Taxing Applesauce Won't Solve Obesity

In an editorial published November 18, the journal declared that "the rise of ultra processed foods in human diets is damaging public health, fueling chronic diseases worldwide, and deepening health inequalities." The prescription follows predictably: mandatory warning labels, marketing bans, restrictions in public institutions and higher taxes on ultra processed foods. Policymakers will listen. They always do when The Lancet speaks. And therein lies the problem.

The Lancet is not alone in this framing. The World Health Organization is now assembling an expert panel to develop global guidance on ultra processed foods for UN member states. The 2025 U.S. Dietary Guidelines Advisory Committee has for the first time included ultra processed foods as an evaluation topic, defaulting to the NOVA classification system because it is "most widely used in research." Public health advocates from São Paulo to Sydney are calling for taxes, bans and warning labels with increasing urgency. A consensus is forming.

But consensus is not the same as precision. And precision is what this conversation desperately needs.

Job Ubbink, professor and head of the Department of Food Science and Nutrition at the University of Minnesota, has articulated this distinction with precision in peer reviewed research. Writing in Obesity Science & Practice, Ubbink and co author Allen Levine argue that "in most cases it is the formulation more than the processing that results in foods that are not recommended as part of a healthy diet." They propose that we should be talking about "ultra formulated" foods, products engineered with high levels of added sugars, refined carbohydrates, saturated fats and sodium, rather than conflating them with foods that have simply undergone intensive processing.

This is not semantic hairsplitting. It is the difference between coherent policy and regulatory theater.

Consider a simple example: an apple transformed into unsweetened applesauce. The processing involved is significant. The apple is washed, peeled, cored, cooked and puréed. Some heat sensitive nutrients are lost. Enzymes are denatured. By many European classification schemes, this product registers as highly processed, potentially a NOVA Group 4 ultra processed food depending on interpretation. Yet nutritionally, unsweetened applesauce remains remarkably close to its source material. It is the same apple, transformed for preservation and distribution.

Now add sugar. Add high fructose corn syrup. Add artificial colors and flavors. The processing has not changed, but the formulation has, and that formulation is what makes the product problematic for human health. The NOVA system and the broader public health consensus fail to adequately distinguish between these two products, and that failure will propagate into policy.

The International Union of Food Science and Technology formed a task force specifically to address what they identified as ten key issues with the NOVA classification system, including its lack of quantification of processing levels, weak scientific grounding and, critically, its "misrepresentation of processing as inherently negative." Britannica's own assessment of NOVA notes its "lack of clear criteria, its oversimplification of relationships between processed foods and health, and its lack of accounting for food access, affordability, and cultural dietary patterns." These are not fringe criticisms. They represent a growing scientific recognition that NOVA, while influential, is not fit for the policy purposes to which it is being applied.

There is a troubling disregard for history in the current discourse around ultra processed foods. Food processing is not a twentieth century corporate invention designed to extract profit at the expense of public health. It is the foundation of civilization itself.

In 1795, the French government offered 12,000 francs to anyone who could develop a reliable method of preserving food for Napoleon's armies. Nicolas Appert, a confectioner and brewer, spent fifteen years experimenting before claiming the prize in 1809 with a method that seems remarkably familiar today: sealing food in glass jars and heating them. What makes Appert's achievement remarkable is that he had absolutely no idea why his method worked. He thought air caused spoilage, so he simply aimed to keep it out. It took another fifty years for Louis Pasteur to demonstrate that microorganisms were the real culprits. Appert's process worked perfectly despite his lack of understanding of the underlying science.

This was not an attempt to create an obesity epidemic. It was an attempt to solve one of humanity's oldest problems: how to feed populations beyond the immediate reach of fresh harvests. Appert's work led to canning. It led, indirectly, to pasteurization. It made possible the feeding of cities, the survival of populations through winters and famines and eventually the global food distribution networks that now provision eight billion people.

Processing has always been about preservation, shelf stability, safety and access. Drying, salting, smoking, fermenting, pickling: every culture on earth developed processing techniques because the alternative was starvation during seasons when fresh food was unavailable. The industrial revolution did not invent food processing; it scaled it.

The Lancet acknowledges that "industrial production, processing, and transport of agri commodities are fossil fuel intensive systems," positioning this as an indictment. But the alternative, a world in which food cannot travel, cannot be preserved, cannot reach populations distant from agricultural production, is not a return to pastoral health. It is a return to food insecurity on a scale that modern populations have largely forgotten.

Here is where the emerging policy consensus becomes not merely incomplete but actively dangerous. The Lancet states that "equity must be central when addressing the challenge of UPFs" and notes that "consumption tends to be higher among people facing economic hardship." The proposed solution? Taxation, with revenues redirected toward "cash transfers for whole foods and other minimally processed foods to help protect low income households."

This sounds progressive. It is, in practice, regressive, and reveals a fundamental misunderstanding of how food systems actually function in underserved communities.

Food deserts are real. They are growing. In many urban and rural areas across America and globally, access to fresh produce is limited by geography, infrastructure and economics. For populations in these areas, shelf stable processed foods are not a lifestyle choice reflecting poor judgment or corporate manipulation. They are a nutritional lifeline.

Return to the applesauce example. A fresh apple, while arguably more nutritious, has limited distribution range and a perishability window measured in weeks. It requires cold chain logistics that many food deserts lack. It costs more per calorie delivered. Unsweetened applesauce, by contrast, can sit on a shelf for years, travel anywhere and deliver fruit based nutrition at a fraction of the cost. A policy regime that taxes this product as "ultra processed" does not protect vulnerable consumers. It prices them out of nutritional options while subsidizing fresh produce they may have no practical means to access.

The Lancet calls for "redirecting agricultural subsidies away from large, transnational corporations" toward "a diverse range of food producers" creating "locally sourced, affordable, minimally processed foods." This vision is appealing in its romanticism. It is also logistically fantastical for the billions of people who do not live within reach of local food producers and cannot afford the premium that "minimally processed" commands at retail.

Even if we could agree on a coherent definition of ultra processed foods, which we cannot given the current state of classification science, we would face another problem the emerging policy consensus does not address: any classification system will be gamed.

Food manufacturers are sophisticated formulators. They employ food scientists and regulatory specialists whose job is to optimize products for market success within whatever constraints regulators impose. If a classification system defines a product as Class 4 (most ultra processed) but a minor reformulation can achieve Class 2 (moderately processed), companies will reformulate. They will do so not because Class 2 is meaningfully healthier, but because Class 2 sells better in a regulatory environment that stigmatizes Class 4.

The result is products that appear healthier by classification while remaining problematic by formulation. A consumer with Type 2 diabetes needs to know whether a product contains high levels of added sugars that will spike their blood glucose. A classification number that has been optimized for marketing rather than metabolic reality tells them nothing useful. They may purchase a Class 2 product believing it aligns with their health needs, when in fact it was simply engineered to achieve that number while retaining the formulation characteristics that drive sales.

This is not speculation. It is the predictable outcome of every nutrient based classification system that has been implemented to date. "Low fat" products proliferated in the 1990s, often replacing fat with sugar to maintain palatability. "No added sugar" products substitute artificial sweeteners that may carry their own metabolic concerns. Classification systems create targets. Formulators hit targets. Consumer health does not necessarily improve.

The Lancet is correct that action is needed. Ultra formulated foods, products engineered for hyperpalatability through combinations of refined sugars, fats and sodium, are contributing to chronic disease burdens worldwide. But the action needed is more sophisticated than the emerging consensus suggests.

First, we need classification systems that distinguish processing from formulation. Ubbink's framework offers a starting point: evaluate foods not by the intensity of their processing but by the health implications of their ingredient matrices. A twin screw extruded corn product with no added oils, salt or sugars is fundamentally different from the same product laden with processed cheese powder and artificial flavors, even though the processing is identical.

Second, we need systems that account for consumer context. A product that is perfectly acceptable for a metabolically healthy adult may be contraindicated for someone with diabetes, hypertension or other diet sensitive conditions. Static classification numbers cannot capture this variability. What we need are dynamic filtering systems that help individual consumers understand whether a specific product aligns with their specific health profile, not whether it meets some abstract threshold of processing.

Third, we need to acknowledge that modern computational systems make this kind of nuanced classification possible in ways that were not feasible even a decade ago. The challenge is no longer technical capacity. It is regulatory and commercial will. Systems capable of analyzing formulations, cross referencing individual health parameters and providing contextual guidance are within reach. The question is whether we will build them, or whether we will instead implement blunt policy instruments that feel decisive while missing the point entirely.

The Lancet describes the UPF industry as "emblematic of a food system that is increasingly controlled by transnational corporations that prioritize corporate profit ahead of public health." This framing positions the debate as a morality play: virtuous public health advocates versus rapacious corporate interests. Reality is more complicated.

Food manufacturers respond to incentives. If the incentive structure rewards hyperpalatable formulations that drive repeat consumption, that is what they will produce. If the incentive structure rewards genuinely healthful products that support consumer wellbeing, many will pivot accordingly, not from altruism but from market logic. The policy question is how to construct that incentive structure, and the answer requires precision that the current recommendations lack.

Taxing "ultra processed foods" as a category will not achieve this precision. It will raise costs on products that range from nutritionally vacant to genuinely healthful, harm consumers who depend on affordable shelf stable options and create classification targets that sophisticated formulators will game. It will generate revenue, create the appearance of action and fail to materially improve public health outcomes.

What would work is harder. It requires classification systems grounded in formulation science rather than processing stigma. It requires consumer facing tools that provide contextual guidance rather than categorical judgments. It requires acknowledging that food processing is not the enemy, it is the infrastructure that feeds the world, while holding the food industry accountable for formulation choices that prioritize palatability over nutrition.

The Lancet has issued its call. Policymakers will respond. The question is whether they will implement policies that match the complexity of the problem, or whether they will settle for the satisfaction of having done something, anything, while the consumers they claim to protect pay the price.

The infrastructure for meaningful UPF classification does not exist yet. But it could. Systems that distinguish processing from formulation, that account for individual health context, that give consumers actual transparency rather than categorical stigma. These are within reach. The question is whether we build them before blunt policy instruments lock in the wrong framework.

I'm working on this problem. If you are too, or if this piece raised questions you want to think through, join the conversation on our Discord server: This debate is just getting started.