that each effect can have multiple causes, and each cause can have multiple effects. Outcomes are a product of mutual and multiple interactions within the system. Multiple causation is the recognition that there are no ultimate causes for any given system outcome. Rather, outcomes are the product of many causes interacting over time. Mutual causation is the recognition that outcomes are often the result of causes acting together to produce an effect.

An emphasis on searching for a single cause of inequity causes us to overlook the ways in which systematic advantages and disadvantages are produced through the interaction of many causes. For example, the educational funding system in the United States is based on local property taxes which has resulted in large disparities in per-pupil spending between predominantly White districts and predominantly Black districts. A systems approach accounts for the linkages between past housing discrimination, housing value, and school segregation, and poor educational outcomes. The search for a single cause, however, such as interpersonal racism or explicit racist intent, might overlook the way in which a school funding formula is grounded in racial and economic segregation and thereby fostered racial disparities.

3. Outcomes are cumulative and delayed

In the linear model, we often use cues such as proximity of a cause to an effect in time and space to determine if that cause led to the observed outcome. The attempt to isolate causation to a single domain at a single point in time, however, is difficult in complex systems. Delayed effects are common as actions work their way through the system. Cumulative causation occurs where a cause within one domain (education) may influence other causes within that domain over time, or in other domains (school pushout in education influencing juvenile justice outcomes); thus, the initial causes produce much larger systemic effects. A systems perspective shows us that inequities that exist within one domain tends to accumulate across domains and over time because of the relationships and interactions that exist among them. Therefore, the cumulative “level” of inequity is not simply the sum of instances of discrimination or disadvantage but is often much greater.

The practice of redlining in the United States and its connection to the subprime loan crisis in the early 2000s is an example of a time delayed, cumulative inequity. Redlining was a practice that was common in the 1900s where mortgage banks would only approve loans for Black applicants in racially homogeneous neighborhoods, resulting in the residential segregation and neighborhood effects which still exist today. The relationship between race and neighborhood that was created in part by redlining created space for the practice of reverse redlining to develop in the late 1990s and early 2000s. Mortgage companies targeted racially segregated neighborhoods for discriminatory lending yet without the original redlining policy, reverse redlining would not have been possible. As a result, Black homeowners disproportionately experienced the brunt of the subprime loan crisis and resulting foreclosures.

4. Systems do not want to change but the right action can have big effects

Complex systems have feedback loops where the output of the system, or part of the system, is a new input back into the system. These negative feedback loops generally promote stability and settling to equilibrium Thus, negative feedback loops can allow the system to adapt to changing conditions and undermine efforts to create change.

Consider the United States’ legacy of school desegregation efforts. Courts ordered racial integration within districts across the country to correct legalized school segregation. In response, however, White families fled the inner cities to move beyond the district boundaries or placed their children in private schools. Today, years after Brown v. Board of Education was decided, many school districts remain highly segregated by race despite the removal of legal segregation.

This does not mean, however, that one cannot change a system. When a system fails to adapt to new inputs or its self-correction is interrupted, a system can be transformed. This can often be accomplished through a positive feedback loop. Positive feedback is an intervention that has a self-reinforcing, magnifying, or snowball effect.

Leverage points are the places within a complex system where a small shift can produce a large change throughout the system. The key to finding leverage points is to not look at the inequity problem in isolation, but to incorporate the history, dynamics, and structure of a given problem to craft effective and lasting solutions. Positive feedback loops are important leverage points because minor actions can create large results. Weakening or strengthening positive feedback loops can adjust relationships that in turn change or create system behaviors.

The area of wealth inequity is an example of a positive feedback loop where intervention at key leverage points can have large effects. In essence, the positive feedback loop is grounded in the fact that existing wealth through the structure of the American economic system creates more wealth. Interventions such as progressive taxation, high quality public education not tied to property taxes, and community ownership are all measures that can attenuate the positive feedback loop that increases wealth for those that already have wealth, widening access to economic resources.

5. Act with caution and reflect on the results

Change agents should take great care when acting to reduce the potential harmful effects of their (well-intended) systems interventions. Just as a problem in one domain may create problems in another domain, solving a problem in one domain may alleviate problems in another domain. Similarly, because effects are multiple, often difficult to predict, interconnected, and delayed in time, what seems like a promising solution could ultimately exacerbate inequities within another domain or for a specific group. Changes agents should undergo a deliberate process to identify root causes of the inequities and assess the equity impact of existing policies, practices, budgets, plans, and decisions as well as proposed systems interventions. They should also undergo a process of continuous strategy improvement by tracking outcomes to assess progress in reducing inequities, identify and address systems responses to change, and then adapt equity strategies as needed to address those system responses.


The principles outlined in this article are essential to understanding the production of more equitable systems in American society. Although interpersonal discrimination persists, inequity is primarily a function of structures of access and opportunity. A systems perspective helps us understand how inequity is born, builds, and resists efforts to change by allowing us to see the world in terms of wholes, rather than parts as well as and how the parts work together to produce inequitable outcomes. It also helps us see how some efforts, even if well-intentioned, can undermine the vision of advancing equity if they are not well designed. Systems thinking does not mean that we cannot act or should fear action but that we need to think about the way we act differently.

Share |