When tariffs are announced, headlines tend to follow the drama: winners, losers, political posturing. The reality for those tasked with analysis is less theatrical and more intricate. In the food processing sector—ISIC code 10, “Manufacture of food products”—the effects of new tariffs are anything but uniform, and quantifying them calls for methodical work rather than big declarations.
It starts, as so often, with definitions. ISIC 10 covers a landscape of activity: meat, fish, fruit, vegetables, oils, dairy, bakery, animal feed—the list is long. Each segment relies on a basket of raw inputs, many of which are imported. Some, like wheat or soy, might account for the majority of total costs in a subsector. Others, like packaging or food additives, matter more at the margins but can still tip the balance when prices move. When a government imposes tariffs on one or more of these inputs, the first task is to trace exactly which parts of ISIC 10 are exposed, and to what degree.
Suppose, for example, a country introduces a 15% tariff on imported wheat. For bread manufacturers, this is immediately consequential. For meat processors, less so, unless feed costs are affected downstream. The goal, then, is to build a model that captures these ripple effects step by step—clear enough for policymakers, detailed enough for industry specialists.
A partial equilibrium framework works best. Start by assembling baseline data: output, prices, and employment in ISIC 10, ideally broken down by major subcategories. Input-output tables are indispensable here. They reveal, for instance, what share of bakery costs come from imported wheat, how much of the dairy sector’s cost base is exposed to global sugar prices, or how sensitive animal feed producers are to soybean imports.
Once the tariff is mapped, calculate the direct cost increase for each subsector. This requires multiplying the tariff-induced price hike by the relevant input share. If wheat comprises 30% of a bakery’s input costs, a 15% wheat tariff pushes up overall costs by 4.5%. Not all costs are passed on to consumers, of course—some are absorbed as slimmer margins, some are mitigated through input substitution or efficiency gains. The next step is to estimate the likely split, drawing on elasticity estimates or firm-level financial data where available.
What happens to output? Here, price elasticity of demand is critical. If consumers are sensitive to bread prices, higher costs will reduce demand, with knock-on effects for output and employment. If demand is relatively inelastic, volume losses may be small but margins squeezed harder. Elasticities vary by product and country—sometimes by orders of magnitude. There is no substitute for country- and product-specific data, though in their absence, sensitivity analysis using a range of plausible values is better than false precision.
Employment effects follow naturally. Lower output means less labor required—although, again, not always in a one-to-one ratio. Some processes are automated, others labor-intensive. Examining firm-level employment and productivity data within ISIC 10 helps refine these estimates. It’s also worth considering lag effects: firms may hold onto workers longer than output justifies, hoping for policy reversals or market recovery.
There’s a temptation, especially under political pressure, to reach for a single headline number: “X jobs lost,” “Y percent decline in output.” The reality is more nuanced. Some subsectors may barely notice the tariff; others may restructure or even thrive if domestic input suppliers benefit from less foreign competition. Substitution is a live possibility—manufacturers might switch from imported to local wheat, or from one type of oil to another, blunting the intended effect of the tariff (or amplifying unintended consequences elsewhere).
Care must be taken with data quality. Not all countries maintain detailed, up-to-date input-output tables or granular ISIC-level employment records. In their absence, analysts may need to piece together estimates from firm surveys, industry associations, or customs data. Each source brings its own caveats—self-reporting bias, time lags, definitional inconsistencies—but triangulation is almost always preferable to relying on a single number.
Finally, it is worth acknowledging what a partial equilibrium model cannot do. It does not capture secondary effects in related sectors (upstream and downstream), changes in household consumption patterns outside food processing, or broader macroeconomic impacts. For some questions, especially those concerning long-term or economy-wide adjustment, a general equilibrium approach is needed. But for targeted, timely analysis of policy shocks in a well-defined sector, partial equilibrium rooted in ISIC codes is often the best tool available.
In the end, the purpose is not just to predict, but to inform better decision-making. By grounding the analysis in ISIC 10, economists can provide clarity amid the noise—offering policymakers a map of who stands to gain, who stands to lose, and what might shift if policies change again tomorrow. The value lies not in perfect foresight, but in disciplined, transparent reasoning—something always in short supply, especially when the subject is as politically charged as food.