In the United States, a significant share of adults wear a connected wristband or use an app to track their health daily. This figure masks a fracture: According to the GSMA 2025 report, 93% of people not connected to mobile internet live in low- or middle-income countries. Biometric data is being used increasingly in certain insurance models, and those unable to produce it bear standard rates.

The essentials

  • Individual biometric data, captured by connected wristbands, is being used with growing frequency in certain insurance models in the United States. The federal framework regulates wellness programs for collective plans subject to HIPAA and the ACA, but its scope is not universal.
  • According to the GSMA 2025 report, 93% of people not connected to mobile internet live in low- or middle-income countries; women represent 58% of the mobile usage gap and are 12% less likely than men to use the internet on mobile.
  • The prices of tracking devices vary significantly by model; some start around $249 and others exceed $400, adding a pricing barrier to the digital divide.
  • The fork in the road remains open: regulation of insurance uses or durable privatization of a health good that has so far been collective.

An insurance market learning to read bodies

The traditional model of health insurance rests on risk pooling. People pay based on broad categories—age, sex, medical history—because that is all an insurer can observe. Connected wristbands change this equation. They produce new granularity: resting heart rate, heart rate variability, sleep hours, daily steps, estimated stress levels. An insurer with access to these data streams can substitute a precise individual profile for a statistical group estimate.

Several major American companies have already taken the step. John Hancock has integrated activity tracking data into its life insurance contracts since 2018. UnitedHealthcare Motion is a voluntary rewards program tied to activity goals, with the form depending on the plan. Biometric data do not yet serve to set health insurance premiums in real time for most contracts, but the direction is clear: the market rewards those who produce favorable data.

What Luigi Zingales documented about regulatory capture applies here with remarkable precision. His work on capitalism and political power, published through ProMarket, the University of Chicago’s forum, describes how dominant actors convert collective goods into private rents as soon as public regulation is absent or lagging. Preventive health has long been treated as a collective good: prevention programs, health screenings, epidemiological data serve the entire population. When these data migrate to private platforms and individual insurance contracts, distributive effects could result depending on the applicable regulatory framework.

The digital divide is also a health divide

Americans who track their health daily are not a representative sample of the population. According to GSMA 2025, globally, 93% of people not connected to mobile internet live in low- or middle-income countries. Even within the United States, inequalities in digital equipment follow the lines of income and ethnicity.

The cost and eligibility of devices depend on the program and model; no universal threshold of $200–$400 is set. At that price, connected devices are unequally adopted across society, but are not reserved solely for high-income households. Women, overrepresented in the lowest income brackets, face two documented disadvantages according to GSMA: they represent 58% of the usage gap between men and women on mobile internet, and are 12% less likely than men to use the same internet on their phone. Pricing incentives can create gaps. In collective programs conditioned on a health factor, the absence of data cannot automatically result in a health-based premium when a reasonable alternative or exception is required.

The mechanism is all the more perverse because it tends to be self-reinforcing. Populations with the least access to tracking devices often face more difficult living conditions, work conditions, food access. Collective wellness programs conditioned on a health factor must offer a reasonable alternative or exception to earn the reward. Those who can prove they walk ten thousand steps a day earn a reward. Populations with less flexible work conditions encounter additional obstacles to accessing these devices and their associated rewards.

This mechanism joins the dynamics described in our analysis of polarization by robots: when a technology amplifies the productivity of already well-equipped workers while remaining inaccessible to others, it deepens the gap instead of closing it.

The absence of federal regulation is not an oversight

There is a federal framework for collective plans: it limits and conditions pricing incentives tied to wellness programs. The most well-known medical privacy law, the federal health information privacy law adopted in 1996, covers data held by healthcare providers, hospitals, and insurers in their administrative functions. Data from a consumer app or wristband are generally not covered by this law, except when the service acts on behalf of a covered entity. The path of data depends on each app, platform, and program; it cannot be generalized from Apple, Google, Garmin, or Fitbit to the insurer.

This regulatory void structures access to biometric data. Insurers can access favorable biometric data through voluntary incentive programs. Collective wellness programs conditioned on a health factor must provide a reasonable alternative or exception according to applicable federal rules.

A competing reading deserves to be stated plainly. Economists close to the competitive markets school, whose work addresses risk pricing based on information, argue that fine differentiation of premiums based on actual health behavior is economically desirable: it aligns incentives, reduces moral hazard, and can in theory lower costs for the most active populations. The argument has internal coherence. It holds as long as one assumes that the decision not to wear a wristband is a free choice. When $200 represents an impossible expense for a household at the bottom of the income distribution, freedom of choice is merely formal.

Differentiated pricing can reflect inequalities of means rather than differences in behavioral choices.

Zingales’s thesis goes beyond the diagnosis of inequality alone. Regulatory capture could concentrate benefits toward certain sector actors, while low-income populations face risks of differentiated pricing. A functioning competitive market would require identical rules of the game for all participants. Unequal access to devices can produce unequal effects under formally identical rules.

Collective data and private data: two opposing logics

A 2026 American Heart Association study suggests that an experimental wristband could detect cardiac arrest in a clinical setting and hints at potential for faster assistance, without establishing the general claim about improved arrhythmia detection or measurable reduction in response times. This clinical value is real. The issue concerns its beneficiaries and the form its organization takes.

A public health program that collects anonymized, aggregated biometric data can feed epidemiological models, identify pockets of cardiovascular risk in certain neighborhoods, direct preventive resources toward populations that need them most. This is a logic of common good. A private insurance program that collects the same data at the individual level to adjust a premium produces the opposite effect: it tends to direct pricing advantages toward those whose risk profile is already favorable. The biometric data is the same object in both cases. Its public or private use can produce distinct effects depending on the applicable regulatory framework.

The connection to digital governance issues is direct here. As shown by the analysis of autonomous robotics regulation, technologies that interact directly with bodies or individual rights require a civil and regulatory responsibility framework that the market alone cannot construct.

2030: what remains possible and what is closing

The fork between two possible trajectories has not yet closed. But regulatory windows have finite lifespans. Two scenarios are taking shape by 2030–2035, and the signals that will allow us to distinguish them are already measurable.

In the first scenario, biometric data remain captive to private platforms. In markets subject to the ACA, access to platform data does not allow the insurer to differentiate premiums based on that data. Some programs offer activity-related rewards. The absence of data cannot, in affected ACA markets, result in a health-based premium; collective programs are subject to specific protections.

Supplemental insurance also covers part of low-income Medicare beneficiaries, but its existence does not demonstrate that stratification linked to biometric data affects them. The signal to watch is the growth of contracts incorporating pricing modulation clauses tied to biometric data, which sector regulators can document.

In the second scenario, a regulatory framework extends protection of biometric data and limits their insurance use. Some states classify certain biometric data as sensitive, but a precise insurance rule must be identified to claim that it governs pricing models. Classification of biometric data as sensitive data exists in some states. European data protection regulation limits automated processing that produces legal effects based on health data. In the United States, the federal level remains the only one that can create a uniform floor.

Public programs for subsidized distribution of health tracking devices, conditioned on data management under public and non-commercial control, could complement this regulation on the supply side. The equipment rate gap between the wealthiest and least wealthy households, in countries where such policies have been implemented, is the second signal to follow.

The first scenario has the force of inertia. Insurance and health technology sector actors have substantial resources and an established presence in regulatory processes. The second scenario needs a political coalition that does not yet clearly exist: health regulators, personal data advocates, and consumer organizations that agree on the common diagnosis before diverging on remedies. This diagnosis—biometric data as private insurance rent—is beginning to be raised in U.S. health policy debates, particularly around discussions on extending health data protection to the app economy. This is an opening, not a guarantee.

International comparison adds useful pressure. The GDPR heavily restricts processing of health data and solely automated decisions, but does not establish a general prohibition on insurance profiling. If this rule produces measurable effects on insurance access equity in Europe over the next few years, it will provide a credible counterexample to the argument that regulation of this type is economically infeasible. The digital disconnection data documented by GSMA could then serve to calibrate where the public effort to make access available should focus as a priority.

What is clear right now: the absence of a regulatory framework could allow a multiplication of contracts incorporating pricing modulation clauses tied to biometric data. The window to choose remains open, but narrows as such models become embedded in the market.

Sources

  1. GSMA, State of Mobile Connectivity 2026, via Telefónica Communication Room: https://www.telefonica.com/en/communication-room/blog/coverage-usage-gsma-state-mobile-connectivity-2026/
  2. Luigi Zingales, work on regulatory capture and political capitalism, ProMarket (Stigler Center, University of Chicago Booth School of Business): https://promarket.org/contributor/luigi-zingales/
  3. American Heart Association, studies on continuous biometric monitoring and arrhythmia detection, 2025–2026 (publications in Circulation and JAHA, no directly verifiable URL here)
  4. John Hancock, Vitality program and life insurance based on activity data (public announcement 2018, no directly verifiable URL here)
  5. UnitedHealthcare, Motion program and premium modulation tied to physical activity data (sector documentation)